autonomous master thesis

• Current Members
• Off-Campus Students
• Robot Videos
• Funded Projects
• Publications by Year
• Publications by Type
• Robot Learning Lecture
• Robot Learning IP
• Humanoid Robotics Seminar
• Research Oberseminar
• New, Open Topics
• Ongoing Theses
• Completed Theses
• External Theses
• Advice for Thesis Students
• Thesis Checklist and Template
• Jobs and Open Positions
• Current Openings
• Information for Applicants
• Apply Here!
• TU Darmstadt Student Hiwi Jobs
• Contact Information

Currently Available Theses Topics

We offer these current topics directly for Bachelor and Master students at TU Darmstadt who can feel free to DIRECTLY contact the thesis advisor if you are interested in one of these topics. Excellent external students from another university may be accepted but are required to first email Jan Peters before contacting any other lab member for a thesis topic. Note that we cannot provide funding for any of these theses projects.

We highly recommend that you do either our robotics and machine learning lectures ( Robot Learning , Statistical Machine Learning ) or our colleagues ( Grundlagen der Robotik , Probabilistic Graphical Models and/or Deep Learning). Even more important to us is that you take both Robot Learning: Integrated Project, Part 1 (Literature Review and Simulation Studies) and Part 2 (Evaluation and Submission to a Conference) before doing a thesis with us.

In addition, we are usually happy to devise new topics on request to suit the abilities of excellent students. Please DIRECTLY contact the thesis advisor if you are interested in one of these topics. When you contact the advisor, it would be nice if you could mention (1) WHY you are interested in the topic (dreams, parts of the problem, etc), and (2) WHAT makes you special for the projects (e.g., class work, project experience, special programming or math skills, prior work, etc.). Supplementary materials (CV, grades, etc) are highly appreciated. Of course, such materials are not mandatory but they help the advisor to see whether the topic is too easy, just about right or too hard for you.

Only contact *ONE* potential advisor at the same time! If you contact a second one without first concluding discussions with the first advisor (i.e., decide for or against the thesis with her or him), we may not consider you at all. Only if you are super excited for at most two topics send an email to both supervisors, so that the supervisors are aware of the additional interest.

FOR FB16+FB18 STUDENTS: Students from other depts at TU Darmstadt (e.g., ME, EE, IST), you need an additional formal supervisor who officially issues the topic. Please do not try to arrange your home dept advisor by yourself but let the supervising IAS member get in touch with that person instead. Multiple professors from other depts have complained that they were asked to co-supervise before getting contacted by our advising lab member.

NEW THESES START HERE

Blending Deep Generative Models using Stochastic Optimization

Topic: This Master Thesis aims to explore the blending of deep generative models using stochastic optimization techniques [1], focusing on reactive motion generation for robotics. The research will encompass the training of deep generative models, such as Score-based [2], or Flow-based models, specifically utilizing the JAX framework for efficient computation. A significant part of the thesis will involve deriving a mixture of experts algorithm, which leverages these trained generative models in combination with other manually specified objectives to enhance the performance of the motion generator. This integration aims to create more adaptive and responsive robotic behaviors in dynamic environments, offering a substantial advancement over existing methods.

Requirements

• Strong Python programming skills
• Knowledge in Machine Learning
• Experience with deep learning libraries and JAX is a plus

Interested students can apply by sending an e-mail to [email protected] and attaching the documents mentioned below:

• Curriculum Vitae
• Motivation letter explaining why you would like to work on this topic and why you are the perfect candidate

References [1] Hansel, K.; Urain, J.; Peters, J.; Chalvatzaki, G. (2023). Hierarchical Policy Blending as Inference for Reactive Robot Control, 2023 IEEE International Conference on Robotics and Automation (ICRA), IEEE. [2] Urain, J.; Funk, N.; Peters, J.; Chalvatzaki G (2023). SE(3)-DiffusionFields: Learning smooth cost functions for joint grasp and motion optimization through diffusion, International Conference on Robotics and Automation (ICRA).

Self-supervised learning of a visual object-centric representation for robotic manipulation

Scope: External Master Thesis 🇫🇷 This master thesis will be conducted with our French partners at Ecole Centrale de Lyon . Possibility of ERASMUS scholarship. Advisor: Alexandre Chapin , Liming Chen , Emmanuel Dellandrea Added: 2024-07-15 Start: ASAP Topic:

Vision-based learning for robotic manipulation often relies on holistic visual scene representations, where the environment is depicted as a single vector. This method is suboptimal for handling diverse scenes and objects in unconstrained environments. Better representations can improve generalization and data efficiency in robotic learning [1]. Inspired by human perception, object-centric representation has been developed to represent environments with multiple vectors, each corresponding to an object's properties [2]. However, these methods mainly use synthetic datasets [3, 4, 5] and struggle with real-world scenarios [6]. With advances in self-supervised learning for vision models [7, 8], which show promise for object discovery, we propose pre-training an object-centric representation using self-supervised methods to scale to real-world scenarios. This thesis will focus on: Developing and training an object-centric self-supervised model on a real-world dataset. Pre-training the model on a real-world robotic dataset. Applying the pre-trained model to visual-based robotic manipulation tasks.

Interested students can apply by sending the required documents to [email protected] and attaching the required documents mentioned below.

• Experience with the Pytorch library

Preferred Qualifications

• Prior experience in Computer Vision and/or Robotics is preferred
• Use of distributed environment for learning of models (SLURM)
• Knowledge on recent self-supervised learning methods for vision [7, 8]

Required Documents

References [1] O. Kroemer et al. “A review of robot learning for manipulation: Challenges, representations, and algorithms” (2019) [2] F. Locatello et al. “Object-centric learning with Slot Attention” (2020) [3] G. Singh et al. “Illiterate DALL-E learns to compose” (2021) [4] T. Kipf et al. “Conditional object-centric learning from video” (2022) [5] G. Singh et al. “Simple Unsupervised Object-Centric Learning for Complex and Naturalistic Videos” (2022) [6] Z. Wu et al. “SlotDiffusion: Object-Centric Generative Modeling with Diffusion Models” (2023) [7] M. Caron et al. “Emerging Properties in Self-Supervised Vision Transformers” (2021) [8] O. J. Hénaff et al. “Object discovery and representation networks” (2022)

Data-Driven Bimanual Robotic Grasping

Scope: Bachelor/Master thesis Advisor: Vignesh Prasad and Alap Kshirsagar Added: 2024-04-25 Start: ASAP Topic:

Grasping is one of the most fundamental and challenging tasks in the robotic manipulation of objects. Most of the prior work on robotic grasping has focused on grasping with a single gripper and several large-scale datasets have been developed in recent years to tackle the problem of single-arm grasping in 3D by utilizing deep-learning techniques [1,2]. But many tasks in industrial and domestic environments require bimanual grasps. Bimanual grasps are required for manipulation of large, deformable or fragile objects. This project seeks to develop a data-driven technique for bimanual robotic grasp generation from visual input. We will utilize a large-scale dataset of simulated bimanual grasps [3] to train a bimanual grasp pose generation model. The method will be evaluated in simulation as well as on a real robot.

• Knowledge in Machine Learning / Supervised Learning
• Experience with deep learning libraries is a plus

Interested students can apply by sending an e-mail to [email protected] and attaching the documents mentioned below:

References [1] C. Eppner, A. Mousavian, and D. Fox, “ACRONYM: A Large-Scale Grasp Dataset Based on Simulation,” in Proceedings - IEEE International Conference on Robotics and Automation, 2021, vol. 2021-May, pp. 6222–6227, doi: 10.1109/ICRA48506.2021.9560844. [2] A. Mousavian, C. Eppner, and Di. Fox, “6-DOF GraspNet: Variational grasp generation for object manipulation,” in Proceedings of the IEEE International Conference on Computer Vision, 2019, vol. 2019-Octob, pp. 2901–2910, doi: 10.1109/ICCV.2019.00299. [3] G. Zhai et al., “{DA2} Dataset: Toward Dexterity-Aware Dual-Arm Grasping,” IEEE Robot. Autom. Lett., vol. 7, no. 4, pp. 8941–8948, 2022.

Imitation Learning for High-Speed Robot Air Hockey

Scope: Master thesis Advisor: Puze Liu and Julen Urain De Jesus Start: ASAP Topic:

High-speed reactive motion is one of the fundamental capabilities of robots to achieve human-level behavior. Optimization-based methods suffer from real-time requirement when the problem is non-convex and contains constraints. Reinforcement learning requires extensive reward engineering to achieve the desired performance. Imitation learning, on the other hand, gathers human knowledge directly from data collection and enables robots to learn natural movements efficiently. In this paper, we explore how imitation learning can be performed in a complex robot Air Hockey Task. The robot needs to learn not only low-level skills, but also high-level tactics from human demonstrations.

• Good Knowledge in Robotics

References * Chi, Cheng, et al. "Diffusion policy: Visuomotor policy learning via action diffusion." arXiv preprint arXiv:2303.04137 (2023). * Liu, Puze, et al. "Robot reinforcement learning on the constraint manifold." Conference on Robot Learning. PMLR (2022). * Pan, Yunpeng, et al. "Imitation learning for agile autonomous driving." The International Journal of Robotics Research 39.2-3 (2020). Interested students can apply by sending an e-mail to [email protected] and attaching the required documents mentioned above.

Walk your network: investigating neural network’s location in Q-learning methods.

Scope: Master thesis Advisor: Theo Vincent Start: Flexible Topic:

Q-learning methods are at the heart of Reinforcement Learning. They have been shown to outperform humans on some complex tasks such as playing video games [1]. In robotics, where the action space is in most cases continuous, actor-critic methods are relying on Q-learning methods to learn the critic [2]. Although Q-learning methods have been extensively studied in the past, little focus has been placed on the way the online neural network is exploring the space of Q functions. Most approaches focus on crafting a loss that would make the agent learn better policies [3]. Here, we offer a thesis that focuses on the position of the online Q neural network in the space of Q functions. The student will first investigate this idea on simple problems before comparing the performance to strong baselines such as DQN or REM [1, 4] on Atari games. Depending on the result, the student might as well get into MuJoCo and compare the results with SAC [2]. The student will be welcome to propose some ideas as well.

