news
DCL Research Spotlight: Mark L. Mote, Andrew Bylard, Marco Pavone, Magnus Egerstedt, Eric Feron
Primary tabs
Mark L. Mote, Andrew Bylard, Marco Pavone, Magnus Egerstedt, Eric Feron
Abstract:
Consideration of collisional behavior in vehicle trajectories can enable richer set of solutions, allowing both better performance, and a safest course of action in scenarios where collision is inevitable. Here we develop an approach for collision tolerant optimal trajectory planning on a three degree of freedom free-flier spacecraft. First, we extend previous results for generating optimal trajectories in particle motion to a full hybrid model for the spacecraft. Next, a collision model is derived from experimental data, and incorporated along with the nominal dynamics as constraints in a mixed integer quadratic program (MIQP). Simulated scenarios demonstrate improved robustness of the MPC online, and experimental comparisons show the significant reductions in fuel consumption for trajectories that incorporate collisions.
Status
- Workflow Status:Published
- Created By:mamstutz3
- Created:10/16/2018
- Modified By:mamstutz3
- Modified:10/18/2018
Categories
Keywords