针对基本的快速拓展随机树算法(rapidly-exploring random tree,RRT^(*))存在搜索随机性大、效率低、路径非最优的缺点,提出一种引入人工势场法算法(artificial potential field method,APF)和Douglas-Peucker算法的改进RRT^(*)-APF-DP...针对基本的快速拓展随机树算法(rapidly-exploring random tree,RRT^(*))存在搜索随机性大、效率低、路径非最优的缺点,提出一种引入人工势场法算法(artificial potential field method,APF)和Douglas-Peucker算法的改进RRT^(*)-APF-DP路径规划算法.在RRT*算法的采样点生成阶段引入变采样范围偏置搜索与步长自适应调整策略,融合重新设计的APF算法的引力与斥力函数,增强路径扩展导向性与绕过障碍物能力.采用重采样策略改进DP算法,优化避障代价与控制点数量.实验结果表明,本算法规划的避障路径满足机械臂的运动要求,且算法规划的避障路径代价、规划时间和路径控制节点数均得到有效改善.展开更多
针对传统快速扩展随机树(rapidly-exploring random tree star,RRT^(*))算法在全局路径规划过程中存在收敛速度慢、搜索路径不平滑、内存占用多等问题,提出了一种RRT^(*)与人工势场法(artificial potential field,APF)的融合搜索算法。...针对传统快速扩展随机树(rapidly-exploring random tree star,RRT^(*))算法在全局路径规划过程中存在收敛速度慢、搜索路径不平滑、内存占用多等问题,提出了一种RRT^(*)与人工势场法(artificial potential field,APF)的融合搜索算法。为了加快RRT^(*)算法在搜索过程中的收敛速度,在算法中利用人工势场法的思想引导扩展随机树快速向目标点生长;对融合算法在空间中的采样范围做出改进,使算法在APF产生的合力特定范围内进行采样,提高算法在空间中的搜索效率,减少无用节点的扩展。仿真结果表明:相比传统的RRT和RRT^(*)算法以及APF-RRT融合算法,APF-RRT^(*)融合算法能够规划出更短、更平滑的路径,路径距离缩短了1.5%~10.83%;算法的搜索时间也显著缩短了1.97%~49.78%;与其他算法相比,APF-RRT^(*)融合算法的路径节点数量减少了4.66%~41.95%,路径平滑性也得到了提高。展开更多
为解决传统人工势场法(APF)进行水面无人艇路径规划时出现局部最小和路径不平滑的问题,对斥力系数进行动态调整并引入逃逸力。为使水面无人艇在动态避碰中按国际海上避碰规则(international regulations for preventing collisions at s...为解决传统人工势场法(APF)进行水面无人艇路径规划时出现局部最小和路径不平滑的问题,对斥力系数进行动态调整并引入逃逸力。为使水面无人艇在动态避碰中按国际海上避碰规则(international regulations for preventing collisions at sea,COLREGS)航行,结合无人艇性能制定了无人艇的避碰规则,并引入转向力。通过Matlab对静态未知障碍物的仿真实验,验证改进APF算法能解决局部最小问题且规划的路径更加安全平滑。通过对动态未知障碍物4种典型局面的仿真实验表明,改进APF算法在复杂环境下也能引导无人艇在符合避碰规则的前提下对多个动态障碍进行安全规避。展开更多
Formation keeping is important for multiple Unmanned Aerial Vehicles(multi-UAV)to fully play their roles in cooperative combats and improve their mission success rate.However,in practical applications,it is difficult ...Formation keeping is important for multiple Unmanned Aerial Vehicles(multi-UAV)to fully play their roles in cooperative combats and improve their mission success rate.However,in practical applications,it is difficult to achieve formation keeping precisely and obstacle avoidance autonomously at the same time.This paper proposes a joint control method based on robust H∞ controller and improved Artificial Potential Field(APF)method.Firstly,we build a formation flight model based on the “Leader-Follower”structure and design a robust H∞ controller with three channels X,Y and Z to eliminate dynamic uncertainties,so as to realize high-precision formation keeping.Secondly,to fulfill obstacle avoidance efficiently in complex situations where UAVs fly at high speed with high inertia,this paper comes up with the improved APF method with deformation factor considered.The judgment criterion is proposed and applied to ensure flight safety.In the end,the simulation results show that the designed controller is effective with the formation keeping a high accuracy and in the meantime,it enables UAVs to avoid obstacles autonomously and recover the formation rapidly when coming close to obstacles.Therefore,the method proposed here boasts good engineering application prospect.展开更多
针对双方无人机之间的动态对抗博弈问题,提出了动态人工势场法(dynamic artificial potential field method,DAPF)和精英蚁群优化(elite ant colony optimization,EACO)算法相结合的求解方法。首先,采用动态人工势场法,以敌我双方无人...针对双方无人机之间的动态对抗博弈问题,提出了动态人工势场法(dynamic artificial potential field method,DAPF)和精英蚁群优化(elite ant colony optimization,EACO)算法相结合的求解方法。首先,采用动态人工势场法,以敌我双方无人机作为博弈的局中人,构建双方无人机动态对抗博弈模型。其次,提出精英蚁群算法,计算双方博弈的纳什均衡策略。该算法引入对立学习和划分精英蚂蚁加快算法收敛速度,并引入遗传算法中的变异操作以避免局部最优值的问题。最后,仿真验证了所提方法的可行性和有效性。展开更多
文摘针对基本的快速拓展随机树算法(rapidly-exploring random tree,RRT^(*))存在搜索随机性大、效率低、路径非最优的缺点,提出一种引入人工势场法算法(artificial potential field method,APF)和Douglas-Peucker算法的改进RRT^(*)-APF-DP路径规划算法.在RRT*算法的采样点生成阶段引入变采样范围偏置搜索与步长自适应调整策略,融合重新设计的APF算法的引力与斥力函数,增强路径扩展导向性与绕过障碍物能力.采用重采样策略改进DP算法,优化避障代价与控制点数量.实验结果表明,本算法规划的避障路径满足机械臂的运动要求,且算法规划的避障路径代价、规划时间和路径控制节点数均得到有效改善.
