Automatic Sensor Placement For Complex Three-dimensional Inspection and Exploration

This paper describes an end-to-end system for positioning sensors to observe desired parts of arbitrary initially unknown three-dimensional environments. The system finds the appropriate number of sensors and their configuration using a derivative-free cooperative optimization algorithm relying on a visibility estimation cost function. The system is tested of scenarios with different types of sensors and observation tasks. The first scenario implies six degrees of freedom floating robots to inspect a broken part of the international space station, while the second scenario involves three degrees of freedom rovers gathering information on a stuck vehicle.