Highly motivated students can apply by sending an email to [email protected] . Please attach your CV, a grade sheet and clearly state why you are interested in this topic. Students who have followed the Reinforcement Learning or Robot Learning course will be prioritized.

• Knowledge in Reinforcement Learning

References [1] Mnih, Volodymyr, et al. "Human-level control through deep reinforcement learning." nature 518.7540 (2015): 529-533. [2] Haarnoja, Tuomas, et al. "Soft actor-critic: Off-policy maximum entropy deep reinforcement learning with a stochastic actor." International conference on machine learning. PMLR, 2018. [3] Hessel, Matteo, et al. "Rainbow: Combining improvements in deep reinforcement learning." Proceedings of the AAAI conference on artificial intelligence. Vol. 32. No. 1. 2018. [4] Agarwal, R., Schuurmans, D. & Norouzi, M.. (2020). An Optimistic Perspective on Offline Reinforcement Learning International Conference on Machine Learning (ICML).

Co-optimizing Hand and Action for Robotic Grasping of Deformable objects

This project aims to advance deformable object manipulation by co-optimizing robot gripper morphology and control policies. The project will involve utilizing existing simulation environments for deformable object manipulation [2] and implementing a method to jointly optimize gripper morphology and grasp policies within the simulation.

Required Qualification:

• Familiarity with deep learning libraries such as PyTorch or Tensorflow

Preferred Qualification:

• Attendance of the lectures "Statistical Machine Learning", "Computational Engineering and Robotics" and "Robot Learning"

Application Requirements:

Interested students can apply by sending an e-mail to [email protected] and attaching the required documents mentioned above.

References: [1] Xu, Jie, et al. "An End-to-End Differentiable Framework for Contact-Aware Robot Design." Robotics: Science & Systems. 2021. [2] Huang, Isabella, et al. "DefGraspNets: Grasp Planning on 3D Fields with Graph Neural Nets." arXiv preprint arXiv:2303.16138 (2023).

Geometry-Aware Diffusion Models for Robotics

In this thesis, you will work on developing an imitation learning algorithm using diffusion models for robotic manipulation tasks, such as the ones in [2, 3, 4], but taking into account the geometry of the task space.

If this sounds interesting, please send an email to [email protected] and [email protected] , and possibly attach your CV, highlighting the relevant courses you took in robotics and machine learning.

What's in it for you:

• You get to work on an exciting topic at the intersection of deep-learning and robotics
• We will supervise you closely throughout your thesis
• Depending on the results, we will aim for an international conference publication

Requirements:

• Be motivated -- we will support you a lot, but we expect you to contribute a lot too
• Robotics knowledge
• Experience setting up deep learning pipelines -- from data collection, architecture design, training, and evaluation
• PyTorch -- especially experience writing good parallelizable code (i.e., runs fast in the GPU)

References: [1] https://arxiv.org/abs/2112.10752 [2] https://arxiv.org/abs/2308.01557 [3] https://arxiv.org/abs/2209.03855 [4] https://arxiv.org/abs/2303.04137 [5] https://arxiv.org/abs/2205.09991

Learning Latent Representations for Embodied Agents

Interested students can apply by sending an E-Mail to [email protected] and attaching the required documents mentioned below.

• Experience with TensorFlow/PyTorch
• Familiarity with core Machine Learning topics
• Experience programming/controlling robots (either simulated or real world)
• Knowledgeable about different robot platforms (quadrupeds and bipedal robots)
• Resume / CV
• Cover letter explaining why this topic fits you well and why you are an ideal candidate

References: [1] Ho and Ermon. "Generative adversarial imitation learning" [2] Arenz, et al. "Efficient Gradient-Free Variational Inference using Policy Search"

Characterizing Fear-induced Adaptation of Balance by Inverse Reinforcement Learning

Interested students can apply by sending an E-Mail to [email protected] and attaching the required documents mentioned below.

• Basic knowledge of reinforcement learning
• Hand-on experience with reinforcement learning or inverse reinforcement learning
• Cognitive science background

References: [1] Maki, et al. "Fear of Falling and Postural Performance in the Elderly" [2] Davis et al. "The relationship between fear of falling and human postural control" [3] Ho and Ermon. "Generative adversarial imitation learning"

Timing is Key: CPGs for regularizing Quadruped Gaits learned with DRL

To tackle this problem we want to utilize Central Pattern Generators (CPGs), which can generate timings for ground contacts for the four feet. The policy gets rewarded for complying with the contact patterns of the CPGs. This leads to a straightforward way of regularizing and steering the policy to a natural gait without posing too strong restrictions on it. We first want to manually find fitting CPG parameters for different gait velocities and later move to learning those parameters in an end-to-end fashion.

Highly motivated students can apply by sending an E-Mail to [email protected] and attaching the required documents mentioned below.

Minimum Qualification:

• Good Python programming skills
• Basic knowledge of the PyTorch library
• Basic knowledge of Reinforcement Learning
• Good knowledge of the PyTorch library
• Basic knowledge of the MuJoCo simulator

References: [1] Cheng, Xuxin, et al. "Extreme Parkour with Legged Robots."

Damage-aware Reinforcement Learning for Deformable and Fragile Objects

Goal of this thesis will be the development and application of a model-based reinforcement learning method on real robots. Your tasks will include: 1. Setting up a simulation environment for deformable object manipulation 2. Utilizing existing models for stress and deformability prediction[1] 3. Implementing a reinforcement learning method to work in simulation and, if possible, on the real robot methods.

If you are interested in this thesis topic and believe you possess the necessary skills and qualifications, please submit your application, including a resume and a brief motivation letter explaining your interest and relevant experience. Please send your application to [email protected].

Required Qualification :

• Enthusiasm for and experience in robotics, machine learning, and simulation
• Strong programming skills in Python

Desired Qualification :

• Attendance of the lectures "Statistical Machine Learning", "Computational Engineering and Robotics" and (optionally) "Robot Learning"

References: [1] Huang, I., Narang, Y., Bajcsy, R., Ramos, F., Hermans, T., & Fox, D. (2023). DefGraspNets: Grasp Planning on 3D Fields with Graph Neural Nets. arXiv preprint arXiv:2303.16138.

Imitation Learning meets Diffusion Models for Robotics

The objective of this thesis is to build upon prior research [2, 3] to establish a connection between Diffusion Models and Imitation Learning. We aim to explore how to exploit Diffusion Models and improve the performance of Imitation learning algorithms that interact with the world.

We welcome highly motivated students to apply for this opportunity by sending an email expressing their interest to Firas Al-Hafez ( [email protected] ) Julen Urain ( [email protected] ). Please attach your letter of motivation and CV, and clearly state why you are interested in this topic and why you are the ideal candidate for this position.

Required Qualification : 1. Strong Python programming skills 2. Basic Knowledge in Imitation Learning 3. Interest in Diffusion models, Reinforcement Learning

Desired Qualification : 1. Attendance of the lectures "Statistical Machine Learning", "Computational Engineering and Robotics" and/or "Reinforcement Learning: From Fundamentals to the Deep Approaches"

References: [1] Song, Yang, and Stefano Ermon. "Generative modeling by estimating gradients of the data distribution." Advances in neural information processing systems 32 (2019). [2] Ho, Jonathan, and Stefano Ermon. "Generative adversarial imitation learning." Advances in neural information processing systems 29 (2016). [3] Garg, D., Chakraborty, S., Cundy, C., Song, J., & Ermon, S. (2021). Iq-learn: Inverse soft-q learning for imitation. Advances in Neural Information Processing Systems, 34, 4028-4039. [4] Chen, R. T., & Lipman, Y. (2023). Riemannian flow matching on general geometries. arXiv preprint arXiv:2302.03660.

• Be extremely motivated -- we will support you a lot, but we expect you to contribute a lot too

Scaling Behavior Cloning to Humanoid Locomotion

Scope: Bachelor / Master thesis Advisor: Joe Watson Added: 2023-10-07 Start: ASAP Topic: In a previous project [1], I found that behavior cloning (BC) was a surprisingly poor baseline for imitating humanoid locomotion. I suspect the issue may lie in the challenges of regularizing high-dimensional regression.

The goal of this project is to investigate BC for humanoid imitation, understand the scaling issues present, and evaluate possible solutions, e.g. regularization strategies from the regression literature.

The project will be building off Google Deepmind's Acme library [2], which has BC algorithms and humanoid demonstration datasets [3] already implemented, and will serve as the foundation of the project.

To apply, email [email protected] , ideally with a CV and transcript so I can assess your suitability.

• Experience, interest and enthusiasm for the intersection of robot learning and machine learning
• Experience with Acme and JAX would be a benefit, but not necessary

References: [1] https://arxiv.org/abs/2305.16498 [2] https://github.com/google-deepmind/acme [3] https://arxiv.org/abs/2106.00672

Robot Gaze for Communicating Collision Avoidance Intent in Shared Workspaces

Scope: Bachelor/Master thesis Advisor: Alap Kshirsagar , Dorothea Koert Added: 2023-09-27 Start: ASAP

Topic: In order to operate close to non-experts, future robots require both an intuitive form of instruction accessible to lay users and the ability to react appropriately to a human co-worker. Instruction by imitation learning with probabilistic movement primitives (ProMPs) [1] allows capturing tasks by learning robot trajectories from demonstrations including the motion variability. However, appropriate responses to human co-workers during the execution of the learned movements are crucial for fluent task execution, perceived safety, and subjective comfort. To facilitate such appropriate responsive behaviors in human-robot interaction, the robot needs to be able to react to its human workspace co-inhabitant online during the execution. Also, the robot needs to communicate its motion intent to the human through non-verbal gestures such as eye and head gazes [2][3]. In particular for humanoid robots, combining motions of arms with expressive head and gaze directions is a promising approach that has not yet been extensively studied in related work.

