摘要
针对建筑机器人进入施工位置全局路径规划最优且可实时避障的要求,提出了一种新型的导航地图建立方法,通过建筑物BIM(Building Information Modeling)模型建立导航地图,通过优化A*算法搜索点选取策略,以及删除了路径上的冗余转折点,缩短算法运行时间,且规划出的全局路径不会紧贴着建筑物墙壁,有效降低了机器人与墙体发生碰撞的可能性。结合动态窗口算法进行局部路径规划,在全局路径关键点之间使用动态窗口法,且加入了新的刹车判定条件,使得机器人的运动连续化。实验结果表明改进后的A*算法运行时间相较于原始算法减少50%以上,每去除一个冗余转折点路径长度减少0.4 m,路径与障碍物之间的距离比原始算法增加了2倍,结合了动态窗口法之后能够较好地避开障碍且输出的控制参数连续化。
Aiming at the requirements of the optimal global path planning for building robots entering the construction site and real-time obstacle avoidance,a new navigation map construction method is proposed.The navigation map is established through the building BIM model,and the search point selection strategy is optimized by optimizing the A*algorithm.The redundant turning point on the path is deleted,the running time of the algorithm is shortened,and the planned global path does not closely follow the building wall,which effectively reduces the possibility of collision between the robot and the wall.Combined with the dynamic window algorithm for local path planning,the dynamic window method is used between the global path key points,and new brake determination conditions are added to make the robot’s motion continuous.The experimental results show that the improved A*algorithm running time is reduced by more than 50%compared with the original algorithm.The path length of each redundant turning point is reduced by 0.4 m,and the distance between the path and the obstacle is increased by 2 times compared with the original algorithm.After the dynamic window method,the obstacle can be better avoided and the output control parameters are continuous.
作者
王凡
李铁军
刘今越
赵海文
WANG Fan;LI Tiejun;LIU Jinyue;ZHAO Haiwen(School of Mechanical Engineering,Hebei University of Technology,Tianjin 300131,China)
出处
《计算机工程与应用》
CSCD
北大核心
2020年第17期224-230,共7页
Computer Engineering and Applications
基金
国家自然科学基金(No.U1813222)
国家重点研发计划(No.2017YFB1301002)。
关键词
建筑机器人
建筑信息模型(BIM)
路径规划
A*算法
动态窗口法
避障
construction robot
Building Information Modeling(BIM)
path planning
A*algorithm
dynamic window approach
obstacle avoidance
作者简介
王凡(1995—),男,硕士研究生,研究领域为建筑机器人导航,E-mail:van_w2094@sina.com;李铁军(1967—),男,博士,教授,研究领域为机器人、人机协作技术;通信作者:刘今越(1977—),男,博士,教授,研究领域为机器人技术、智能检测与控制;赵海文(1974—),男,博士,副教授,研究领域为机电一体化、智能检测与控制。
