Iterative temporal motion planning for hybrid systems in partially unknown environments

Abstract

This paper considers the problem of motion planning for a\nhybrid robotic system with complex and nonlinear dynamics\nin a partially unknown environment given a temporal logic\nspecification. We employ a multi-layered synergistic framework\nthat can deal with general robot dynamics and combine\nit with an iterative planning strategy. Our work allows us\nto deal with the unknown environmental restrictions only\nwhen they are discovered and without the need to repeat\nthe computation that is related to the temporal logic specification.\nIn addition, we define a metric for satisfaction of\na specification. We use this metric to plan a trajectory that\nsatisfies the specification as closely as possible in cases in\nwhich the discovered constraint in the environment renders\nthe specification unsatisfiable. We demonstrate the efficacy\nof our framework on a simulation of a hybrid second-order\ncar-like robot moving in an office environment with unknown\nobstacles. The results show that our framework is successful\nin generating a trajectory whose satisfaction measure of the\nspecification is optimal. They also show that, when new obstacles\nare discovered, the reinitialization of our framework\nis computationally inexpensive.