Goals of the thesis:

• Develop a method to combine robot head/gaze motion with ProMPs for online collision avoidance
• Implement the method on a Franka-Emika Panda Robot
• Evaluate and compare the implemented behaviors in a study with human participants

Highly motivated students can apply by sending an email to [email protected]. Please attach your CV and transcript, and clearly state your prior experiences and why you are interested in this topic.

• Strong Programming Skills in python
• Prior experience with Robot Operating System (ROS) and user studies would be beneficial
• Strong motivation for human-centered robotics including design and implementation of a user study

References : [1] Koert, Dorothea, et al. "Learning intention aware online adaptation of movement primitives." IEEE Robotics and Automation Letters 4.4 (2019): 3719-3726. [2] Admoni, Henny, and Brian Scassellati. "Social eye gaze in human-robot interaction: a review." Journal of Human-Robot Interaction 6.1 (2017): 25-63. [3] Lemasurier, Gregory, et al. "Methods for expressing robot intent for human–robot collaboration in shared workspaces." ACM Transactions on Human-Robot Interaction (THRI) 10.4 (2021): 1-27.

Tactile Sensing for the Real World

Topic: Tactile sensing is a crucial sensing modality that allows humans to perform dexterous manipulation[1]. In recent years, the development of artificial tactile sensors has made substantial progress, with current models relying on cameras inside the fingertips to extract information about the points of contact [2]. However, robotic tactile sensing is still a largely unsolved topic despite these developments. A central challenge of tactile sensing is the extraction of usable representations of sensor readings, especially since these generally contain an incomplete view of the environment.

Recent model-based reinforcement learning methods like Dreamer [3] leverage latent state-space models to reason about the environment from partial and noisy observations. However, more work has yet to be done to apply such methods to real-world manipulation tasks. Hence, this thesis will explore whether Dreamer can solve challenging real-world manipulation tasks by leveraging tactile information. Initial results suggest that tasks like peg-in-a-hole can indeed be solved with Dreamer in simulation (see figure above), but the applicability of this method in the real world has yet to be shown.

In this work, you will work with state-of-the-art hardware and compute resources on a hot research topic with the option of publishing your work at a scientific conference.

Highly motivated students can apply by sending an email to [email protected]. Please attach a transcript of records and clearly state your prior experiences and why you are interested in this topic.

• Ideally experience with deep learning libraries like JAX or PyTorch
• Experience with reinforcement learning is a plus
• Experience with Linux

References [1] 2S Match Anest2, Roland Johansson Lab (2005), https://www.youtube.com/watch?v=HH6QD0MgqDQ [2] Gelsight Inc., Gelsight Mini, https://www.gelsight.com/gelsightmini/ [3] Hafner, D., Lillicrap, T., Ba, J., & Norouzi, M. (2019). Dream to control: Learning behaviors by latent imagination. arXiv preprint arXiv:1912.01603.

Large Vision-Language Neural Networks for Open-Vocabulary Robotic Manipulation

Robots are expected to soon leave their factory/laboratory enclosures and operate autonomously in everyday unstructured environments such as households. Semantic information is especially important when considering real-world robotic applications where the robot needs to re-arrange objects as per a set of language instructions or human inputs (as shown in the figure). Many sophisticated semantic segmentation networks exist [1]. However, a challenge when using such methods in the real world is that the semantic classes rarely align perfectly with the language input received by the robot. For instance, a human language instruction might request a ‘glass’ or ‘water’, but the semantic classes detected might be ‘cup’ or ‘drink’.

Nevertheless, with the rise of large language and vision-language models, we now have capable segmentation models that do not directly predict semantic classes but use learned associations between language queries and classes to give us ’open-vocabulary’ segmentation [2]. Some models are especially powerful since they can be used with arbitrary language queries.

In this thesis, we aim to build on advances in 3D vision-based robot manipulation and large open-vocabulary vision models [2] to build a full pick-and-place pipeline for real-world manipulation. We also aim to find synergies between scene reconstruction and semantic segmentation to determine if knowing the object semantics can aid the reconstruction of the objects and, in turn, aid manipulation.

Highly motivated students can apply by sending an e-mail expressing their interest to Snehal Jauhri (email: [email protected]) or Ali Younes (email: [email protected]), attaching your letter of motivation and possibly your CV.

Topic in detail : Thesis_Doc.pdf

Requirements: Enthusiasm, ambition, and a curious mind go a long way. There will be ample supervision provided to help the student understand basic as well as advanced concepts. However, prior knowledge of computer vision, robotics, and Python programming would be a plus.

References: [1] Y. Wu, A. Kirillov, F. Massa, W.-Y. Lo, and R. Girshick, “Detectron2”, https://github.com/facebookresearch/detectron2 , 2019. [2] F. Liang, B. Wu, X. Dai, K. Li, Y. Zhao, H. Zhang, P. Zhang, P. Vajda, and D. Marculescu, “Open-vocabulary semantic segmentation with mask-adapted clip,” in CVPR, 2023, pp. 7061–7070, https://github.com/facebookresearch/ov-seg

Dynamic Tiles for Deep Reinforcement Learning

Linear approximators in Reinforcement Learning are well-studied and come with an in-depth theoretical analysis. However, linear methods require defining a set of features of the state to be used by the linear approximation. Unfortunately, the feature construction process is a particularly problematic and challenging task. Deep Reinforcement learning methods have been introduced to mitigate the feature construction problem: these methods do not require handcrafted features, as features are extracted automatically by the network during learning, using gradient descent techniques.

In simple reinforcement learning tasks, however, it is possible to use tile coding as features: Tiles are simply a convenient discretization of the state space that allows us to easily control the generalization capabilities of the linear approximator. The objective of this thesis is to design a novel algorithm for automatic feature extraction that generates a set of features similar to tile coding, but that can arbitrarily partition the state space and deal with arbitrary complex state space, such as images. The idea is to combine the feature extraction problem directly with Linear Reinforcement Learning methods, defining an algorithm that is able both to have the theoretical guarantees and good convergence properties of these methods and the flexibility of Deep Learning approaches.

• Curriculum Vitae (CV);
• A motivation letter explaining the reason for applying for this thesis and academic/career objectives.

Minimum knowledge

• Good Python programming skills;
• Basic knowledge of Reinforcement Learning.

Preferred knowledge

• Knowledge of the PyTorch library;
• Knowledge of the Atari environments (ale-py library).
• Knowledge of the MushroomRL library.

Accepted candidate will

• Define a generalization of tile coding working with an arbitrary input set (including images);
• Design a learning algorithm to adapt the tiles using data of interaction with the environment;
• Combine feature learning with standard linear methods for Reinforcement Learning;
• Verify the novel methodology in simple continuous state and discrete actions environments;
• (Optionally) Extend the experimental analysis to the Atari environment setting.

Deep Learning Meets Teleoperation: Constructing Learnable and Stable Inductive Guidance for Shared Control

This work considers policies as learnable inductive guidance for shared control. In particular, we use the class of Riemannian motion policies [3] and consider them as differentiable optimization layers [4]. We analyze (i) if RMPs can be pre-trained by learning from demonstrations [5] or reinforcement learning [6] given a specific context; (ii) and subsequently employed seamlessly for human-guided teleoperation thanks to their physically consistent properties, such as stability [3]. We believe this step eliminates the laborious process of constructing complex policies and leads to improved and generalizable shared control architectures.

Highly motivated students can apply by sending an e-mail expressing your interest to [email protected] and [email protected] , attaching your letter of motivation and possibly your CV.

• Experience with deep learning libraries (in particular Pytorch)
• Knowledge in reinforcement learning and/or machine learning

References: [1] Niemeyer, Günter, et al. "Telerobotics." Springer handbook of robotics (2016); [2] Selvaggio, Mario, et al. "Autonomy in physical human-robot interaction: A brief survey." IEEE RAL (2021); [3] Cheng, Ching-An, et al. "RMP flow: A Computational Graph for Automatic Motion Policy Generation." Springer (2020); [4] Jaquier, Noémie, et al. "Learning to sequence and blend robot skills via differentiable optimization." IEEE RAL (2022); [5] Mukadam, Mustafa, et al. "Riemannian motion policy fusion through learnable lyapunov function reshaping." CoRL (2020); [6] Xie, Mandy, et al. "Neural geometric fabrics: Efficiently learning high-dimensional policies from demonstration." CoRL (2023).

Dynamic symphony: Seamless human-robot collaboration through hierarchical policy blending

This work focuses on arbitration between the user and assistive policy, i.e., shared autonomy. Various works allow the user to influence the dynamic behavior explicitly and, therefore, could not satisfy stability guarantees [3]. We pursue the idea of formulating arbitration as a trajectory-tracking problem that implicitly considers the user's desired behavior as an objective [4]. Therefore, we extend the work of Hansel et al. [5], who employed probabilistic inference for policy blending in robot motion control. The proposed method corresponds to a sampling-based online planner that superposes reactive policies given a predefined objective. This method enables the user to implicitly influence the behavior without injecting energy into the system, thus satisfying stability properties. We believe this step leads to an alternative view of shared autonomy with an improved and generalizable framework.

Highly motivated students can apply by sending an e-mail expressing your interest to [email protected] or [email protected] , attaching your letter of motivation and possibly your CV.

References: [1] Niemeyer, Günter, et al. "Telerobotics." Springer handbook of robotics (2016); [2] Selvaggio, Mario, et al. "Autonomy in physical human-robot interaction: A brief survey." IEEE RAL (2021); [3] Dragan, Anca D., and Siddhartha S. Srinivasa. "A policy-blending formalism for shared control." IJRR (2013); [4] Javdani, Shervin, et al. "Shared autonomy via hindsight optimization for teleoperation and teaming." IJRR (2018); [5] Hansel, Kay, et al. "Hierarchical Policy Blending as Inference for Reactive Robot Control." IEEE ICRA (2023).

