Probabilistic car-like robot path planning from temporal logic specifications

Abstract

An algorithmic framework is developed for automatic deployment of car-like robots based on Linear Temporal Logic (LTL) formulae over a set of regions of interest in the environment. The environment and the regions of interest are a priori known, and the robot has non-negligible size and restricted steering capabilities. The approach relies on constructing a probabilistic finite-state abstraction of the car-like robot and on finding a trajectory (run) in this abstraction such that the probability of satisfying the LTL formula is maximized. The feasibility of our approach is supported by simulations under Matlab environment.