Statistical Approach for the Definition of Satellite Pattern of Life

In an overly crowded space environment, the tracking of orbital satellites, and thus the Space Situational Awareness is becoming a critical sector. Given the increasing challenge of tracking objects in the presence of maneuvers and detecting them, establishing a recursive behaviour of the satellite, called Pattern of Life, can help in pursuing this task. In fact, knowledge of the Pattern of Life of a Resident Space Object also plays a critical role in predicting when the satellite is expected to maneuver, which can serve both civilian (improving tracking capability) and military purposes (identifying anomalous maneuvers performed by unfriendly assets). In this context, many techniques for generating the Pattern of Life of an active Space Object are developed using deep learning approaches, whereas this paper presents a method for Pattern of Life generation which is based on a statistical analysis of both the detected maneuvers and satellite orbital parameters, and can also detect possible changes in the behaviour of a satellite.