文摘针对传统快速扩展随机树(rapidly-exploring random tree star,RRT^(*))算法在全局路径规划过程中存在收敛速度慢、搜索路径不平滑、内存占用多等问题,提出了一种RRT^(*)与人工势场法(artificial potential field,APF)的融合搜索算法。为了加快RRT^(*)算法在搜索过程中的收敛速度,在算法中利用人工势场法的思想引导扩展随机树快速向目标点生长;对融合算法在空间中的采样范围做出改进,使算法在APF产生的合力特定范围内进行采样,提高算法在空间中的搜索效率,减少无用节点的扩展。仿真结果表明:相比传统的RRT和RRT^(*)算法以及APF-RRT融合算法,APF-RRT^(*)融合算法能够规划出更短、更平滑的路径,路径距离缩短了1.5%~10.83%;算法的搜索时间也显著缩短了1.97%~49.78%;与其他算法相比,APF-RRT^(*)融合算法的路径节点数量减少了4.66%~41.95%,路径平滑性也得到了提高。
文摘为解决传统人工势场法(APF)进行水面无人艇路径规划时出现局部最小和路径不平滑的问题,对斥力系数进行动态调整并引入逃逸力。为使水面无人艇在动态避碰中按国际海上避碰规则(international regulations for preventing collisions at sea,COLREGS)航行,结合无人艇性能制定了无人艇的避碰规则,并引入转向力。通过Matlab对静态未知障碍物的仿真实验,验证改进APF算法能解决局部最小问题且规划的路径更加安全平滑。通过对动态未知障碍物4种典型局面的仿真实验表明,改进APF算法在复杂环境下也能引导无人艇在符合避碰规则的前提下对多个动态障碍进行安全规避。
基金supported by Funding from the National Key Laboratory of Rotorcraft Aeromechanics,China(No.61422202108)the National Natural Science Foundation of China(No.52176009).
文摘Formation keeping is important for multiple Unmanned Aerial Vehicles(multi-UAV)to fully play their roles in cooperative combats and improve their mission success rate.However,in practical applications,it is difficult to achieve formation keeping precisely and obstacle avoidance autonomously at the same time.This paper proposes a joint control method based on robust H∞ controller and improved Artificial Potential Field(APF)method.Firstly,we build a formation flight model based on the “Leader-Follower”structure and design a robust H∞ controller with three channels X,Y and Z to eliminate dynamic uncertainties,so as to realize high-precision formation keeping.Secondly,to fulfill obstacle avoidance efficiently in complex situations where UAVs fly at high speed with high inertia,this paper comes up with the improved APF method with deformation factor considered.The judgment criterion is proposed and applied to ensure flight safety.In the end,the simulation results show that the designed controller is effective with the formation keeping a high accuracy and in the meantime,it enables UAVs to avoid obstacles autonomously and recover the formation rapidly when coming close to obstacles.Therefore,the method proposed here boasts good engineering application prospect.
文摘针对双方无人机之间的动态对抗博弈问题,提出了动态人工势场法(dynamic artificial potential field method,DAPF)和精英蚁群优化(elite ant colony optimization,EACO)算法相结合的求解方法。首先,采用动态人工势场法,以敌我双方无人机作为博弈的局中人,构建双方无人机动态对抗博弈模型。其次,提出精英蚁群算法,计算双方博弈的纳什均衡策略。该算法引入对立学习和划分精英蚂蚁加快算法收敛速度,并引入遗传算法中的变异操作以避免局部最优值的问题。最后,仿真验证了所提方法的可行性和有效性。