Feeling the Heat: Igniting Matches via Tactile Sensing and Human Demonstrations

In this thesis, we want to investigate the effectiveness of vision-based tactile sensors for solving dynamic tasks (igniting matches). Since the whole task is difficult to simulate, we directly collect real-world data to learn policies from the human demonstrations [2,3]. We believe that this work is an important step towards more advanced tactile skills.

Highly motivated students can apply by sending an e-mail expressing your interest to [email protected] and [email protected] , attaching your letter of motivation and possibly your CV.

• Good knowledge of Python
• Prior experience with real robots and Linux is a plus

References: [1] https://www.youtube.com/watch?v=HH6QD0MgqDQ [2] Learning Compliant Manipulation through Kinesthetic and Tactile Human-Robot Interaction; Klas Kronander and Aude Billard. [3] https://www.youtube.com/watch?v=jAtNvfPrKH8

Inverse Reinforcement Learning for Neuromuscular Control of Humanoids

Within this thesis, the problems of learning from observations and efficient exploration in overactued systems should be addressed. Regarding the former, novel methods incorporating inverse dynamics models into the inverse reinforcement learning problem [1] should be adapted and applied. To address the problem of efficient exploration in overactuted systems, two approaches should be implemented and compared. The first approach uses a handcrafted action space, which disables and modulates actions in different phases of the gait based on biomechanics knowledge [2]. The second approach uses a stateful policy to incorporate an inductive bias into the policy [3]. The thesis will be supervised in conjunction with Guoping Zhao ( [email protected] ) from the locomotion lab.

Highly motivated students can apply by sending an e-mail expressing their interest to Firas Al-Hafez ( [email protected] ), attaching your letter of motivation and possibly your CV. Try to make clear why you would like to work on this topic, and why you would be the perfect candidate for the latter.

Required Qualification : 1. Strong Python programming skills 2. Knowledge in Reinforcement Learning 3. Interest in understanding human locomotion

Desired Qualification : 1. Hands-on experience on robotics-related RL projects 2. Prior experience with different simulators 3. Attendance of the lectures "Statistical Machine Learning", "Computational Engineering and Robotics" and/or "Reinforcement Learning: From Fundamentals to the Deep Approaches"

References: [1] Al-Hafez, F.; Tateo, D.; Arenz, O.; Zhao, G.; Peters, J. (2023). LS-IQ: Implicit Reward Regularization for Inverse Reinforcement Learning, International Conference on Learning Representations (ICLR). [2] Ong CF; Geijtenbeek T.; Hicks JL; Delp SL (2019) Predicting gait adaptations due to ankle plantarflexor muscle weakness and contracture using physics-based musculoskeletal simulations. PLoS Computational Biology [3] Srouji, M.; Zhang, J:;Salakhutdinow, R. (2018) Structured Control Nets for Deep Reinforcement Learning, International Conference on Machine Learning (ICML)

Robotic Tactile Exploratory Procedures for Identifying Object Properties

Goals of the thesis

• Literature review of robotic EPs for identifying object properties [2,3,4]
• Develop and implement robotic EPs for a Digit tactile sensor
• Compare performance of robotic EPs with human EPs

Desired Qualifications

• Interested in working with real robotic systems
• Python programming skills

Literature [1] Lederman and Klatzky, “Haptic perception: a tutorial” [2] Seminara et al., “Active Haptic Perception in Robots: A Review” [3] Chu et al., “Using robotic exploratory procedures to learn the meaning of haptic adjectives” [4] Kerzel et al., “Neuro-Robotic Haptic Object Classification by Active Exploration on a Novel Dataset”

Scaling learned, graph-based assembly policies

• scaling our previous methods to incorporate mobile manipulators or the Kobo bi-manual manipulation platform. The increased workspace of both would allow for handling a wider range of objects
• [2] has shown more powerful, yet, it includes running a MILP for every desired structure. Thus another idea could be to investigate approaches aiming to approximate this solution
• adapting the methods to handle more irregular-shaped objects / investigate curriculum learning

Highly motivated students can apply by sending an e-mail expressing your interest to [email protected] , attaching your letter of motivation and possibly your CV.

• Experience with deep learning libraries (in particular Pytorch) is a plus
• Experience with reinforcement learning / having taken Robot Learning is also a plus

References: [1] Learn2Assemble with Structured Representations and Search for Robotic Architectural Construction; Niklas Funk et al. [2] Graph-based Reinforcement Learning meets Mixed Integer Programs: An application to 3D robot assembly discovery; Niklas Funk et al. [3] Structured agents for physical construction; Victor Bapst et al.

Long-Horizon Manipulation Tasks from Visual Imitation Learning (LHMT-VIL): Algorithm

The proposed architecture can be broken down into the following sub-tasks: 1. Multi-object 6D pose estimation from video: Identify the object 6D poses in each video frame to generate the object trajectories 2. Action segmentation from video: Classify the action being performed in each video frame 3. High-level task representation learning: Learn the sequence of robotic movement primitives with the associated object poses such that the robot completes the demonstrated task 4. Low-level movement primitives: Create a database of low-level robotic movement primitives which can be sequenced to solve the long-horizon task

Desired Qualification: 1. Strong Python programming skills 2. Prior experience in Computer Vision and/or Robotics is preferred

Long-Horizon Manipulation Tasks from Visual Imitation Learning (LHMT-VIL): Dataset

During the project, we will create a large-scale dataset of videos of humans demonstrating industrial assembly sequences. The dataset will contain information of the 6D poses of the objects, the hand and body poses of the human, the action sequences among numerous other features. The dataset will be open-sourced to encourage further research on VIL.

[1] F. Sener, et al. "Assembly101: A Large-Scale Multi-View Video Dataset for Understanding Procedural Activities". CVPR 2022. [2] P. Sharma, et al. "Multiple Interactions Made Easy (MIME) : Large Scale Demonstrations Data for Imitation." CoRL, 2018.

Adaptive Human-Robot Interactions with Human Trust Maximization

• Good knowledge of Python and/or C++;
• Good knowledge in Robotics and Machine Learning;
• Good knowledge of Deep Learning frameworks, e.g, PyTorch;

References: [1] Xu, Anqi, and Gregory Dudek. "Optimo: Online probabilistic trust inference model for asymmetric human-robot collaborations." ACM/IEEE HRI, IEEE, 2015; [2] Kwon, Minae, et al. "When humans aren’t optimal: Robots that collaborate with risk-aware humans." ACM/IEEE HRI, IEEE, 2020; [3] Chen, Min, et al. "Planning with trust for human-robot collaboration." ACM/IEEE HRI, IEEE, 2018; [4] Poole, Ben et al. “On variational bounds of mutual information”. ICML, PMLR, 2019.

Causal inference of human behavior dynamics for physical Human-Robot Interactions

Highly motivated students can apply by sending an e-mail expressing your interest to [email protected] , attaching your a letter of motivation and possibly your CV.

• Good knowledge of Robotics;
• Good knowledge of Deep Learning frameworks, e.g, PyTorch
• Li, Q., Chalvatzaki, G., Peters, J., Wang, Y., Directed Acyclic Graph Neural Network for Human Motion Prediction, 2021 IEEE International Conference on Robotics and Automation (ICRA).
• Löwe, S., Madras, D., Zemel, R. and Welling, M., 2020. Amortized causal discovery: Learning to infer causal graphs from time-series data. arXiv preprint arXiv:2006.10833.
• Yang, W., Paxton, C., Mousavian, A., Chao, Y.W., Cakmak, M. and Fox, D., 2020. Reactive human-to-robot handovers of arbitrary objects. arXiv preprint arXiv:2011.08961.

Incorporating First and Second Order Mental Models for Human-Robot Cooperative Manipulation Under Partial Observability

Scope: Master Thesis Advisor: Dorothea Koert , Joni Pajarinen Added: 2021-06-08 Start: ASAP

The ability to model the beliefs and goals of a partner is an essential part of cooperative tasks. While humans develop theory of mind models for this aim already at a very early age [1] it is still an open question how to implement and make use of such models for cooperative robots [2,3,4]. In particular, in shared workspaces human robot collaboration could potentially profit from the use of such models e.g. if the robot can detect and react to planned human goals or a human's false beliefs during task execution. To make such robots a reality, the goal of this thesis is to investigate the use of first and second order mental models in a cooperative manipulation task under partial observability. Partially observable Markov decision processes (POMDPs) and interactive POMDPs (I-POMDPs) [5] define an optimal solution to the mental modeling task and may provide a solid theoretical basis for modelling. The thesis may also compare related approaches from the literature and setup an experimental design for evaluation with the bi-manual robot platform Kobo.

Highly motivated students can apply by sending an e-mail expressing your interest to [email protected] attaching your CV and transcripts.

References:

• Wimmer, H., & Perner, J. Beliefs about beliefs: Representation and constraining function of wrong beliefs in young children's understanding of deception (1983)
• Sandra Devin and Rachid Alami. An implemented theory of mind to improve human-robot shared plans execution (2016)
• Neil Rabinowitz, Frank Perbet, Francis Song, Chiyuan Zhang, SM Ali Eslami,and Matthew Botvinick. Machine theory of mind (2018)
• Connor Brooks and Daniel Szafir. Building second-order mental models for human-robot interaction. (2019)
• Prashant Doshi, Xia Qu, Adam Goodie, and Diana Young. Modeling recursive reasoning by humans using empirically informed interactive pomdps. (2010)

• DSpace@MIT Home
• MIT Libraries

This collection of MIT Theses in DSpace contains selected theses and dissertations from all MIT departments. Please note that this is NOT a complete collection of MIT theses. To search all MIT theses, use MIT Libraries' catalog .

MIT's DSpace contains more than 58,000 theses completed at MIT dating as far back as the mid 1800's. Theses in this collection have been scanned by the MIT Libraries or submitted in electronic format by thesis authors. Since 2004 all new Masters and Ph.D. theses are scanned and added to this collection after degrees are awarded.

MIT Theses are openly available to all readers. Please share how this access affects or benefits you. Your story matters.

