Motion Planning and Decision Making in Robotics

Garant předmětu:

Jan Faigl

Lecturer:

Jan Faigl

Tutor:

Jan Faigl

Supervisor:

Department of Computer Science

Synopsis:

The course aims to acquaint students with the use of planning approaches and decision-making techniques for autonomous robotic systems. Students in the course will use the knowledge of motion planning algorithms with the focus on randomized sampling-based algorithms and their theoretical properties. The further topic deals with employing the planning methods in the data collection missions such as environment monitoring, coverage, surveillance and inspections tasks where the motion planning is embedded into variants of the combinatorial routing problems as robotic variants of the traveling salesman problem, which is called the sequencing tasks in robotics. The next part of the course is dedicated to a discussion of robotic information gathering problems where on-line decision making has to be made during the mission. Informative adaptive path planning algorithms and existing solutions will be discussed. The last part of the course is dedicated to planning with uncertainty and sequential decision-making under uncertainty. The course will be composed of lectures providing an introduction to the topics and reading groups where students will be assigned state-of-the-art papers and present them to the group.

Requirements:

Syllabus of lectures:

1. Robotic paradigm and control architectures

2. Path and motion planning - notation, terminology, and Grid and Graph-based path planning methods

3. Reading group: Lifelong planning approaches and real-time heuristic search - D* lite, LRTA*, ARA*

4. Robotic information gathering - exploration of unknown environment, information-theoretic approach

5. Multi-goal Motion Planning Problem and Sequencing Tasks - Robotic variants of the traveling salesman problem

6. Data collection planning - TSP(N), PC-TSP(N), and OP(N) and an extension for curvature constrained vehicles

7. Reading group: on-line decision making and adaptive informative path planning

8. Randomized sampling based-motion planning methods

9. Improved asymptotically optimal motion planning algorithms

10. Reading group: Motion planning with uncertainty

11. Sequential decision-making under uncertainty - Markov Decision Processes

12. Partially Observable Markov Decision Processes (POMDPs)

13. Reading group: Designing control policies for robotic systems

Syllabus of tutorials:

Study Objective:

Study materials:

Steven M. LaValle: Planning Algorithms, 2006.

Mykel J. Kochenderfer: Decision Making Under Uncertainty, 2015.

Note:

Further information:

https://cw.fel.cvut.cz/wiki/courses/xep36uir/start

Time-table for winter semester 2024/2025:

Time-table is not available yet

Time-table for summer semester 2024/2025:

Time-table is not available yet

The course is a part of the following study plans: