Teaching
Courses I teach and opportunities for prospective students.
Courses
I design and teach courses on robotic manipulation, mobile robotics, and learning for control.
| Year | Course | Institution | Role |
|---|---|---|---|
| 2026 | Introduction to Robotic Manipulation | University of Turku | Course designer and lecturer: wrote the syllabus and held 7 lectures. |
| 2025 | Perception and Navigation in Mobile Robotics | University of Turku | Course redesigner and lecturer: rebuilt the course, held the lectures and developed the exercises. |
| 2023 | Reinforcement Learning | KTH Royal Institute of Technology | Teaching assistant. |
| 2019, 2021–2022 | Robotic Manipulation | Aalto University | Primary teaching assistant: co-developed the curriculum and exercises. |
I also started and lead the reform of the University of Turku’s Robotics and Autonomous Systems MSc programme (2026–2027). I proposed seven new courses, and the department’s education development group endorsed the reform unanimously.
Student Supervision
I have co-supervised 3 PhD students at KTH, two of whom graduated in 2025, and was principal supervisor of 11 MSc theses at Aalto University, KTH and the Czech Technical University. Their projects covered how robots choose where to grasp an object (grasp synthesis), how they handle cloth and other soft objects, and generative models for robotics.
Prospective Students
I am currently not looking for PhD students. I do supervise Master’s thesis projects in five areas. The example papers show what such a project can look like.
- Grasp generation for robot hands, from two-finger grippers to multi-fingered (dexterous) hands, with the grasp’s full 3D position and orientation (6-DoF). Examples: generating multi-finger grasps with diffusion models and grasping several objects in one hand.
- Manipulating cluttered scenes and deformable objects such as cloth. Examples: multi-finger grasping in clutter from a single depth image, tracking the 3D shape of cloth from camera images and identifying a garment and its full shape when folds hide parts of it.
- Generative models and planning under uncertainty, for when the robot cannot see everything it needs to know. Example: planning manipulation when the robot cannot tell where one object ends and the next begins.
- Perception that combines vision and touch (visuo-haptic perception). Example: choosing where to touch an object to complete its 3D shape.
- Vision-language-action (VLA) models, which turn camera images and written instructions into robot actions, for robots with different grippers and hands.
If one of these fits you, email me at jens.lundell@utu.fi with: