Consensus sensor fusion to estimate the relative attitude during space capture operations

The concept of space capture refers to the capability of a spacecraft to grasp a target satellite with the ultimate goal of executing servicing operations on the captured vehicle. Much research effort is currently focused on the development of technologies that enable such complex operations by means of autonomous vehicles. Sensors and estimation technologies used to retrieve the relative pose of the two satellites in close proximity are of key importance.This work1 presents an algorithm for sensor fusion based on a consensus filter that leverages on the measures provided by a set of four Time-of-Flight distance sensors. The goal is to obtain an estimate of the relative orientation of two planes, one holding the sensors, the other being a target surface. The algorithm is based on the computation of five different orientations on which the consensus filter is applied. Numerical simulations show that the estimated orientation is more accurate than the simple average of the five different orientations.The proposed algorithm is performance is further evaluated using real measures collected from four Time-of-Flight sensors in a simple experimental setup. The experiment allows to compare the obtained estimation with the real orientation and with the output of a classic model-based Kalman filter used as a reference. The results suggests that the consensus algorithm estimates the relative attitude with a lower error, although requiring more measures compared to the Kalman filter.