Robert Pěnička presents Advancing Multi-Goal Motion Planning with Kinodynamic Rapidly-Exploring Random Forest

How can we efficiently compute collision-free, kinodynamically feasible
trajectories for robotic systems operating in cluttered environments? Join us
for a short seminar introducing the Kinodynamic Rapidly-Exploring Random Forest
(KRRF) algorithm — a novel approach to solving the multi-goal motion planning
problem, which combines motion planning for kinodynamic systems with the
Traveling Salesman Problem (TSP) solution. KRRF grows multiple planning trees
simultaneously from target locations, using a heuristic-driven expansion
strategy to efficiently connect waypoints while adhering to robot dynamics.
Unlike conventional methods, KRRF avoids solving costly two-point boundary value
problems, making it applicable to a wide range of robotic platforms with
arbitrary motion models. This seminar will present simulation-based results,
demonstrating KRRF’s performance across different motion models, including
car-like and differential-drive robots, in cluttered environments. Join us as we
explore how KRRF is shaping the future of kinodynamic multi-goal motion
planning!

Reference:
P. Ježek, M. Minařík, V. Vonásek and R. Pěnička, "KRRF: Kinodynamic
Rapidly-Exploring Random Forest Algorithm for Multi-Goal Motion Planning," in
IEEE Robotics and Automation Letters, vol. 9, no. 12, pp. 10724-10731, Dec.
2024, doi: 10.1109/LRA.2024.3478570.
Paper: https://doi.org/10.1109/LRA.2024.3478570
Video: https://youtu.be/KLneA8Mkep4
Code: https://github.com/ctu-mrs/krrf