近期，实验室教师姚蔚然作为第一作者的论文“Evolutionary Utility Prediction Matrix-Based Mission Planning for Unmanned Aerial Vehicles in Complex Urban Environments”被国际权威期刊IEEE Transactions on Intelligent Vehicles录用。

本文提出了一种进化效用预测矩阵（EUPM）方法，用于解决城市场景中执行异构任务的无人机群的输入输出耦合任务规划问题。采用分布式任务规划体系结构，建立了城市任务的结构化框架。根据典型的城市任务模式，分析了任务规划子问题之间的输入-输出耦合关系。任务规划子问题的模块设计具有改进的输入-输出耦合关系。仿真和硬件在环实验表明，EUPM方法相比于传统方法，更有效地实现了效用预测和高效的任务规划。

Abstract

An unmanned aerial vehicle (UAV) swarm that self-organizes to provide superior intelligence and overwhelming effects is a promising technology in urban scenarios. Complex terrain constraints of urban environment increase mission planning difficulties, and bring a strong input-output coupling between subproblems of mission planning, which affect the computing performance and effectiveness of the UAV swarm. In this paper, an evolutionary utility prediction matrix (EUPM) method is presented to solve the input-output-coupled mission planning problem for a UAV swarm executing heterogeneous tasks in an urban scenario with complex constraints. A structured framework of urban missions is established with a distributed mission planning architecture. The input-output coupling relationships between the subproblems of mission planning are analyzed in terms of typical urban mission patterns. Modules for the mission planning subproblems are designed with improved input-output coupling relationships. Simulations and hardware-in-loop experiments demonstrate that the EUPM method achieves accurate prediction of the utility and high effective mission planning solutions than the traditional methods.