MOTION PLANNING FOR MULTIPLE SYSTEMS UNDER COORDINATED CONSTRAINTS

摘要

Two problems of motion planning for controlled systems which are required to attain a given target set under coordinated constraints are formulated and solved using dynamic optimization techniques. Constraint coordination arises when the state of each system is mapped onto state constraints for other systems. These problems are formulated in terms of backward reach sets which are the sub-zero level sets of appropriate value functions for non-standard cost functions. The value functions are the solutions of Hamilton-Jacobi-Bellman type PDEs. For linear dynamics and ellipsoidal constraints the value functions are calculated through duality techniques from convex analysis.