TUM · Robotics

Robot Motion Planning

Course Syllabus

Summer Semester 2024

About the course

Many confuse the recordings with actual lectures per week. The online courses are from previous years and have the advantage that you can stop the lecture at any time and restart it at any slide later by choosing it from the slide listing displayed on the left of the recording.

I will repeat the outline of the course here, so you can check our progress throughout the semester.

I give you some material ahead of time. You should finish processing the corresponding videos when the corresponding homework set gets presented.

Syllabus

Section Topic Video Homework
Representation
How to represent the robot to simplify the planning task — configuration space. Ways to reduce the robot to a point. Video 1 HmSet 1
Direct Planning Methods
Planning of the complete path online — Bug algorithms Video 2 HmSet 2
Planning of the complete path online — Wavefront planner Video 2 HmSet 2
Planning Strategies
Complete Map Knowledge Offline construction of roadmaps — Visibility Graph Video 2 HmSet 1
Complete Map Knowledge Offline construction of roadmaps — Voronoi Graph Video 2 HmSet 1
Complete Map Knowledge Offline construction of roadmaps — Trapezoidal Cell Decomposition and Boustrophedon Video 2 HmSet 2, 3
Partial Map Knowledge Heuristic method based on Potential Field Video 3 HmSet 2, 3
No Map Knowledge Sampling-based method — Multiple Query Probabilistic Roadmap Video 4 HmSet 3
No Map Knowledge Sampling-based method — Single Query Probabilistic Roadmap Video 5 HmSet 3
Analysis of the expansiveness (connectivity) of a graph Video 6 HmSet 3
Parametrization of the PRM (number of nodes etc.) based on expansiveness Video 6 HmSet 3
No Map Knowledge Improvements of sampling-based methods — Obstacle-Based PRM, RRT Video 7
Expansiveness of the Space (PRM parametrization)
Methods to improve connectivity of probabilistic roadmaps Video 7
Fusion of Uncertain Data for Mapping
Gaussian · Linear Fusion with Kalman Filter
Gaussian · Non-Linear Fusion with Extended Kalman Filter
Gaussian · Highly Non-Linear Fusion with Unscented Kalman Filter
Arbitrary Uncertainties Fusion with Bayesian and Particle Filter
Simultaneous Localization and Mapping (SLAM)

🎥 How to use the recordings

The recordings are from previous years and are intended as supporting material. You can pause the lecture at any time and jump to individual slides using the slide listing.

📝 Homework workflow

Material is provided ahead of time. Students should complete the corresponding videos when the relevant homework set is presented.