A multi-person cooperative task simulation and data integration method and platform for performance evaluation

By constructing a multi-layered variable space through an integrated simulation platform, introducing a composite novelty metric and a hierarchical attention fusion network, and generating a multi-dimensional performance feature time series matrix, the problem of real-time performance evaluation for multi-agent collaborative tasks is solved, achieving efficient and robust performance evaluation and decision support.

CN122134044APending Publication Date: 2026-06-02CHINA STATE SHIPBUILDING CORP LTD RESEARCH INSTITUTE 719
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
CHINA STATE SHIPBUILDING CORP LTD RESEARCH INSTITUTE 719
Filing Date
2026-03-18
Publication Date
2026-06-02

AI Technical Summary

Technical Problem

Existing technologies struggle to achieve real-time performance evaluation of multi-agent collaborative tasks, and traditional methods cannot effectively integrate massive amounts of heterogeneous data, resulting in poor robustness of evaluation results, high computational resource consumption, and a lack of attribution explanation for key influencing factors.

Method used

An integrated simulation platform is adopted. By constructing a four-layer variable space of task, personnel, equipment and environment, a composite novelty metric function and a hierarchical attention fusion network are introduced to generate a robust multi-dimensional performance feature time series matrix. New profiles are generated through multi-objective Bayesian optimization iteration, and an interactive dashboard is provided to display performance comparison and attribution heatmap.

Benefits of technology

It achieves intelligent fusion driven by data quality, improves the accuracy of feature extraction, reduces the consumption of computing resources, provides efficient multi-objective optimization and deep attribution explanation capabilities, and supports real-time decision support.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122134044A_ABST
    Figure CN122134044A_ABST
Patent Text Reader

Abstract

This invention discloses a method and platform for multi-person collaborative task simulation and data integration for performance evaluation. The method includes: constructing a multi-level variable space and generating experimental design profiles with high information content and large distribution differences using the Composite Novelty Metric (CNMF); collecting simulation participant operation logs, equipment entity status, and environmental parameters in real time, and dynamically scoring each data source; generating a standardized multi-dimensional performance feature time series matrix; employing a multi-task deep learning network to output overall performance scores, classification results, and attribution heatmaps; using a multi-objective Bayesian optimization (MOBO) algorithm to find the optimal solution on the Pareto front to drive a new round of simulation process; and generating an interactive dashboard to provide users with intuitive decision-making assistance. This invention solves the technical problems in traditional simulation evaluation, such as data fusion being greatly affected by noise, low experimental design efficiency, difficulty in multi-objective optimization, and lack of deep attribution explanation, thus realizing intelligent and automated performance evaluation.
Need to check novelty before this filing date? Find Prior Art