Experimental Evaluation of a COTS Time-of-Flight Camera as Rendezvous Sensor for Small Satellites

With the growing number of space debris objects and the advent of launching 20,000 or more new satellites into “mega constellations”, the idea of on-orbit servicing and space debris removal has gained renewed interest. The big challenge of debris removal and servicing is to remain cost efficient. In order to keep costs low, the servicing and removal vehicles must be cheap to produce, launch and operate. Therefore, small satellites, ideally multi-unit cubesats equipped with simple sensors, propulsion systems and capture mechanisms must be developed. The mass, volume and power restrictions for these systems all but rule out the use of the typical sensor systems for rendezvous and capture, such as radar or lidar. Instead, camera-based systems functioning in the challenging optical, thermal and radiation conditions of space must be developed. The paper describes an experimental study evaluating the use of a Microsoft Kinect v2 sensor as rendezvous and capture sensor. The sensor is used in a hardware-in-the-loop spacecraft maneuver simulator to detect position, velocity and attitude of an object representing a non-cooperative servicing or removal target. The paper describes the implementation of a sensor system built around the Kinect v2 sensor and the open source Open3D point cloud manipulation software. The paper then proceeds to describe the laboratory experiments run to test the performance of the system in determining the position of a chaser vehicle relative to the target object, with a twelve-camera OptiTrack system serving to establish ground truth. The tests in the controlled lab environment show that the system is readily able to measure relative position to within 0.005 m along its boresight axis at a rate on the order of 1 Hz, without any knowledge of the target geometry and without distinct reflectors or other fiducials placed on the target.