If you have questions about MIT theses in DSpace, [email protected] . See also Access & Availability Questions or About MIT Theses in DSpace .

If you are a recent MIT graduate, your thesis will be added to DSpace within 3-6 months after your graduation date. Please email [email protected] with any questions.

Permissions

MIT Theses may be protected by copyright. Please refer to the MIT Libraries Permissions Policy for permission information. Note that the copyright holder for most MIT theses is identified on the title page of the thesis.

Theses by Department

• Comparative Media Studies
• Computation for Design and Optimization
• Computational and Systems Biology
• Department of Aeronautics and Astronautics
• Department of Architecture
• Department of Biological Engineering
• Department of Biology
• Department of Brain and Cognitive Sciences
• Department of Chemical Engineering
• Department of Chemistry
• Department of Civil and Environmental Engineering
• Department of Earth, Atmospheric, and Planetary Sciences
• Department of Economics
• Department of Electrical Engineering and Computer Sciences
• Department of Humanities
• Department of Linguistics and Philosophy
• Department of Materials Science and Engineering
• Department of Mathematics
• Department of Mechanical Engineering
• Department of Nuclear Science and Engineering
• Department of Ocean Engineering
• Department of Physics
• Department of Political Science
• Department of Urban Studies and Planning
• Engineering Systems Division
• Harvard-MIT Program of Health Sciences and Technology
• Institute for Data, Systems, and Society
• Media Arts & Sciences
• Operations Research Center
• Program in Real Estate Development
• Program in Writing and Humanistic Studies
• Science, Technology & Society
• Science Writing
• Sloan School of Management
• Supply Chain Management
• System Design & Management
• Technology and Policy Program

Collections in this community

Doctoral theses, graduate theses, undergraduate theses, recent submissions.

Inferring Microstructural Change via Non-Contact Ultrasonic Property Measurements 

Sensor Evaluation and Fleet Modeling of Long-Range Low-Cost Autonomous Surface Vehicles 

Exploring Trade-offs and Emergent Properties of Heterogeneous Swarms of Maritime Robot Systems through Empirical Analysis and Application-Driven Experiments 

Show Statistical Information

Autonomous Systems Track

Research/thesis topics - autonomous systems track, research and thesis topics.

The NPS faculty is comprised of accomplished scholars and professionals, predominantly civilian and almost all with doctorates. About 10 percent of the faculty members are senior military officers who, along with students, infuse important operational and combat experience into the education and research programs.

Several Federal agencies and defense organizations, such as NASA, NRO, National Security Agency, as well as defense contractors, sponsor academic chair professorships which further strengthen the institution's relevance.

Below you will find information about some challenging projects within the area of autonomous systems NPS faculty is carrying out. Autonomous Systems Track students will be able to participate in them and contribute with their Master’s theses.

High-Fidelity Modeling and Control Algorithms Engineering for Unmanned Systems Prof. Oleg Yakimenko ( email ) The research opportunities cover a wide range of the issues related to high-fidelity modeling, simulation, testing, model identification, guidance, navigation and control algorithms design as applied to a single and multiple autonomous sensor platforms. These include traditional air, surface and underwater vehicles, missiles, and guided aerodynamic decelerator systems (parachutes and parafoils). The Mathworks’ MATLAB / Simulink development environment is used for computer simulations, hardware-in-the-loop experimentations, real-time code development, and post-experiment data analysis and animations.

Autonomous Localization, Planning, and Control of Mobile Robots Prof. Xiaoping Yun ( email ) This research project investigates design and implementation of localization, mapping, planning, and control algorithms for single and multiple ground-based mobile robots. Multiple sensory modules including infrared sensors, ultrasonic sensors, laser range finders, cameras, and inertial/magnetic sensors are considered for localization and obstacle avoidance. Special graphic simulation software and five Nomadic Robots are available for experimentation.

Robust Adaptive Control for Complex Systems and Flexible Structures Prof. Roberto Cristi ( email ) A variety of research topics in Robust Adaptive Control systems to control systems with uncertain dynamics. A particularly important issue is the control of systems with flexible appendages, which result in non-minimum phase systems with zeros on the imaginary axis. These systems are particularly challenging, especially in Model Reference Adaptive Control, which require stable pole zero cancellation for model matching.

Biologically-Inspired Vehicles Prof. Richard Harkins ( email ) The research emphasis is on biologically inspired mobility for surf-zone vehicles in collaboration with Case Western Reserve and Bristol Universities in Ohio and the UK respectively.

Situational Awareness for Surveillance and Interdiction Operations Prof. Timothy Chung ( email ) This research investigates effective deployment and employment of heterogeneous assets, both manned and unmanned, for conducting search, identification, and interception missions in an area of interest. Both broad area and tactical contexts are considered. Situational awareness of dynamic object locations and identities is represented by a probability model, which integrates imperfect observations and captures nondeterministic object motions. Optimization models guide the allocation of assets to best improve the operational picture by optimal choice of search routes and interception tasking. Incorporation of autonomous elements into these operations can enhance their overall effectiveness via decision support and/or analysis.

UAV Vulnerabilities and Limitations Prof. Robert Bluth ( email ) The research project focuses on different aspects of UAV vulnerability and limitations imposed by a variety of factors.

Self-Organizing Tactical Networking and Collaboration Prof. Alex Bordetsky ( email ) Investigation of various topics related to tactical networking with sensors and unmanned systems (UAS) as well collaboration between geographically distributed units with focus on high value target tracking and surveillance missions. Exploring technologies associated with networking and the human aspects of networked forms of organization, including network-controlled unmanned systems, various forms of multiplatform wireless networking, mesh networked tactical vehicles, deployable operations centers, collaborative technologies, situational awareness systems, multi-agent architectures, and management of sensor-unmanned vehicle-decision maker self-organizing environments. Adaptation of emerging and commercially available technologies to military requirements and investigation of new social networking/collaboration elements associated with the addition of such technologies to the battle space and maritime security operational scenarios.

Collaborative Control of Multiple UAVs Prof. Isaac Kaminer ( email ) Research concentrate on control algorithms design for unmanned air vehicles, modeling and simulation, and flight controls.

Energy Independent Intelligent Autonomy Prof. Vlad Dobrokhodov ( email ) The research concentrates on extending operational endurance of intelligent autonomous systems of all domains including aerial, surface, and underwater robots. The key enabling technology is a cooperative ability of robots to adapt their control strategies to the operational environment and harvest energy in various forms during the mission execution. Adaptation mechanism is built around several new technologies that include the energy-aware control algorithms that drive individual robots, the cooperative algorithms that enable efficient exploration of energy-rich environment, and planning of energy-efficient cooperative missions.

Spacecraft Research and Design Prof. Brij Agrawal ( email ) The research opportunities lie in several challenging areas, such as flexible spacecraft control, acquisition, tracking, and pointing; optical beam control; adaptive optics; beam jitter; adaptive control; control moment gyros control; and space systems design. 

Dynamics and Control of Multiple Unmanned Vehicles Prof. Marcello Romano ( email ) The current research efforts include (real-time) guidance, dynamics, control, and on-board-autonomy for autonomous proximity maneuvers of multiple vehicles.

Development of Small-Scale Autonomous Ground Vehicles Prof. Mike Ross ( email ) The research topics include optimal orbit transfer, attitude control design for agile spacecraft, spacecraft formation flying, motion planning and collision avoidance for unmanned ground vehicles, Mars exploration trajectory design, autonomous cooperative control of robotic manipulators, experimental system identification using motion capture technology, optimal control and adaptive optics. 

Design of Micro (Flapping-Wing) Aerial Vehicles Prof. Kevin Jones ( email ) Research deals with (flapping-wing) aerial vehicles aerodynamic design and systems integration.

Autonomous Underwater Vehicles GN&C Algorithms Design Prof. Doug Horner ( email ) Research is conducted into topics falling in the following broad areas: underwater navigation, control and communication; tactical decision aids; collaborative multi-vehicle operations; obstacle avoidance using forward look sonar; and common AUV mission description language.

Robot Mission Playback using Virtual Environments Prof. Don Brutzman ( email ) The NPS Autonomous Unmanned Vehicle Workbench supports physics-based modeling and visualization of autonomous vehicle behavior and sensors. Our “3 R’s” are Rehearsal, Run-time control and Replay for air, surface and underwater robots. Applying physics-based virtual environments supports control-algorithm development, control-constant testing, mission generation and rehearsal, and replay of completed missions in a benign laboratory environment. Import and export translation of mission plans, scripts and telemetry is accomplished using the Autonomous Vehicle Command Language. Generation of mission reports and X3D visualization using the Savage model library enables better understanding of robot capabilities and real-world results.

Sensor Networks for Surveillance and Detection of Suspicious Behavior Prof. Neil Rowe ( email ) Prof. Gurminder Singh ( email ) This work investigates architectures for sensor networks and the kinds of processing necessary to detect and classify important behaviors. Recent work has focused on infrared, acoustic, and magnetic sensors, plus automated analysis of video when appropriate. This work is related to more general research conducted on the detection and planning of deception operations. The Computer Science Department also does considerable work on data mining, techniques of which are important with analysis of sensor data provided by various autonomous systems.

• Teaching & Research Facilities
• Research/Thesis Topics

Google Custom Search

Wir verwenden Google für unsere Suche. Mit Klick auf „Suche aktivieren“ aktivieren Sie das Suchfeld und akzeptieren die Nutzungsbedingungen.

Hinweise zum Einsatz der Google Suche

• Lehrstuhl für Robotik, Künstliche Intelligenz und Echtzeitsysteme
• TUM School of Computation, Information and Technology
• Technische Universität München

Open Student Thesis Offers

When writing your thesis at TUM I6, please follow our thesis guidelines . A guide to writing good thesis can be found  here . A collection of useful material for research can be found here .

