Agile reactive navigation for a non‐holonomic mobile robot using a pixel processor array

This paper presents an agile reactive navigation strategy for driving a non-holonomic ground vehicle around pre-set course of gates in cluttered environment using low-cost processor array sensor. enables machine vision tasks to be performed directly upon the sensor's image plane, rather than separate general-purpose computer. The authors demonstrate small running through or avoiding multiple at high speed minimal computational resources. To achieve this, target tracking algorithms are developed Pixel Processing Array and captured images then processed on sensor acquiring information controlling vehicle. algorithm can run up 2000 fps outdoors 200 indoor illumination levels. Conducting processing level avoids bottleneck transfer encountered conventional sensors. real-time performance on-board robustness is validated experiments. Experimental results algorithm's ability enable navigate average 2.20 m/s passing 3.88 ‘slalom’ task featuring significant visual clutter.