Multi-Modal Motion Planning: A Hybrid Automaton Approach
Informed RRT* path - 1
Informed RRT* path - 2
Hybrid Automaton for Autonomous Driving with 4 states
Informed RRT* path on a map in Rviz and Gazebo
Description
We present a novel approach for Multi-Modal Motion Planning using a Hybrid Automaton. We present a preliminary framework here, with the switching conditions chosen based on cluttering in the environment. Cluttering is classified based on the points within a predefined distance threshold. We then present the simulation environment and evaluation metrics to select various planners as states for the Hybrid Automaton
My Role
• Developed a Hybrid Automaton framework for autonomous vehicles based on sensor data
• Implemented RRT* & Informed RRT* with a C++ node for a high-level plan
• Developed guard and reset conditions for the Hybrid Automaton