脑注意力机制启发的群体智能协同避障方法

doi:10.11959/j.issn.2096-6652.202215

智能科学与技术学报 ›› 2022, Vol. 4 ›› Issue (1): 84-96.doi: 10.11959/j.issn.2096-6652.202215

脑注意力机制启发的群体智能协同避障方法

项羽铭¹, 陈焜¹, 赵志峰¹^,², 李荣鹏¹, 张宏纲¹

¹ 浙江大学信息与电子工程学院，浙江杭州 310027
² 之江实验室，浙江杭州 311121

修回日期:2022-01-14 出版日期:2022-03-15 发布日期:2022-03-01
作者简介:项羽铭（1999− ），男，浙江大学信息与电子工程学院在读，主要研究方向为群体智能、强化学习
陈焜（1997− ），男，浙江大学信息与电子工程学院硕士生，主要研究方向为群体智能、强化学习
赵志峰（1975− ），男，之江实验室总工程师，主要研究方向为认知无线电、无线网格网络和软件定义网络在无线通信中的应用
李荣鹏（1989− ），男，博士，浙江大学信息与电子工程学院副教授，主要研究方向为智能通信网络、网络智能、网络切片
张宏纲（1967− ），男，博士，浙江大学电子与信息工程学院教授，主要研究方向为认知无线电、绿色通信和下一代异构蜂窝网络架构
基金资助:
国家自然科学基金资助项目(61731002);国家自然科学基金资助项目(62071425);浙江省重点研发计划资助项目(2019C01002);浙江省重点研发计划资助项目(2019C03131);浙江省自然科学基金资助项目(LY20F010016);华为合作项目

Crowd intelligent cooperative obstacle avoidance method inspired by brain attention mechanism

Yuming XIANG¹, Kun CHEN¹, Zhifeng ZHAO¹^,², Rongpeng Li¹, Honggang ZHANG¹

¹ College of Information Science ＆Electronic Engineering, Zhejiang University, Hangzhou 310027, China
² Zhejiang Lab, Hangzhou 311121, China

Revised:2022-01-14 Online:2022-03-15 Published:2022-03-01
Supported by:
The National Natural Science Foundation of China(61731002);The National Natural Science Foundation of China(62071425);The Key Research and Development Program of Zhejiang Province(2019C01002);The Key Research and Development Program of Zhejiang Province(2019C03131);The Natural Science Foundation of Zhejiang Province(LY20F010016);Huawei Cooperation Project

摘要/Abstract

摘要：

现有群体智能系统执行探索任务时往往获取、计算和传输大量冗余信息，造成群智系统资源利用效率低下。设计任务驱动的计算和通信资源高效融合利用机理成为亟待解决的关键科学问题。基于此，提出了一种脑注意力机制启发的群体智能协同避障方法。受脑注意力机制启发，群智系统引入基于深度Q网络的传感器智能选择模块，实现在探索未知环境时高效选择传感器的工作状态，利用尽量少的传感器开销获取和计算关键信息；以最佳交互碰撞避免算法为基础，单一智能体融合邻居智能体的少量关键信息，驱动传感器智能选择模块，在完成集群协同避障的同时，大幅降低传感器获取和计算信息的冗余度。在仿真平台及Kehepera IV机器人实际场景中进行验证，结果显示，所提方法可以显著降低集群系统传感器的信息冗余度，并且随着智能体数量以及信息交互量的增加，性能增益更加显著。

关键词: 群体智能, 资源利用效率优化, 深度Q网络, 最佳交互碰撞避免

Abstract:

A crowd intelligent (CI) system often acquires, calculates, and transmits a large amount of redundant information during the performing of exploration tasks, which inevitably results in inefficient use of the limited resources.Therefore, it emerges a strong incentive to design a task-driven mechanism for efficient utilization of computing and communication resources.A crowd intelligent cooperative obstacle avoidance method inspired by brain attention mechanism was proposed.Inspired by brain attention mechanism, the CI system introduced an intelligent selection module based on the deep Q network, by efficiently tuning the working state of sensors exploring the unknown environment and realized the acquisition and calculation of key necessary information with as little sensor overhead as possible.Meanwhile, based on the optimal reciprocal collision avoidance algorithm, a single agent fuses a small amount of limited information from neighbor agents to drive the intelligent selection module, so as to greatly reduce the redundancy of sensor acquisition and information calculation required for the obstacle avoidance task.The effectiveness of this proposed method was verified through extensive simulation analyses and practical realization empowered with Kehepera IV robots.The results show that the proposed method can significantly reduce the redundancy of sensor information in the CI system.More importantly, as the number of agents and the amount of information interaction increase, there also emerged a clear trend in the increase of performance gains.

Key words: crowd intelligent, optimization of resource utilization efficiency, deep Q network, optimal reciprocal collision avoidance

中图分类号:

TP18

项羽铭, 陈焜, 赵志峰, 等. 脑注意力机制启发的群体智能协同避障方法[J]. 智能科学与技术学报, 2022, 4(1): 84-96.

