基于神经网络和人工势场的协同博弈路径规划

张菁; 何友; 彭应宁; 李刚

doi:10.7527/S1000-6893.2018.22493

航空学报 >

2019 , Vol. 40 >Issue 3: 322493 - 322493

DOI: https://doi.org/10.7527/S1000-6893.2018.22493

电子电气工程与控制

基于神经网络和人工势场的协同博弈路径规划

张菁 ,
何友 ,
彭应宁 ,
李刚

展开

1. 清华大学电子工程系, 北京 100084;
2. 复杂航空系统仿真重点实验室, 北京 100076;
3. 海军航空大学信息融合研究所, 烟台 264001

收稿日期: 2018-06-28

修回日期: 2018-07-31

网络出版日期: 2018-09-17

基金资助

中国博士后科学基金（2018M631483）

收起

Neural network and artificial potential field based cooperative and adversarial path planning

ZHANG Jing ,
HE You ,
PENG Yingning ,
LI Gang

Expand

1. Department of Electronic Engineering, Tsinghua University, Beijing 100084, China;
2. Science and Technology on Complex Aviation Systems Simulation Laboratory, Beijing 100076, China;
3. Research Institute of Information Fusion, Naval Aeronautical University, Yantai 264001, China

Received date: 2018-06-28

Revised date: 2018-07-31

Online published: 2018-09-17

Supported by

Postdoctoral Science Foundation of China (2018M631483)

Fold

摘要

协同博弈路径规划是空战自主决策、机器人体育比赛等应用场景中的重要问题，其难点在于对环境对抗性反馈的实时自适应和多智能体的相互配合。提出一种基于神经网络和人工势场的协同博弈路径规划方法，使用反向传播（BP）神经网络自适应调整人工势场函数系数，并将人工势场作为神经网络输出端的特征提取。为解决真实样本质量和数量不足的问题，基于遗传算法仿真生成样本数据用于神经网络训练，并通过滚动时域的思路面向动态博弈优化样本性能。从样本数据中提炼出距离差与航向差以反映协同和博弈特性，利用神经网络的黑盒特性和学习能力解决协同博弈问题。应用于二对一反隐身超视距空战路径规划，比经典人工势场法有明显性能提升，且计算开销可接受，计算复杂度分析表明该方法可以较好扩展到多机对抗场景。

关键词： 协同博弈路径规划; 仿真生成样本; 神经网络; 人工势场; 遗传算法; 滚动时域优化; 超视距空战; 反隐身

本文引用格式

张菁 , 何友 , 彭应宁 , 李刚 . 基于神经网络和人工势场的协同博弈路径规划[J]. 航空学报, 2019 , 40(3) : 322493 -322493 . DOI: 10.7527/S1000-6893.2018.22493

Abstract

Cooperative and adversarial path planning is a significant issue in scenarios such as air combat and sport games. The challenge is the adaptation for dynamic feedback and the cooperation between multi-agents. A neural network and artificial potential field based method is proposed, in which the potential gain coefficient of the artificial potential field is adaptively adjusted by the Back Propagation (BP) neural network. The artificial potential field can be seen as feature extraction for the neural network on its output phase. To face the issue of insufficient natural samples for training the neural network, the sample is generated by simulation and optimized by a genetic algorithm and receding horizon optimization. The "different of distance" and "different of heading" are defined to show the characters of cooperative and adversarial path planning, and the black box feature and learning capacity of neural network are well exploited for cooperative and adversarial path planning. This method is evaluated in a two on one anti-stealth beyond-visual-range air combat, and shows significant improvement of performance and affordable costs. The computational complex analysis shows that our algorithm is scalable for multi-aircrafts cases.

Key words： cooperative and adversarial path planning; simulated sample generation; neural network; artificial potential field; genetic algorithm; receding horizon optimization; beyond-visual-range air combat; anti-stealth

