摘要
面向地震场景,提出了一种有效的空地协同搜救框架,高空侦察机获取地面受损建筑物位置信息传递给无人机搜索群体,无人机根据此信息做分布式任务规划,到达目标建筑对受灾人群做进一步侦察,并将伤员密度分布信息传递给地面无人车群体。无人车做集中式规划,到达伤员地点执行救援任务。针对无人机群体的任务规划,根据其小型、廉价、可大规模部署以及通信能力弱的特点,改进传统的拍卖任务规划方法,提出了一种新的基于通信保持的拍卖方法(CMBA);救援无人车群体虽然载荷能力强,但是在灾区恶劣道路环境条件下,无法大规模部署,必须发挥其执行任务的最大效用,传统遗传算法适用于中心化的精确任务规划,但是存在易陷入局部最优解的缺点,提出了一种自适应反馈调节遗传算法(AFBA-GA)改进这一缺点。通过在机器人仿真器中和标准数据集中的测试,验证了任务规划方法的有效性。
A novel method of swarm robot task planning based on air and ground coordination is proposed for earthquake search and rescue. A surveillance aircraft gets the positions of the damaged buildings, and passes this information to the UAV swarm. The UAV swarm uses a decentralized method to plan the searching paths based on this information. They get the positions and number of the injured people, and pass this information to the unmanned vehicles. The vehicles rescue the people according to it. A method of communication-maintain-based auction(CMBA) is provided for the UAV planning, considering the UAV features including that they are cheap, small and easy to be deployed massively. A planning method for the unmanned vehicles is provided, which is a centralized approach of adaptive-feedback-based adjustment-GA(AFBA-GA). Compared with traditional GA, our method is easier to escape from local optimum. By the experiments in the simulated robot environment and the tests on a CVRP benchmark, the effectiveness of the planning methods is validated.
引文
[1]CLAES D,OLIEHOEK F,BAIER H,et al.Decentralised online planning for multi-robot warehouse commissioning[C]//AAMAS,2017:492-500.
[2]BLENDER T,BUCHNER T,FERNANDEZ B,et al.Managing a mobile agricultural robot swarm for a seeding task[C]//IECON 2016-42nd Annual Conference of the IEEE Industrial Electronics Society.IEEE,2016:6879-6886.
[3]BAKER C A B,RAMCHURN S,TEACY W T,et al.Planning search and rescue missions for UAV teams[C]//Proceedings of the Twenty-second European Conference on Artificial Intelligence.IOS Press,2016:1777-1778.
[4]WEISS G.Multiagent systems:A modern approach to distributed artificial intelligence[M].MIT Press,2000.
[5]BROOKS R S.A robust layered control system for a mobile robot[J].IEEE J.robot.autom,1986,2(1):14-23.
[6]MENDOZA J P,BISWAS J,COOKSEY P,et al.Selectively reactive coordination for a team of robot Soccer Champions[C]//AAAI,2016:3354-3360.
[7]PARKER L E.ALLIANCE:An architecture for fault tolerant multi-robot cooperation[J].IEEE Trans.Robotics Automat.,1998,14(2):220-240.
[8]葛斯楠.基于改进蚁群算法的多机器人任务分配[D].沈阳:沈阳工业大学,2016.GE Sinan.Path planning for robots based on ant colony algorithm[D].Shenyang:Shenyang University of Technology,2016.
[9]JIAN J,ZANG X Z,JIE Z.Cooperative multi-robot foraging based on anxiety conception and auction method[J].Control&Decision,2008,23(5):541-545.
[10]NUNES E,GINI M L.Multi-robot auctions for allocation of tasks with temporal constraints[C]//AAAI,2015:2110-2116.
[11]TALEBPOUR Z,SAVARE S,MARTINOLI A.Market-based coordination in dynamic environments based on the Hoplites framework[C]//International Conference on Intelligent Robots and Systems,IEEE,2017:1105-1112.
[12]SILVER D,HUANG A,MADDISON C J,et al.Mastering the game of Go with deep neural networks and tree search[J].Nature,2016,529(7587):484-489.
[13]KARTAL B,NUNES E,GODOY J,et al.Monte Carlo tree search with branch and bound for multi-robot task allocation[C]//The IJCAI'16 Workshop on Autonomous Mobile Service Robots,2016:1.
[14]KARTAL B,GODOY J,KARAMOUZAS I,et al.Stochastic tree search with useful cycles for patrolling problems[C]//IEEE International Conference on Robotics and Automation,IEEE,2015:1289-1294.
[15]GOLDBERG D E.Genetic algorithm in search optimization and machine learning[J].Addison Wesley,1989,13(7):2104-2116.
[16]BRIAN G,WILLIAM W.ROS stage robot simulator[EB/OL].http://wiki.ros.org/stage_ros,2013.
[17]QUIGLEY M,GERKEY B P,CONLEY K,et al.ROS:An open-source robot operating system[C]//Proc.IEEEICRA Workshop on Open Source Robotics,2009.
[18]Augerat.CVRP测试数据集[DB/OL].http://neo.lcc.uma.es/vrp/vrp-instances/capacitated-vrp-instances/,2013.Augerat.CVRP test data set[DB/OL].http://neo.lcc.uma.es/vrp/vrp-instances/capacitated-vrp-instances/,2013.
