Development of an autonomous spacecraft for planetary exploration

摘要

The purpose of this paper is to present a new approach in the concept and implementation of autonomous micro-spacecraft. The one true "artificial agent" approach to autonomy requires the micro-spacecraft to interact in a direct manner with the environment through the use of sensors and actuators. As such, little computational effort is required to implement such an approach, which is clearly of great benefit for limited microsatellites. Rather than using complex world models, which have to be updated, the agent is allowed to exploit the dynamics of its environment for cues as to appropriate actions to achieve mission goals. The particular artificial agent implementation used here has been borrowed from studies of biological systems, where it has been used successfully to provide models of motivation and opportunistic behaviour. The so-called "cue-deficit" action selection algorithm considers the micro-spacecraft to be a nonlinear dynamical system with a number of observable states. Using optimal control theory rules are derived which determine which of a finite repertoire of behaviours the satellite should select and perform. The principal benefits of this approach is that the micro-spacecraft is endowed with selfsufficiency, defined here to be the ability to achieve mission goals, while never placing itself in an irrecoverable position.