Autonomous Driving, Robotic Grasping, and Dense Prediction

I am looking for highly self-motivated students to work on projects related to autonomous driving, robotic grasping, and dense prediction (classification, detection, and segmentation). The topics include:

• [GR/FP/IDP/SA/BA/MA/etc.] Topic 1: Vision language model for robotic grasp pose estimation (4D and 6D).
• [GR/FP/IDP/SA/BA/MA/etc.] Topic 2: Object detection and semantic segmentation for autonomous driving.
• [GR/FP/IDP/SA/BA/MA/etc.] Topic 3: Multi-modal or multi-sensor fusion for robust object perception.
• [GR/FP/IDP/SA/BA/MA/etc.] Topic 4: Correspondence-free absolute pose estimation algorithm.
• [GR/FP/IDP/SA/BA/MA/etc.] Topic 5: Large language model and vision language model for robotics (autonomous vehicles or robotic grasping).
• [GR/FP/IDP/SA/BA/MA/etc.] Topic 6: Medical image segmentation.

We also have several other open topics in robotics and welcome proposals or ideas from your side. For more information, please contact  Dr. rer. nat. Hu Cao .

Embodied AI

• [SA/BA/MA]:   Robotic Lifelong Reinforcement Learning
• [SA/BA/MA]: Development of GPU Accelerated Robotic Reinforcement Learning Benchmark

AUTOtech.agil - Future of Autonomous Driving and Intelligent Traffic Infrastructure

• [MA/GR/BA] Visual Language Model in Autonomous Driving
• [MA/GR/BA] Domain Adaptation of Synthetic Data for 3D Traffic Environment Perception 
• Real-Time and Robust 3D Object Detection on the Autonomous Driving Test Stretch Using LiDAR Point Cloud Data
• Real-Time and Multi-Modal 3D Object Detection on the Autonomous Driving Test Stretch Using Camera and LiDAR Sensors
• Deep Traffic Scenario Mining, Detection, Classification and Generation on the Autonomous Driving Test Stretch using the CARLA Simulator
• Automated Camera Stabilization and Calibration for Intelligent Transportation Systems

Spiking Neural Networks - Next Generation AI for Autonomous Driving

At the KI-ASIC project we are researching about the application of bio-inspired neural networks to real-world applications.

If you are interested in learning about neuroscience and how neuromorphic engineering is trying to narrow the gap between biology an technology, do not hesitate to contact us.

• [MA|GR] Implementation of a novel Spiking Neural Network on Neuromorphic Hardware

Previous topics:

[BA|MA|GR]  Efficient radar collision detection on neuromorphic hardware

Whisker-Inspired Tactile Sensor

We aim to build a non-intrusive tactile sensor based on the biological structure of rodent's vibrissae, and apply them on the robot arm, biomimetic rodent robot and other platform to provide perception and self-estimate ability.

Available now:

• [MA|SA|BA] Dynamic Contact Estimate along A Whisker-inspired Tactile Sensor.
• [MA|SA|BA] Reconstruction of Contacts in Search Space from A Whisker Sensor Array.

For more information about this topic, please contact Yixuan Dang .

CeCaS: Autonomous driving - Systems and Software Engineering

As part of the research project CeCaS, a group has come up to build a new system architecture for future vehicles with a focus on autonomous driving.

• [MA] Automated Design Space Exploration for Automotive Resource Allocation
• [MA] Exploring in-vehicle TIme-Sensitive Network scheduling based on formal requirements 

Reinforcement Learning, Representation Learning, Meta-RL and Robotics

Three of our DEMOs  can be found at  https://sites.google.com/view/kuka-environment/  ,  https://sites.google.com/view/cemrl , and  https://videoviewsite.wixsite.com/rlsnake .

• Context-based Meta-Reinforcement Learning with Bayesian Nonparametric Models
• Exploiting DATA Symmetries in Context-Based Meta-Reinforcement Learning
• [BA/MA] Meta Reinforcement Learning in Dynamic Environment  
• [BA/MA] Complex Robotic Manipulation via Actionable Representation Learning Guided Exploration 
• [BA/MA] Reinforcement Learning via Hindsight Experience Replay (HER)
• [BA/MA] Reinforcement Learning for Adaptive Locomotion of Snake-like Robot
• [BA/MA] Energy-Efficient Gait Exploration for Snake-like Robots Based on Adversarial Reinforcement Learning
• [BA/MA] Imitation Learning via Demonstration
• [BA/MA] Reinforcement Learning via Hindsight Goal Generation (HGG)
• ​​​​ [BA/MA] Language-conditioned Meta-reinforcement Learning for Multi-manipulation tasks
• [BA/MA] Safe Robotic Manipulation Control with TENG Sensor and Control Barrier function Based on Machine Learning
• [BA/MA] Reinforcement Learning for Medical IoT
• [BA/MA] Offline Reinforcement Learning based on Robotics Scenarios

We have several other open topics in the domain of reinforcement learning in robotics and we also accept open proposals or ideas with yourselves. For more information, please contact Zhenshan Bing .

Neural SLAM and Biomimetic Rodent Robot

• [MA/BA]  Design and Control of a Rat Robot with Actuated Spine and Ribs
• [MA/BA]  Biologically Plausible Spatial Navigation (NeuralSLAM)
• [MA/BA]  Brain-inspired Localization and Mapping based on LiDAR Sensor
• [MA/BA] Pathological gait generation for rat robot with spine-based damage control
• [MA/BA] Rewiring the CPG controller during rat robot error behaviors

For more information about this topic, please contact Zhenshan Bing and Florian Walter .

KI.FABRIK: Future AI & Robotic Factory

 1. Perception and Manipulation of Deformable Objects:

• Motion Planning and Collision Avoidance in Dual-arm Manipulation of Deformable Linear Objects
• 3D Tracking of Deformable Linear Objects with Multi-sensory Integration

 2. Path Planning:

• Path Planning Algorithms via Open Motion Planning Library (OMPL)
• DARKO - Mobile Robot Manipulator Simulation and Trajectory Planning
• Visualization of the mobile robot manipulator for the DARKO project

 3. Obstacle Avoidance:

Algorithms for Multiple Dynamic/Static Obstacle Avoidance

 4. Dynamics Learning:

• Data-Driven Identification and Model Reduction of Dynamic Model

 5. Visual Servoing Control:

• Deep Reinforcement Learning Enhanced Image-based Visual Servoing for Robotic Manipulators

Systems and Software Engineering

Our group provides a range of topics related to systems and software engineering with applications in robotics and automotive.

Currently, the following thesis proposals are open:

• [BA/MA] Containerizing ROS: Efficient Deployment and Management of Robotic Applications
• [BA/MA] Performance study on transporting large-scale dataset.
• [MA] Multi-robot cooperation under signal temporal logic
• [MA] From Natural language to Formal Automotive Architectural Requirements and Vice Versa

Simulation-Based Learning Control for Real-World Robotic Manipulation and Navigation

• [BA/MA/GR/IDP] Comparative Sim2Real Path Planning for AGV Navigation

For more information about other available topics in reinforcement learning, data-driven control and sim2real transfer, please contact Hossein Malmir .

DeepSLAM: Deep Learning based Localization and Mapping (Vision-based Perception and Navigation)

• [MA/BA/GR] DeepSLAM: Deep Learning based Localization and Mapping
• [MA/BA/GR] Simultaneous Localization and Mapping in Dynamic Environments

Autonomous Vehicles

The Cyber Physical Systems group is pursuing a wide range of research directions related to safe decision making, motion planning and control for autonomous vehicles, involving both formal methods, sampling- and optimization-based methods as well as deep learning-based methods.

• [MA] Real-time Motion Planning for Autonomous Driving
• [GR/MA] Learning Model Predictive Robustness of Probabilistic Signal Temporal Logic
• [BA/MA] Encoding the Future: Deep Representations for Traffic using Graph Neural Networks
• [BA/MA] Learning Isometric Embeddings of Road Networks using Multidimensional Scaling
• [MA] Deep Multi-Step Planning for Autonomous Driving
• [MA] Graph Neural Networks for Deep Behavior Prediction in Traffic Scenes
• [BA/SA] Development of a Route Planner for Autonomous Driving
• [MA/SA] Development of a Negotiation Algorithm for Multi-Agent Driving in the Context of Falsification
• [BA/SA/MA] Development of an Algorithm for Attacker Coordination in the Context of Falsification in Autonomous Driving
• [BA/SA] Evaluating Trajectories with Criticality Measures and Robustness Metrics for Autonomous Driving

Safe Reinforcement Learning in Single Robot and Multi-Robot Systems

• [MA/BA]  Safe Massively Multi-Agent Reinforcement Learning
• [MA/BA] Multi-Robot Manipulation and Navigation with Safe Multi-Agent Reinforcement Learning 

We have several topics about Reinforcement Learning, Robotics,  Autonomous Driving and AI Safety, for more information, please contact Shangding Gu .

Safe Reinforcement Learning in Robotics

Currently open positions:

• [MA] Ensuring Human Safety for AI-based Robot Control in ROS 2

Additionally, we will have open topics in safe reinforcement learning for manipulators and mobile platforms in the future. If you are interested in these topics, you can contact: jakob (dot) thumm (at) tum.de

Modular Robotics

For interest in a BA/MA thesis in machine learning for modular robotics, please contact Jonathan Kuelz or Matthias Mayer .

Currently open:

• [MA/SA/BA] Efficient Path Planning for Modular Robots
• [MA/SA/BA] Solving Real-world Robotics Tasks within 3D Scans

Robust and Nonlinear Motion Planning & Control

• [MA] Safe and Efficient Inspection of Power Lines using UAVs
• [MA] Robust Control of Linear Systems with Uncertain Parameters
• [MA] Safety Certification for Learning-Based Control

Formal Methods and Reachability Analysis

• [BA or MA]  Data-driven Identification of Uncertainty Sets for Autonomous Systems
• [BA or MA] Uniform Trajectory Planning for Cyber-Physical Systems
• [BA or MA] Optimization-based Verification of Cyber-Physical Systems
• [MA] Ensuring Safety of Large-Scale Structures
• [MA] Exploiting Mixed-Monotonicity in Reachability Analysis
• [MA] Errors of Trajectories for Autonomous Vehicles and Cyber-Physical Systems

Offline Reinforcement Learning

[Research internship] Mastering the game of Skat using decision transformers

Safe Reinforcement Learning, Multi-Agent Reinforcement Learning

• [BA/MA]: Provably Safe Reinforcement Learning Control of a Quadrotor

Reinforcement Learning for Safe and Efficient Combustion Engine Control

Neurorobotics in the human brain project.

