本書主要介紹了滿足多個時空約束要求的多智能體運動規劃技術，全書共8章：第1～2章闡述了運動規劃的基本概念、相關技術，重點講述了智能體多層次行為模型；第3～6章分別從不同角度對多智能體時空約束建模方法進行了講解；第7章介紹了相關仿真應用案例；第8章對深度強化學習在運動規劃中的應用進行了探索研究。 本書適合從事智能無人平臺運動規劃研究、開發的技術人員，以及相關專業的高校師生閱讀參考。
李石磊，海軍工程大學信息安全系講師、中國仿真學會仿真技術應用專業委員會委員。主要從事復雜系統建模與仿真、信息安全技術研究與教學，主持了國家自然科學基金青年項目“物理仿真虛擬人運動控制技術研究”、博士后基金項目“復雜動態場景多智能體運動規劃技術”，參與了國家自然科學基金面上項目“數據和模型混合驅動的虛擬人群仿真及其在軍事中的應用研究”等項目。
第1 章 智能體運動規劃研究概述·······································································（1）
1.1 研究背景···························································································（1）
1.2 研究現狀···························································································（1）
1.3 研究思路···························································································（8）
第2 章 運動規劃中的智能體多層次行為模型框架設計·········································.（11）
2.1 環境建模方法··················································································.（11）
2.1.1 三角剖分法···········································································.（11）
2.1.2 可視圖方法···········································································.（12）
2.1.3 Voronoi 圖法··········································································.（12）
2.1.4 隨機采樣法···········································································.（13）
2.2 運動規劃方法··················································································.（16）
2.2.1 全局運動規劃········································································.（16）
2.2.2 局部運動規劃········································································.（20）
2.3 智能體多層次行為模型設計································································.（23）
2.3.1 智能體介紹···········································································.（23）
2.3.2 現有智能體行為模型框架·························································.（24）
2.3.3 智能體多層次行為模型框架······················································.（27）
第3 章 基于多信息域、多分辨率場景描述模型的多層次運動規劃算法研究··············.（31）
3.1 相關工作························································································.（31）
3.2 層次化運動規劃算法總體思路·····························································.（34）
3.2.1 問題描述··············································································.（34）
3.2.2 總體思路··············································································.（35）
3.3 全局層次上的多樣化引導路徑生成·······················································.（36）
3.3.1 高層時空約束的表示·······························································.（36）
3.3.2 全局概率路徑圖的生成····························································.（38）
3.3.3 全局多樣化引導路徑的生成······················································.（39）
3.4 局部層次上的多樣化運動路徑生成·······················································.（41）
3.4.1 局部高分辨率概率路徑圖的生成················································.（41）
3.4.2 局部多樣化運動路徑的生成······················································.（42）
3.5 動態障礙情況下的運動路徑重新規劃問題··············································.（43）
3.5.1 運動路徑的重新規劃·······························································.（43）
3.5.2 概率路徑圖的動態更新····························································.（44）
3.6 仿真實驗與結果分析·········································································.（45）
第4 章 多智能體避碰行為研究·······································································.（48）
4.1 相關工作························································································.（48）
4.2 避碰行為的基本概念與問題描述··························································.（50）
4.2.1 基本概念··············································································.（50）
4.2.2 問題描述··············································································.（51）
4.3 反應式避碰行為模型·········································································.（52）
4.4 基于最小代價原則的預測式避碰行為模型··············································.（57）
4.4.1 最小代價原則········································································.（57）
4.4.2 基于最小代價原則的預測式避碰行為建�！ぁぁぁぁぁぁぁぁぁぁぁぁぁぁぁぁぁぁぁぁぁぁぁぁぁぁぁぁぁぁぁぁぁぁぁ�.（58）
4.5 仿真實驗與結果分析·········································································.（63）
第5 章 基于增廣物理仿真的運動規劃多任務約束建模分析與優化求解····················.（67）
5.1 相關工作························································································.（67）
5.2 問題描述························································································.（70）
5.3 控制圍欄函數基礎背景知識································································.（70）
5.4 基于控制圍欄函數的多任務約束統一描述與運動路徑動力學物理仿真優化生成···.（71）
5.4.1 基于控制圍欄函數的多任務約束統一描述····································.（72）
5.4.2 增廣物理仿真框架下的智能體運動路徑優化生成···························.（74）
5.5 仿真實驗與結果分析·········································································.（77）
5.5.1 可生成多樣化行為的智能體運動路徑優化計算······························.（77）
5.5.2 自動駕駛中自適應巡航運動路徑優化計算····································.（81）
5.6 多智能體運動規劃仿真演示軟件設計開發··············································.（83）
5.6.1 演示軟件設計開發··································································.（83）
5.6.2 多智能體運動路徑生成測試······················································.（85）
第6 章 多任務約束時空融合處理機制研究························································.（87）
6.1 相關工作························································································.（87）
6.2 具有速度和加速度約束的多智能體時間最優軌跡規劃·······························.（88）
6.2.1 問題描述··············································································.（88）
6.2.2 算法思路及流程·····································································.（92）
6.2.3 可達集與可達速度區間的計算···················································.（95）
6.2.4 利用VIP 算法計算時間最優軌跡················································.（97）
6.2.5 仿真實驗與結果分析·······························································.（99）
6.3 基于控制圍欄函數描述的多任務約束時空融合處理··································（107）
6.3.1 同時段下多任務約束時序沖突的自動優化處理······························（107）
6.3.2 仿真實驗與結果分析·······························································（111）
第7 章 滿足時空約束要求的多智能體運動規劃在城市維穩處突行動仿真場景下的應用····.（115）
7.1 應用場景描述··················································································（115）
7.2 可視化仿真·····················································································（117）
7.2.1 仿真參數設置········································································（117）
7.2.2 仿真結果演示········································································（118）
第8 章 深度強化學習在運動規劃中的應用探索研究············································（124）
8.1 相關工作························································································（124）
8.2 深度強化學習算法探索效率提升策略研究··············································（126）
8.2.1 基本思路··············································································（127）
8.2.2 遺傳算法和DDPG 算法相結合的DRL 算法探索效率提升策略··········（129）
8.2.3 仿真實驗與結果分析·······························································（132）
8.3 基于深度強化學習的多智能體避碰行為生成···········································（134）
8.3.1 問題描述與基本思路·······························································（134）
8.3.2 仿真實驗與結果分析·······························································（135）
8.4 下一步展望·····················································································（136）
參考文獻······································································································（137）
新書資訊