复杂时空约束下的多智能体运动规划

定　　价：79 元

• 作者：李石磊 等
• 出版时间：2023/12/1
• ISBN：9787121470110
• 出 版 社：电子工业出版社

• 中图法分类：TP24 
• 页码：164
• 版次：01
• 开本：16开

内容简介

本书主要介绍了满足多个时空约束要求的多智能体运动规划技术，全书共8章：第1～2章阐述了运动规划的基本概念、相关技术，重点讲述了智能体多层次行为模型；第3～6章分别从不同角度对多智能体时空约束建模方法进行了讲解；第7章介绍了相关仿真应用案例；第8章对深度强化学习在运动规划中的应用进行了探索研究。 本书适合从事智能无人平台运动规划研究、开发的技术人员，以及相关专业的高校师生阅读参考。

作者介绍

李石磊，海军工程大学信息安全系讲师、中国仿真学会仿真技术应用专业委员会委员。主要从事复杂系统建模与仿真、信息安全技术研究与教学，主持了国家自然科学基金青年项目“物理仿真虚拟人运动控制技术研究”、博士后基金项目“复杂动态场景多智能体运动规划技术”，参与了国家自然科学基金面上项目“数据和模型混合驱动的虚拟人群仿真及其在军事中的应用研究”等项目。

目 录

第1 章 智能体运动规划研究概述·······································································（1）
1.1 研究背景···························································································（1）
1.2 研究现状···························································································（1）
1.3 研究思路···························································································（8）
第2 章 运动规划中的智能体多层次行为模型框架设计·········································.（11）
2.1 环境建模方法··················································································.（11）
2.1.1 三角剖分法···········································································.（11）
2.1.2 可视图方法···········································································.（12）
2.1.3 Voronoi 图法··········································································.（12）
2.1.4 随机采样法···········································································.（13）
2.2 运动规划方法··················································································.（16）
2.2.1 全局运动规划········································································.（16）
2.2.2 局部运动规划········································································.（20）
2.3 智能体多层次行为模型设计································································.（23）
2.3.1 智能体介绍···········································································.（23）
2.3.2 现有智能体行为模型框架·························································.（24）
2.3.3 智能体多层次行为模型框架······················································.（27）
第3 章 基于多信息域、多分辨率场景描述模型的多层次运动规划算法研究··············.（31）
3.1 相关工作························································································.（31）
3.2 层次化运动规划算法总体思路·····························································.（34）
3.2.1 问题描述··············································································.（34）
3.2.2 总体思路··············································································.（35）
3.3 全局层次上的多样化引导路径生成·······················································.（36）
3.3.1 高层时空约束的表示·······························································.（36）
3.3.2 全局概率路径图的生成····························································.（38）
3.3.3 全局多样化引导路径的生成······················································.（39）
3.4 局部层次上的多样化运动路径生成·······················································.（41）
3.4.1 局部高分辨率概率路径图的生成················································.（41）
3.4.2 局部多样化运动路径的生成······················································.（42）
3.5 动态障碍情况下的运动路径重新规划问题··············································.（43）
3.5.1 运动路径的重新规划·······························································.（43）
3.5.2 概率路径图的动态更新····························································.（44）
3.6 仿真实验与结果分析·········································································.（45）
第4 章 多智能体避碰行为研究·······································································.（48）
4.1 相关工作························································································.（48）
4.2 避碰行为的基本概念与问题描述··························································.（50）
4.2.1 基本概念··············································································.（50）
4.2.2 问题描述··············································································.（51）
4.3 反应式避碰行为模型·········································································.（52）
4.4 基于最小代价原则的预测式避碰行为模型··············································.（57）
4.4.1 最小代价原则········································································.（57）
4.4.2 基于最小代价原则的预测式避碰行为建模····································.（58）
4.5 仿真实验与结果分析·········································································.（63）
第5 章 基于增广物理仿真的运动规划多任务约束建模分析与优化求解····················.（67）
5.1 相关工作························································································.（67）
5.2 问题描述························································································.（70）
5.3 控制围栏函数基础背景知识································································.（70）
5.4 基于控制围栏函数的多任务约束统一描述与运动路径动力学物理仿真优化生成···.（71）
5.4.1 基于控制围栏函数的多任务约束统一描述····································.（72）
5.4.2 增广物理仿真框架下的智能体运动路径优化生成···························.（74）
5.5 仿真实验与结果分析·········································································.（77）
5.5.1 可生成多样化行为的智能体运动路径优化计算······························.（77）
5.5.2 自动驾驶中自适应巡航运动路径优化计算····································.（81）
5.6 多智能体运动规划仿真演示软件设计开发··············································.（83）
5.6.1 演示软件设计开发··································································.（83）
5.6.2 多智能体运动路径生成测试······················································.（85）
第6 章 多任务约束时空融合处理机制研究························································.（87）
6.1 相关工作························································································.（87）
6.2 具有速度和加速度约束的多智能体时间最优轨迹规划·······························.（88）
6.2.1 问题描述··············································································.（88）
6.2.2 算法思路及流程·····································································.（92）
6.2.3 可达集与可达速度区间的计算···················································.（95）
6.2.4 利用VIP 算法计算时间最优轨迹················································.（97）
6.2.5 仿真实验与结果分析·······························································.（99）
6.3 基于控制围栏函数描述的多任务约束时空融合处理··································（107）
6.3.1 同时段下多任务约束时序冲突的自动优化处理······························（107）
6.3.2 仿真实验与结果分析·······························································（111）
第7 章 满足时空约束要求的多智能体运动规划在城市维稳处突行动仿真场景下的应用····.（115）
7.1 应用场景描述··················································································（115）
7.2 可视化仿真·····················································································（117）
7.2.1 仿真参数设置········································································（117）
7.2.2 仿真结果演示········································································（118）
第8 章 深度强化学习在运动规划中的应用探索研究············································（124）
8.1 相关工作························································································（124）
8.2 深度强化学习算法探索效率提升策略研究··············································（126）
8.2.1 基本思路··············································································（127）
8.2.2 遗传算法和DDPG 算法相结合的DRL 算法探索效率提升策略··········（129）
8.2.3 仿真实验与结果分析·······························································（132）
8.3 基于深度强化学习的多智能体避碰行为生成···········································（134）
8.3.1 问题描述与基本思路·······························································（134）
8.3.2 仿真实验与结果分析·······························································（135）
8.4 下一步展望·····················································································（136）
参考文献······································································································（137）