• Developmental Body Modeling in Soft Robotics
• Cloud-Based Robotics for Machine Learning
• Virtual Neurorobotics with Intel Loihi
• Spiking Compliant Robot Control with Intel Loihi
• Integration of the Neural Simulator NEST into the Neurorobotics Platform
• Deep Spiking Q-Networks
• Autonomous Locomotion Control for Snake Robot Based on Bio inspired Vision Sensor and Spiking Neural Network
• Advanced Autonomous Driving Control Based on Bio inspired Vision Sensor and Spiking Neural Network
• Spiking Neural Network for Autonomous Navigation based on LiDAR Sensor
• Deep Spiking Reinforcement Learning
• Learning adaptive target reaching with Recurrent Neural Networks
• Biologically-inspired Perception for Autonomous Vehicles based on LiDAR Sensor

SPEED-CARGO Project

• Knowledge Base and Inference Models for AI-based Optimiza-tion and Control of Air Cargo ProcessesMathematical Programming Models Processes Availability
• Deep learning applied to Real World Robotic System
• Mathematical Programming Models and Optimization for Air Cargo Processes

Machine Learning Algorithms for Hybrid Vehicle Data

Please see this page for the available topics about 3D Object Detection and Tracking.

Autonomous Robot & Visual Servo & Deep Learning & Robot Design & Medical Robotics

For more information, please visit my homepage Mingchuan Zhou or contact me via email ([email protected]).

OSBORNE (Future Automotive E/E Architectures for Autonomous Cars)

We have a set of open topics in the domain of affective computing and multimodal emotion recognition, within the context of OSBORNE project, for more information please contact Sina .

Collaboration with Chair for Product Development and Lightweight Design

• Design and Control of a Humanoid Robot Arm

External thesis proposals

• Masterthesis – Grey-Box Modellierung und Validierung des Motoraufheizverhaltens auf dem Rollenprüfstand und im Fahrzeug mittels Machine Learning
• [Master Thesis or Working Student] LLM for robotics with multi-fingered hand (Agile Robots AG)

Low-level vision

For more information, please visit my homepage ( Yuning Cui ) or contact me via email ([email protected]).

Master of Science (MSc) in Autonomous Systems

Specializations.

There are no specializations within this programme.

• With a bachelor from Denmark
• With a bachelor from outside of Denmark

Autonomous systems are entering our working, urban, and domestic environments. Imagine fleets of smart robots that collaborate in manufacturing facilities, advanced warehouse logistics solutions such as the Amazon warehouses, crew-less cargo ships, smart grids, or interconnected smart home appliances. Those are just a few examples of what we call autonomous systems.

Oops, this feature requires cookies

To view the content, you must change your cookie consent to allow functionality and targeting cookies

Study Programme

The MSc in Autonomous Systems equips you to participate in and lead work related to the design and deployment of autonomous systems, exploiting the latest advances in artificial intelligence, big data analysis, digitization, sensing, optimization, information technologies, and systems engineering.

The MSc programme in Autonomous Systems has been developed to build both mathematical and theoretical understanding, as well as practical experience of students, applicable to Industry 4.0, robotics, and autonomous software agents, among other application areas.

Focus is on systematic problem solving and synthesis, and you will acquire specialized knowledge in the area. More specifically, the study programme will provide you with sound hands-on experience through laboratory work and the application of IT-based tools.

Presentation of the MSc programme in Autonomous Systems

You should be aware that your choice of courses in your individual study plan offers a high degree of flexibility. Therefore, you have every opportunity to design your own study programme and career by choosing from the wide range of courses offered at DTU. You are responsible for ensuring consistency when designing your study plan.

Read more about the structure of the MSc programmes at DTU.

Combine work with studies

This study programme is also available as an Industry study programme where you can combine work and study over a 4-year period. Special rules apply.

Industry MSc Eng programmes

With an MSc Eng in Autonomous Systems, you can pursue a career in various production and service industries in both private and public organizations. For instance, you can have a career as an AI developer, production manager, systems engineer, IT consultant - to mention but a few. Read more about your career options with a master's degree in Autonomous Systems

More opportunities at DTU

Honours Programmes

Get access to a particularly challenging course of studies with DTU's Honours Programmes.

Industry Master of Science in Engineering

Join an Industry Master of Science in Engineering - a part-time study which can be combined with a part-time job.

Study abroad

As a DTU student, you have many opportunities to study abroad as part of your degree - everything from a semester abroad to an international joint degree or a summer school exchange.

Learn more about your opportunities to go abroad

Make a sustainable difference

With a degree from DTU you are prepared to make a sustainable difference. DTU Charter helps to ensure that.

Student life at DTU

Get an interactive tour around campus

Enter our interactive online universe at DTU and explore different locations on Campus.

Study relevant projects

At DTU you have the chance to experiment and develop innovative solutions to problems or challenges - in collaboration with students from all academic fields of DTU.

Sports and social events

DTU has a large selection of indoor and outdoor sports facilities. You are also invited to attend the Friday bars and the many other social events and festive occasions at DTU.

Get your master's at DTU

An international university

Lying in one of the world’s happiest and safest countries, DTU welcomes graduate students from all over the world to be part of an international, innovative and open-minded learning environment.

Any questions?

• Study Guidance

Get help with questions regarding the study programmes - including course planning, credit transfer applications, exemption applications, and DTU’s rules and regulations.

• International Admissions Office

For applicants with a bachelors degree from outside of Denmark. Get guidance on the application process.

• International Student Services

Offers guidance to DTU students on exchange options and as well as administrative and cultural issues relating to being new in Denmark and new at DTU.

How to apply

• Apply with a Danish bachelor degree
• Apply with a non-Danish bachelor degree

Similar programmes

• Computer Science and Engineering
• Human-Centered Artificial Intelligence
• Prerequisites
• Student life

Digital Commons @ Kettering University

• < Previous

Home > THESIS > GRADUATE_THESES > 1

Graduate Theses

Layout option.

Master of Science in Engineering

Autonomous Vehicle Lateral Controller Analysis and Comparison for Multiple Controller Strategies

Alexander L. Garrow , Kettering University

Date of Award

Document type, degree name.

Engineering

First Advisor

Dr. Diane Peters

Second Advisor

Dr. Ahmed Mekky

Third Advisor

Dr. Jennifer Bastiaan

In an autonomous vehicle there needs to be a robust control strategy capable of calculating the optimal inputs to the vehicle. Autonomous systems control the vehicle through the turning of the steering wheel, application of the brakes, and generating torque at the powertrain. Lateral and longitudinal vehicle motion are the two main control considerations for autonomous vehicles with the former focusing on the steering wheel and the latter on the brakes and throttle. The Kettering University AutoDrive team is tasked with creating these control strategies and applying them to a Chevrolet Bolt. The focus of this thesis is on the creation and testing of various lateral control strategies in simulation. The simulation environment and controls are both implemented in Simulink and MATLAB software. The performance of each controller tested is compared and analyzed in terms of how well they can follow a path as well as how much lateral acceleration the controller creates. Through the process of completing this thesis a novel form of lateral control was created, named Garrow Control, and the optimal controller is selected for the purposes of use in the fourth year of the SAE AutoDrive competition. The novel contributions of this thesis are the comparison of various lateral control strategies in the application of the AutoDrive competition and the creation of a new form of lateral control.

Recommended Citation

Garrow, Alexander L., "Autonomous Vehicle Lateral Controller Analysis and Comparison for Multiple Controller Strategies" (2021). Graduate Theses . 1. https://digitalcommons.kettering.edu/graduate_theses/1

Since March 02, 2022

• Collections
• Disciplines

Advanced Search

• Notify me via email or RSS

Author Corner

• Sign the Author Agreement

Library Resources

• About Digital Commons @ KU
• About Faculty Works @ KU
• Open Access
• Research Data Management

Home | About | FAQ | My Account | Accessibility Statement

Privacy Copyright

• Svenska Close Den här sidan är ej översatt Startsida på svenska

Planning and Motion Control in Autonomous Heavy-Duty Vehicles

Autonomous driving has been an important topic of research in recent years. In this thesis we study its application to Heavy Duty Vehicles, more specifically, vehicles consisting of a truck and a trailer. An overview study is done on three fundamental steps of an autonomous driving system, planning, trajectory tracking and obstacle avoidance.

In the planning part, we use RRT, and two other variants of the algorithm to find trajectories in an unstructured environment, e.g., a mining site. A novel path optimization post-processing technique well suited for use with RRT solutions was also developed. For the trajectory tracking task several well-known controllers were tested, and their performance compared. An extension is proposed to one of the controllers in order to take into account the trailer. The performance evaluation was done on scaled truck systems in the Smart Mobility Lab at KTH. The obstacle avoidance is done with the aid of a simple, yet functional Model Predictive Controller. For this purpose, we developed di erent formulations of the optimization problem, corresponding to distinct optimization goals and vehicle models, in order to assess both the quality of the MPC, and of the assumed truck model. The outcome of this thesis is a fully autonomous system, able to plan and move in constrained environments, while avoiding unpredicted obstacles. It was implemented using a 1:32 scale remote controlled truck, commanded by a desktop computer.

Rui Oliveira

Rui Oliveira's Master's Thesis (pdf 3.7 MB)

• Press Enter to activate screen reader mode.

Department of Earth Sciences

Master in earth sciences.

The Master's degree study programme in Earth Sciences at ETH Zurich is characterised by practical training and a comprehensive science education.

