Algorithm training and verification system for unmanned system cluster cooperative control

Abstract

The present application relates to the technical field of unmanned control, in particular to an algorithm training and verification system for cluster cooperative control of unmanned systems, a cluster control module generates cooperative task instructions and distributes them to each heterogeneous unmanned module; the heterogeneous unmanned module receives the cooperative task instructions and executes them; a model initialization module generates a virtual model based on the initial positions of multiple heterogeneous unmanned modules, and generates predicted movement data of multiple heterogeneous unmanned units based on the cooperative task instructions; a data synchronization module obtains actual execution data of multiple heterogeneous unmanned units, and synchronizes them to the virtual mirror model based on timestamps to align with the predicted movement data; a feature extraction module extracts training features based on the actual execution data and the predicted movement data, the training features including formation keeping degree, cluster overall motion consistency and task progress deviation rate; an efficiency evaluation module evaluates the efficiency of the control algorithm based on the training features, which can automatically verify the cooperative control algorithm and improve work efficiency.

Images