Multi-layer and multi-granularity cluster task planning method based on sequential assignment

Abstract

The invention discloses a multi-layer and multi-granularity cluster task planning method based on sequential assignment, which includes multi-layer dimension reduction belief information representation, inter-layer information interaction model and sequential assignment of online prediction functions, and the multi-layer dimension reduction belief information is expressed as Establish a unified patrol model. In the inter-layer information interaction model, the granularity of information that each layer of drones pays attention to is different, but there is a certain correspondence, and the correspondence between each layer of drones mainly includes bottom-up Belief information transmission and top-down command information transmission; the order distribution of the online prediction function is based on myopia scheduling algorithm, and the myopia scheduling algorithm includes single UAV patrol based on heuristic function, sequential allocation technology based on Centralized scheduling of subgroup patrols and cluster patrols. The invention can better realize the task planning of the continuous reconnaissance of the unmanned aerial vehicle cluster under complicated circumstances, and is beneficial to popularization and application.

Description

technical field [0001] The invention relates to the technical field of unmanned aerial vehicle swarm control, in particular to a multi-layer multi-granularity swarm task planning method based on sequential allocation. Background technique [0002] UAV swarms have developed rapidly in recent years, especially UAV swarms with reconnaissance capabilities, which are becoming a common means of situational awareness. We consider a scenario where a swarm of drones is constantly scouting an area to gather as much situational information as possible. This is very common in reality, such as searching for missing tourists in the mountains, reconnaissance of the battlefield environment, obtaining disaster information and patrolling plant protection, etc. In these areas, the environment is dynamic and uncertain, and each drone can only obtain local information at the current moment. In other words, the environment is partially observable. Therefore, the difficulty lies in how to carry...

AI Technical Summary

Problems solved by technology

However, as the number of drones increases, the number of possible actions increases exponentially, and some baseline algorithms, such as Partially Observable Monte Carlo schemes, struggle to solve our problem

Images