Yuming XIANG, Kun CHEN, Zhifeng ZHAO, et al. Crowd intelligent cooperative obstacle avoidance method inspired by brain attention mechanism[J]. Chinese Journal of Intelligent Science and Technology, 2022, 4(1): 84-96.

图/表 11

图1

图2

图3

图4

图5

图6

图7

图8

表1

图9

图10

参考文献 26

[1]	沈博闻, 于宁波, 刘景泰 . 仓储物流机器人集群的智能调度和路径规划[J]. 智能系统学报, 2014,9(6): 659-664.
	SHEN B W , YU N B , LIU J T . Intelligent scheduling and path planning of warehouse mobile robots[J]. CAAI Transactions on Intelligent Systems, 2014,9(6): 659-664.
[2]	张钟元 . 室外物流配送机器人的多传感器融合同步定位与建图技术[D]. 上海:上海交通大学, 2019.
	ZHANG Z Y . Multi-sensors fusion for simultaneous localization and mapping of outdoor delivery robots[D]. Shanghai:Shanghai Jiao Tong University, 2019.
[3]	陶永, 高赫, 王田苗 ,等. 移动工业机器人在飞机装配生产线中的应用研究[J]. 航空制造技术, 2021,64(5): 32-41,67.
	TAO Y , GAO H , WANG T M ,et al. Application of mobile industrial robot in aircraft assembly production line[J]. Aeronautical Manufacturing Technology, 2021,64(5): 32-41,67.
[4]	顾宝兴, 姬长英, 王海青 ,等. 智能移动水果采摘机器人设计与试验[J]. 农业机械学报, 2012,43(6): 153-160.
	GU B X , JI C Y , WANG H Q ,et al. Design and experiment of intelligent mobile fruit picking robot[J]. Transactions of the Chinese Society for Agricultural Machinery, 2012,43(6): 153-160.
[5]	王艳新, 唐文秀, 吴函 ,等. 基于多传感器融合技术的智能抢险救灾机器人设计[J]. 林业机械与木工设备, 2018,46(5): 17-20.
	WANG Y X , TANG W X , WU H ,et al. Design of intelligent emergency rescue robots based on multi-sensor fusion technology[J]. Forestry Machinery ＆ Woodworking Equipment, 2018,46(5): 17-20.
[6]	王红君, 徐军, 赵辉 ,等. 基于平滑蚁群算法的机器人路径规划[J]. 燕山大学学报, 2017,41(3): 278-282.
	WANG H J , XU J , ZHAO H ,et al. Mobile robot path planning based on smoothing ant colony algorithm[J]. Journal of Yanshan University, 2017,41(3): 278-282.
[7]	武彬, 吴耿锋, 马飞 ,等. 基于免疫算法的移动机器人路径规划系统[J]. 计算机工程, 2004,30(12): 122-123,146.
	WU B , WU G F , MA F ,et al. An immunity algorithm-based motive robot path-planning system[J]. Computer Engineering, 2004,30(12): 122-123,146.
[8]	VADAKKEPAT P , TAN K C , WANG M L . Evolutionary artificial potential fields and their application in real time robot path planning[C]// Proceedings of 2000 Congress on Evolutionary Computation. Piscataway:IEEE Press, 2000: 256-263.
[9]	问泽藤, 温淑慧, 张迪 . 未知环境下移动机器人自主避障算法的研究[J]. 燕山大学学报, 2021,45(3): 274-282.
	WEN Z T , WEN S H , ZHANG D . Research on autonomous obstacle avoidance algorithm for mobile robots in unknown environment[J]. Journal of Yanshan University, 2021,45(3): 274-282.
[10]	LI Y M , KO?ECKA J ,, . Learning view and target invariant visual servoing for navigation[C]// Proceedings of 2020 IEEE International Conference on Robotics and Automation. Piscataway:IEEE Press, 2020: 658-664.
[11]	BERG J V D , GUY S JVD , LIN M ,et al. Reciprocal n-body collision avoidance[M]// Springer tracts in advanced robotics. Heidelberg: Springer Berlin Heidelberg, 2011: 3-19.
[12]	LI H , WENG B W , GUPTA A ,et al. Reciprocal collision avoidance for general nonlinear agents using reinforcement learning[J]. arXiv preprint, 2019,arXiv:1910.10887.
[13]	朱杰斌, 秦世引 . 无人机编队飞行的分布式控制策略与控制器设计[J]. 智能系统学报, 2010,5(5): 392-399.
	ZHU J B , QIN S Y . Distributed control strategy and controller design for UAV formation flight[J]. CAAI Transactions on Intelligent Systems, 2010,5(5): 392-399.
[14]	刘华罡, 方浩, 毛昱天 ,等. 多智能体系统分布式群集运动与避障控制[C]// 第二十九届中国控制会议论文集. 皮斯卡特维:电子与电气工程师协会出版公司, 2010: 4611-4616.
	LIU H G , FANG H , MAO Y T ,et al. Distributed flocking control and obstacle avoidance for multi-agent systems[C]// Proceedings of the 29th Chinese Control Conference. Piscataway:IEEE Press, 2010: 4536-4541.
[15]	BAHDANAU D , CHO K , BENGIO Y . Neural machine translation by jointly learning to align and translate[J]. arXiv preprint, 2014,arXiv:1409.0473.
[16]	VASWANI A , SHAZEER N , PARMAR N ,et al. Attention is all you need[C]// Proceedings of the Advances in neural information processing systems.[S.l.:s.n.], 2017: 5998-6008.
[17]	SUTTON R S , BARTO A G . Reinforcement learning:an introduction[J]. IEEE Transactions on Neural Networks, 1998,9(5): 1054.
[18]	刘朝阳, 穆朝絮, 孙长银 . 深度强化学习算法与应用研究现状综述[J]. 智能科学与技术学报, 2020,2(4): 314-326.
	LIU Z Y , MU C X , SUN C Y . An overview on algorithms and applications of deep reinforcement learning[J]. Chinese Journal of Intelligent Science and Technology, 2020,2(4): 314-326.
[19]	MNIH V , KAVUKCUOGLU K , SILVER D ,et al. Human-level control through deep reinforcement learning[J]. Nature, 2015,518(7540): 529-533.
[20]	XIN J , ZHAO H , LIU D ,et al. Application of deep reinforcement learning in mobile robot path planning[C]// Proceedings of 2017 Chinese Automation Congress. Piscataway:IEEE Press, 2017: 7112-7116.
[21]	YUE P Y , XIN J , ZHAO H ,et al. Experimental research on deep reinforcement learning in autonomous navigation of mobile robot[C]// Proceedings of 2019 14th IEEE Conference on Industrial Electronics and Applications. Piscataway:IEEE Press, 2019: 1612-1616.
[22]	罗鹏, 黄珍, 秦易晋 ,等. 基于 DQN 的车辆驾驶行为决策方法[J]. 交通信息与安全, 2020,38(5): 67-77,112.
	LUO P , HUANG Z , QIN Y J ,et al. A method of vehicle driving behavior decision based on DQN algorithm[J]. Journal of Transport Information and Safety, 2020,38(5): 67-77,112.
[23]	FIORINI P , SHILLER Z . Motion planning in dynamic environments using velocity obstacles[J]. The International Journal of Robotics Research, 1998,17(7): 760-772.
[24]	SUNEHAG P , LEVER G , GRUSLYS A ,et al. Value-decomposition networks for cooperative multi-agent learning[J]. arXiv preprint, 2017,arXiv:1706.05296
[25]	RASHID T , SAMVELYAN M , SCHROEDER C ,et al. Qmix:monotonic value function factorisation for deep multi-agent reinforcement learning[C]// Proceedings of the 35th International Conference on Machine Learning. Stockholm:PMLR, 2018.
[26]	LUO W H , SYCARA K . Adaptive sampling and online learning in multi-robot sensor coverage with mixture of Gaussian processes[C]// Proceedings of 2018 IEEE International Conference on Robotics and Automation. Piscataway:IEEE Press, 2018: 6359-6364.

