Clutter-Resilient Autonomous Mobile Robot Navigation with Computationally Efficient Free-Space Features.

This paper proposes free-space motion tubes, a motion primitive for the local navigation of mobile robots equipped with range sensors. The geometry of a candidate motion tube captures the free-space required such that the robot may execute a maneuver without colliding with obstacles. Computational efficiency is achieved by selecting meaningful samples of the tube and evaluating them at run-time in the sensor space. Increasing the sensor resolution or the number of obstacles do not have any impact in the computational cost. Experimental results with a mobile platform show that free-space motion tubes are well-suited for navigating in cluttered environments and narrow passages.