参考文献

[1] ERNEST N, CARROLL D, SCHUMACHER C, et al. Genetic fuzzy based artificial intelligence for unmanned combat aerial vehicle control in simulated air combat missions[J]. Journal of Defense Management, 2016, 6:144.
[2] SCHVANEVELDT R W, GOLDSMITH T E, BENSON A E, et al. Neural network models of air combat maneuvering:AL-TR-1992-0037[R]. Texas:Air Force Materiel Command, 1992.
[3] 张菁, 何友, 彭应宁, 等. 新一代战斗机技术特征论证[J]. 系统工程与电子技术, 2018, 40(8):1754-1759. ZHANG J, HE Y, PENG Y N, et al. Technical features demon-stration of the next generation fighter[J]. Journal of Systems Engineering and Electronics, 2018, 40(8):1754-1759(in Chinese).
[4] 左家亮, 杨任农, 张滢, 等. 基于启发式强化学习的空战机动智能决策[J]. 航空学报, 2017, 38(10):321168. ZUO J L, YANG R N, ZHANG Y, et al. Intelligent decision-making in air combat maneuvering based on heuristic reinforcement learning[J]. Acta Aeronautica et Astronautica Sinica, 2017, 38(10):321168(in Chinese).
[5] 符小卫, 李金亮, 高晓光. 威胁联网下无人作战飞机突防作战航迹规划[J]. 航空学报, 2014, 35(4):1042-1052. FU X W, LI J L, GAO X G. Defense penetration path planning for UCAV based on threat netting[J]. Acta Aeronautica et Astronautica Sinica, 2014, 35(4):1042-1052(in Chinese).
[6] 周金良, 黄彦文, 曹其新. 对抗环境下足球机器人路径规划[J]. 上海交通大学学报, 2006, 40(11):1827-1831. ZHOU J L, HUANG Y W, CAO Q X. The path planning for robot soccer under antagonistic environment[J]. Journal of Shanghai Jiaotong University, 2006, 40(11):1827-1831(in Chinese).
[7] SU W, MENG R, YU C. A study on soccer robot path planning with fuzzy artificial potential field[C]//International Conference on Computing, Control and Industrial Engineering. Piscataway, NJ:IEEE Press, 2010:386-390.
[8] 朱毅, 张涛, 宋靖雁. 未知环境下势场法路径规划的局部极小问题研究[J]. 自动化学报, 2010, 36(8):1122-1130. ZHU Y, ZHANG T, SONG J Y, et al. Study on the local minima problem of path planning using potential field method in unknown environments[J]. Acta Automatica Sinica, 2010, 36(8):1122-1130(in Chinese).
[9] GE S S, CUI Y J. New potential functions for mobile robot path planning[J]. IEEE Transactions on Robotics and Automation, 2000, 16(5):615-620.
[10] 张建英, 刘暾. 基于人工势场法的移动机器人最优路径规划[J]. 航空学报, 2007, 28(S1):S183-S188. ZHANG J Y, LIU T. Optimized path planning of mobile robot based on artificial potential field[J]. Acta Aeronautica et Astronautica Sinica, 2007, 28(S1):S183-S188(in Chinese).
[11] KOREN Y, BORENSTEIN J. Potential field methods and their inherent limitations for mobile robot navigation[C]//IEEE International Conference on Robotics and Automation. Piscataway, NJ:IEEE Press, 1991(2):1398-1404.
[12] TOIT N E D, BURDICK J W. Robot motion planning in dynamic, uncertain environments[J]. IEEE Transactions on Robotics, 2012, 28(1):101-115.
[13] LUO C, YANG S X. A bioinspired neural network for real-time concurrent map building and complete coverage robot navigation in unknown environments[J]. IEEE Transactions on Neural Networks, 2008, 19(7):1279-1298.
[14] YANG S X, MENG M. Neural network approaches to dynamic collision-free trajectory generation[J]. IEEE Transactions on Systems Man & Cybernetics Part B Cybernetics, 2001, 31(3):302-318.
[15] KASSIM A A, KUMAR B V K V. A neural network architecture for path planning[C]//International Joint Conference on Neural Networks. Piscataway, NJ:IEEE Press, 2002:787-792.
[16] KUMAR B V. Neural network architecture for generating potential fields for path planning[C]//Proceedings of SPIE-the International Society for Optical Engineering. Bellingham, WA:SPIE, 1992, 1766.
[17] ZHU A, YANG S X. A neural network approach to dynamic task assignment of multirobots[J]. IEEE Transactions on Neural Networks, 2006, 17(5):1278-1287.
[18] GLASIUS R, KOMODA A, GIELEN S C A M. Neural network dynamics for path planning and obstacle avoidance[J]. Neural Networks, 1995, 8(1):125-133.
[19] NA Y K, OH S Y. Hybrid control for autonomous mobile robot navigation using neural network based behavior modules and environment classification[J]. Autonomous Robots, 2003, 15(2):193-206.
[20] 何炎祥, 范收平. 基于神经网络和人工势场的滚动规划[J]. 计算机工程与应用, 2005, 41(31):66-68. HE Y X, FAN S P. Rolling planning based on neural networks and artificial potential field[J]. Computer Engineering & Applications, 2005, 41(31):66-68(in Chinese).
[21] 禹建丽, 孙增圻, 成久洋之. 一种快速神经网络路径规划算法[J]. 机器人, 2001, 23(3):201-205. YU J L, SUN Z Q, KROUMOV V. Fast algorithm for path planning based on neural network[J]. Robot, 2001, 23(3):201-205(in Chinese).
[22] 陈世春, 黄沛霖, 姬金祖. 典型隐身飞机的RCS起伏统计特性[J]. 航空学报, 2014, 35(12):3304-3314. CHEN S C, HUANG P L, JI J J. Radar cross section fluctuation characteristics of typical stealth aircraft[J]. Acta Aeronautica et Astronautica Sinica, 2014, 35(12):3304-3314(in Chinese).
[23] PARSCH A. Raytheon (Hughes) AIM-120 AMRAAM[EB/OL]. (2007-07-25)[2018-09-16].http://www.designation-systems.net/dusrm/m-120.html.
[24] KHATIB O. Real-time obstacle avoidance for manipulators and mobile robots[J]. International Journal of Robotics Research, 1986, 5(5):500-505.

Options

文章导航

模态框（Modal）标题

摘要

本文引用格式

Abstract

参考文献