Guidance Strategy for Autonomous Inspection of Unknown Non-Cooperative Resident Space Objects

Abstract

As space debris has become a cause of concern for space operations around Earth, active debris removal and satellite servicing missions have gained increasing attention. Within this framework, in specific scenarios, the chaser might be asked to operate autonomously in the vicinity of a non-cooperative, unknown target. This paper presents a sampling-based receding-horizon motion planning algorithm that selects inspection maneuvers while taking many complex constraints into account. The proposed guidance solution is compared with classical approaches and it is shown to take advantage of the characteristics of the natural dynamics of the relative motion to outperform them. In addition, the impact of different input sampling exploration strategies is explored to propose a simple and more robust approach based on subset simulation.

Instance Segmentation for Feature Recognition on Noncooperative Resident Space Objects

“…In fact, even when not supported by proper hardware, the method could still be employed to gather additional information on the geometry, configuration, and conditions of the unknown target, e.g. during the observation phases scheduled in the mission paradigm envisioned by Maestrini et al [7].…”

“…This information may turn out to be paramount for the determination of attitude, geometrical properties, and mass distribution of the observed RSO, should it be in an off-nominal configuration. Think, for example, of promising inspection techniques as the one explained in [7] or other based on the concept of safety ellipses. Since the center of mass of the object is assumed to coincide with the center of the ellipsoid, these techniques would fail in the case of highly asymmetric or elongated objects.…”

“…However, it has been demonstrated that such investigation techniques could suffer from considerable shortcomings when dealing with unknown objects [7], whose tumbling attitude can't be assessed in advance. Indeed, should the target be controlled (i.e., not tumbling) this would impair the visibility of certain regions of its surface if the relative trajectory is kept constant.…”

Design and guidance of a multi-active debris removal mission

“…The space environment in the low Earth orbit (LEO) is becoming increasingly congested with space debris. As a result, the average rate of debris collisions has increased to four or five objects per year [1]. Most debris consist of artificial objects, such as derelict satellites, discarded rocket stages, and fragments originating from collisions.…”

Approximate Quantiles for Stochastic Optimal Control of LTI Systems with Arbitrary Disturbances

“…Techniques for calculating collision avoidance probabilities have been posed [22], but are difficult to embed in control algorithms and may be limited by the shape of the satellite. Methods for rendezvous and proximity operations between a controlled satellite and a stationary or potentially non-cooperative satellite have been proposed [23][24][25] but lack in their ability to accommodate multiple controlled vehicles. Previous work of ours [26] has solved this problem in Gaussian environments with difference-of-convex function optimization but reliance on predetermining risk allocation between constraints led to sub-optimal solutions, and the method cannot accommodate non-Gaussian disturbances.…”