The Master's degree study programme in Earth Sciences provides multi-disciplinary approach: field work, analytical methods, lab experiments, numerical simulation, e-learning, case study analysis, and teamwork.

Four different majors

By choosing one of the offered majors the students define the main area of their educational path:

• chevron_right Major in Geology
• chevron_right Major in Engineering Geology
• chevron_right Major in Mineralogy and Geochemistry
• chevron_right Major in Geophysics

Students will receive personalised advice in order to plan their studies according to their interests and career goals.

Career opportunities

Earth Science graduates work in a wide variety of jobs, all over the world in sectors like environmental consulting firms, mining and exploration companies, oil companies, research institutes, government, and educational institutions.

Structure of the programme

Courses are organised into topical blocks (modules) of 12 ECTS. Each major has four compulsory modules and numerous elective courses from the complete offerings of ETH Zurich and of the University of Zurich.

Master's thesis in Earth Sciences

The Master's programme in Earth Sciences is concluded with the Master's thesis. The subject of the thesis is in the major study area and represents either an applied or fundamental research project.

Application for Master's studies

The application for Master's degree programmes with start in Autumn Semester 2024 is closed.

Download Study Guide: Master in Earth Sciences (PDF, 603 KB) vertical_align_bottom

Study Coordination

• Location location_on NO D 55
• Phone phone +41 44 632 64 83

Dep. Erdwissenschaften Sonneggstrasse 5 8092 Zürich Switzerland

Science in Perspective (Wissenschaft im Kontext)

The study programme “Science in Perspective (Wissenschaft im Kontext)” enables students at ETH Zurich to develop new perspectives on the contents of their core subjects.

Google: Almost The Cheapest Magnificent 7 Stock, Double-Digit Growth Expected

• Despite the previous market concerns about the longevity about its Google Search segment attributed to the rise of ChatGPT, the segment remains GOOG's top/ bottom-line driver in FQ2'24.
• The management has demonstrated their ability to push Google Cloud into sustained profitability, despite our original skepticism surrounding the new accounting methodology.
• GOOG is also nearing its projected $100B revenue run-rate from both YouTube Ads and Google Cloud by end 2024, as observed from the cumulative FQ2'24 revenue of $19B.
• Its in-house AI capabilities, vertically integrated platform building expertise, and innovation across its existing/new offerings have led to its promising results indeed.
• Combined with its market-leading streaming share and monetization of autonomous driving through Waymo, it is apparent that the market is sleeping on GOOG's well-diversified capabilities.

Jonathan Kitchen

We previously covered Alphabet Inc. aka Google ( NASDAQ: GOOG ) in May 2024, discussing its promising prospects in the search engine/ generative AI capabilities, along with robust profitability and rich balance sheet in FQ1'24.

Combined with its ability to consistently deliver profitable growth and robust shareholder returns, we had upgraded our Buy rating to Strong Buy then, with the stock still fairly valued while offering an excellent upside potential.

Since then, GOOG has further rallied by +16.1%, well outperforming the wider market at +8.4%, before the recent rotation away from high growth stock over the past two weeks.

Even so, we are maintaining our optimism surrounding its long-term prospects, thanks to its double beat FQ2'24 earnings call and accelerating Cloud performance metrics.

Combined with relatively cheap P/E valuations compared to its Magnificent 7/streaming/autonomous driving peers, it goes without saying that the stock remains a long-term winner for growth oriented investors.

We shall discuss further.

GOOG's Integrated AI Investment Thesis Has Paid Off Handsomely

GOOG 3Y Stock Price

TradingView

GOOG has had a volatile three years indeed, as observed in the stock price chart above, attributed to the pre-pandemic boom, subsequent correction after November 2021, and the eventual recovery after the January 2023 bottom.

Despite the previous market concerns about the longevity about its Google Search segment, as discussed in our GOOG article here in January 2023, it is apparent that those fears have been unfounded.

This is why.

GOOG continues to command the largest global search engine market share at 91.06% as of June 2024 (+0.26 points MoM/ -1.58 YoY/ -1.65 points from December 2019 levels of 92.71% ), despite the minimal share losses from November 2022 levels of 92.21% when ChatGPT was launched.

While part of the losses are attributed to Microsoft ( MSFT ) Bing's gains, we are not overly concerned indeed, since Google Search continues to generate impressive revenues of $48.5B in FQ2'24 ( +5% QoQ / +13.7% YoY/ +105.1% from FQ2'19 levels of $23.64B ).

At the same time, GOOG is also nearing its projected $100B revenue run-rate from both YouTube Ads and Google Cloud by end 2024, as observed from the cumulative FQ2'24 revenue of $19B (+7.5% QoQ/ +21% YoY/ +233.5% from FQ2'19 levels of $5.7B).

Most importantly, the management has demonstrated their ability to push Google Cloud into sustained profitability, despite our original skepticism surrounding the new accounting methodology , attributed to the extended "useful life of its servers from four to six years and network equipment from five to six years."

This has led to Google Cloud's expanding operating margins of 11.3% in FQ2'24 (+1.9 points QoQ/ +6.4 YoY), further underscoring why their ongoing innovation across every layer of the AI stack has paid off handsomely.

This is also why we believe that as an AI-first company, GOOG has demonstrated robust monetization efforts across its search/ generative AI/ cloud offerings, thanks to its multi-modal integrated AI capabilities, which in turn enhances the user experience and increases search engine loyalty.

We must remind readers that GOOG boasts an in-house AI team compared to Microsoft's partnership with OpenAI, which enables the former to vertically integrate the platform building expertise while innovating across its existing/ new offerings.

This further underscores why the management has been able to rapidly deploy/ monetize Gemini across its developer tools and data center to edge use cases.

Combined with the sustained rationalization of its offerings and headcounts , it is unsurprising that we have seen GOOG's robust top-line growth flow into its increasingly richer adj operating margins of 32.3% (+0.7 points QoQ/ +3.1 YoY/ +8.8 from FQ2'19 levels of 23.5%) in FQ2'24.

GOOG Valuations

Tikr Terminal

With consensus forward estimates remaining stable over the past three months and GOOG still expected to report an accelerated top/ bottom-line expansion at a CAGR of +11.6%/ +19.4% through FY2026, it is apparent that the stock remains cheap here.

This is especially since GOOG at FWD P/E valuations of 21.29x is still trading well below its Magnificent 7 peers (aside from META at 22.06x) and its 5Y P/E mean of 25x, as observed in the chart above.

Even compared to their bottom-line growth rates, such as Meta at 22.06x at +20.9% through FY2026, AAPL at 31.46x/ +10.3%, MSFT at 33.88x/ +16.5%, AMZN at 37.91x/ +37.5%, NVDA at 38.93x/ +49.6%, and TSLA at 72.15x/ +13.3%, respectively, it is apparent that GOOG is on the cheap side compared to the other Mag 7 peers.

Given YouTube's leading streaming share of 8.4% in June 2024 ( +0.8 points MoM / +0.2 YoY ) in a growing overall market of 40.3% (+1.5 points MoM/ +2.6 YoY), it is apparent that the secular transition from TV Media to streaming is still ongoing.

Even when compared to its streaming peer, Netflix ( NFLX ) at FWD P/E valuations of 33.57x with the projected adj EPS growth at +31.1% through FY2026, it is apparent that GOOG remains highly attractive at current levels.

Lastly, readers must not forget GOOG's autonomous driving capability, Waymo , which has started to deliver over 50K weekly paid public rides in San Francisco and Phoenix, along with its ongoing fully autonomous testing (without driver).

While TSLA has touted robo-taxi capabilities with an event now planned in October 10, 2024, it remains to be seen when monetization may actually occur, with Waymo currently leading the race.

As a result of these developments, we maintain our belief that the market is sleeping on GOOG's well-diversified capabilities, with it offering interested investors with an excellent margin of safety despite the massive rally since the January 2023 bottom.

So, Is GOOG Stock A Buy , Sell, or Hold?

For now, GOOG has charted a new peak of $190s and is trading above its 100/ 200 day moving averages, thanks to the robust FQ2'24 earnings results and the market's conviction surrounding its AI monetization.

Despite the recent correction, the stock continues to trade not too far from our previous article levels of $170s, with current levels still near to our updated fair value estimates of $174.20, based on the LTM adj EPS of $6.97 (+6.9% from the previous levels of $6.52) and the 5Y P/E valuations of 25x.

At the same time, there remains an excellent upside potential of +40.6% to our reiterated long-term price target of $258.30, based on the flattish consensus FY2026 adj EPS estimates.

In addition, GOOG has executed well on its long-term shareholder returns, based on the 269M shares, or the equivalent 2.1% of its float retired over the LTM, and 1.35B/ 9.7% since FY2019, with approximately $54B still remaining in its repurchase authorization.

While minimal, the annualized $0.80 dividend per share allows long-term shareholders to DRIP and accumulate additional shares on a quarterly basis.

As a result of its robust AI monetization prospects and dual pronged prospective returns through capital appreciation/ dividend payouts, we are maintaining our Buy rating for the GOOG stock here.

Do not sleep on this giant.

This article was written by

Analyst’s Disclosure: I/we have a beneficial long position in the shares of GOOG, TSLA, META, MSFT, NVDA, NFLX, AMZN, AAPL either through stock ownership, options, or other derivatives. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article. The analysis is provided exclusively for informational purposes and should not be considered professional investment advice. Before investing, please conduct personal in-depth research and utmost due diligence, as there are many risks associated with the trade, including capital loss.

Seeking Alpha's Disclosure: Past performance is no guarantee of future results. No recommendation or advice is being given as to whether any investment is suitable for a particular investor. Any views or opinions expressed above may not reflect those of Seeking Alpha as a whole. Seeking Alpha is not a licensed securities dealer, broker or US investment adviser or investment bank. Our analysts are third party authors that include both professional investors and individual investors who may not be licensed or certified by any institute or regulatory body.

Recommended For You

About goog stock.

More on GOOG

Related stocks.

Trending Analysis

Trending news.