方法	3个机器人	4个机器人	8个机器人
持续采样法（基线）	315.0	383.75	586.86
混合高斯模型选择	238.67 / 24.2%	303.75 / 20.8%	419.75 / 28.5%
DQN智能选择	198.0 / 37.1%	257.25 / 33.0%	—
混合高斯模型选择 + {障碍物，传感器}信息（实验4）	174.0 / 44.8%	172.0 / 55.2%	241.0 / 58.9%
DQN智能选择 + {障碍物，传感器}信息（实验5）	147.0 / 53.3%	155.5 / 59.5%	220.86 / 62.4%

脑注意力机制启发的群体智能协同避障方法

Crowd intelligent cooperative obstacle avoidance method inspired by brain attention mechanism

在线阅读

PDF下载

可视化

摘要/Abstract

引用本文

使用本文

图/表 11

参考文献 26

相关文章 5

Metrics

推荐阅读 0

[1]	郁洲, 毕敬, 苑海涛. 基于改进DQN算法的复杂海战场路径规划方法[J]. 智能科学与技术学报, 2022, 4(3): 418-425.
[2]	王雨倩, 丁嵘. 基于群体智能成果的路径规划程序自动生成系统[J]. 智能科学与技术学报, 2022, 4(2): 255-263.
[3]	杨强, 童咏昕, 王晏晟, 范力欣, 王薇, 陈雷, 王魏, 康焱. 群体智能中的联邦学习算法综述[J]. 智能科学与技术学报, 2022, 4(1): 29-44.
[4]	冯埔, 吴文峻, 罗杰, 于鑫, 田雍恺. 基于群体熵的机器人群体智能汇聚度量[J]. 智能科学与技术学报, 2022, 4(1): 65-74.
[5]	宿帅, 朱擎阳, 魏庆来, 唐涛, 阴佳腾. 基于DQN的列车节能驾驶控制方法[J]. 智能科学与技术学报, 2020, 2(4): 372-384.