Integrating explicit path planning with reactive control of mobile robots using Trulla

摘要: Abstract This article describes and investigates a method of interleaving explicit path planning with reactive control. The Trulla all-paths planner computes an priori set optimal paths. Minor reactions to obstacles terrain changes serve switch the robot from precomputed new path, eliminating subgoal obsession. Major deviations suggest that map is significantly wrong; replanning should be triggered ensure continued progress robot. dot product used as intrinsic measure major deviation. methodology particularly well-suited for computationally bound robots such planetary rovers operating in indoor environments large number minor unmodeled obstacles. algorithms, reports on experimental data collected mobile under representative scenarios. compared continuous fixed frequency differing velocities. results show deferred Trulla/dot produced actual paths similar more frequent distance time but up 100 times less computation. reduced computation led 8.75% increase traveled 24% travel time. In presence faulty sensor data, outperformed other methods which radically changed back forth due spurious readings.