A neural network-based autonomous mission planning method for deep space probes

Abstract

The invention relates to an autonomous mission planning method for a deep space detector based on a neural network, belonging to the technical field of deep space detection. The autonomous mission planning method comprises the following steps: establishing a deep space detector system module, defining an autonomous mission planning problem of the deep space detector system, determining input and output of an autonomous mission planning system, establishing a neural network model, training the neutral network by simulated data produced by a simulation system of the deep space detector, giving a group of input by the trained neutral network, and then obtaining ideal output. In case of needing to coordinate a plurality of sub systems, the autonomous mission planning method can be used for meeting real-time needs of the detector, quickly obtaining planning solution and making up the defect that communication controlled by a ground station is delayed, is beneficial to greatly improving the autonomous mission planning efficiency of the deep space detector and meeting the real-time needs of the detector, short in planning time, high in efficiency and relatively suitable for the deep space detector with high real-time needs.

Description

technical field [0001] The invention relates to a neural network-based autonomous task planning method for deep space detectors, belonging to the technical field of deep space detection. Background technique [0002] In the deep space environment, based on the perception and understanding of the space environment and the capabilities and status of the deep space probe itself, the deep space probe uses computer software technology, artificial intelligence technology, etc. The selected actions and the resource constraints provided perform reasoning and automatically generate a set of time-ordered action sequences, called a "plan". Once the plan is executed, the state of the detector can be successfully transferred to the desired goal. state. [0003] For autonomous task planning problems, being able to plan quickly in a dynamically changing environment is a major capability of intelligent behavior. Moreover, the distance between the deep space probe and the target is long, t...

AI Technical Summary

Problems solved by technology

[0005] The present invention aims at the problem of autonomous mission planning of deep space detectors, in order to overcome the shortcomings of existing heuristic-based planning methods such as difficulty in designing heuristics, large search space, and slow planning speed, etc., and proposes a neural network-based deep space detection The autonomous task planning method of the detector can meet the real-time requirements of the detector when multiple subsystems need to be coordinated

Images