Abstract by Ashton Larkin
Dynamically Feasible, Energy-Efficient Planning With RRT*
In the area of autonomous robotics, there are many algorithms that find a path in an unkown environment that the robot can traverse. Finding the shortest path in the environment often requires robot configurations that violate its dynamic constraints, making shortest paths impractical in many motion planning scenarios. We present a way to model the dynamics of a differential drive robot that allows us to define an energy cost function. We will show how this cost function is used as the heuristic in the RRT* (Rapidly-Exploring Random Tree) algorithm and discuss the complexities that arise from using this heuristic.