摘要
针对传统动态窗口法在行人密集环境下动态路径规划存在灵活性差、效率低、安全性缺乏等问题,提出一种社会交互空间下基于社会约束自适应动态窗口法(social_DWA),并采用其解决服务机器人局部路径规划问题。首先,采用非对称高斯公式对单行人以及多人群组交互空间进行模型化描述;其次,在传统动态窗口法的基础上,采用动态行人方位角约束对动态行人进行避让;改进距离评价函数,分类决策与行人、多人群组、一般障碍物的安全距离;最后,提出速度权重自适应调整策略,优化服务机器人在途经不同密集度社会交互区域时的移动速度。为验证算法有效性,在两种模拟社会场景下,先后开展了social_DWA算法与传统DWA算法、FIDWA算法的路径规划仿真对比实验。结果表明:采用social_DWA算法所消耗的运动时间在场景1中较传统DWA和FIDWA算法分别缩短了1.53、0.43 s,在场景2中较传统DWA和FIDWA算法分别缩短了26.3、2.86 s;相较于传统DWA算法和FIDWA算法,social_DWA算法能保持有效的行人安全距离,并使运行轨迹更加合理。social_DWA算法在行人避让、环境适应能力等方面具有一定的优越性。
To address issues such as poor flexibility,low efficiency,and safety concerns in dynamic path planning in the crowded pedestrian environments,an adaptive dynamic window method based on social constraints(social_DWA)in the social interaction space is proposed,and it is used to solve local path planning problems of service robots.Firstly,the asymmetric Gaussian formula is employed to model interaction spaces of single pedestrian and groups.Secondly,based on the original dynamic window method,the dynamic pedestrian azimuth constraint is implemented to steer clear of dynamic pedestrians.Enhancements are made to the distance evaluation function to classify and ascertain the safety distances from pedestrians,pedestrian groups,and general obstacles.Finally,an adaptive adjustment strategy of speed weight is proposed to optimize the movement speed of service robots while passing through social interaction areas with varying densities.To verify the effectiveness of the algorithm,the path planning simulation experiments of the social_DWA algorithm,the traditional DWA algorithm,and the FIDWA algorithm are carried out in two simulated social scenarios featuring different complexities.The results show that the motion time consumed by the social_DWA algorithm is reduced by 1.53 and 0.43 s compared with the traditional DWA algorithms and FIDWA algorithms in scenario 1,and 26.3 and 2.86 s lower than the traditional DWA algorithms and FIDWA algorithms in scenario 2,respectively.Compared with the traditional DWA algorithm and FIDWA algorithm,the social_DWA algorithm maintains an effective pedestrian safety distance and ensures more rational running trajectories.The validation confirms the social_DWA algorithm’s superiority in pedestrian avoidance and environmental adaptability.
作者
何丽
宁子豪
袁亮
刘志强
HE Li;NING Zihao;YUAN Liang;LIU Zhiqiang(School of Intelligent Manufacturing Modern Industry,Xinjiang University,Urumqi 830047,China;School of Information Science and Technology,Beijing University of Chemical Technology,Beijing 100029,China)
出处
《西安交通大学学报》
EI
CAS
CSCD
北大核心
2024年第5期42-51,共10页
Journal of Xi'an Jiaotong University
基金
国家自然科学基金资助项目(62063033)。
关键词
服务机器人
路径规划
动态窗口法
参数自适应
人性化
service robots
path planning
dynamic windows approach
self-adaptation
humanization
