摘要
针对建筑机器人在施工现场获取地图信息时间长且需要规划出一条全局的、能实时避障的路径等问题,该文提出了一种应用建筑信息模型(building information model,BIM)技术建立导航地图并进行路径规划的算法。根据BIM模型中的信息对传统RRT算法进行优化改进,提出了IRRT(improved rapid-exploration random tree)算法。首先将原有的固定步长改为动态步长,通过判断与目标点的远近界定步长大小,避免了节点的盲目扩张;其次,对随机采样点的生成范围进行了约束,并设置一个同时考虑目标点和随机点的权重来解决传统RRT算法中新生成点仅由随机采样点单一决定的问题;算法陷入最小值时选取随机扰动策略进行逃脱;最后在全局路径的相邻节点间使用动态窗口法进行局部避障。实验仿真结果表明IRRT算法比传统RRT算法在搜索速度上快了3倍多,平均路径比改进前减少25.56%,平均节点减少8.92%,加入动态窗口法后有效提高了机器人实时避障能力,更适合多变的室内环境使用。
[Objective]A building robot requires a long duration to acquire map information at the construction site and needs to plan a global path that can avoid obstacles in real time.To address this issue,this paper proposes an algorithm that applies building information model(BIM)technology to establish a navigation map and perform path planning.It can quickly and comprehensively obtain the map information of the construction site and provide more accurate basic data for path planning.Then,global path planning and local path adjustment are comprehensively considered so that the construction robot can quickly search the global path,flexibly respond to changes in obstacles on the construction site and adjust the path in real time.The information in the BIM model is extracted and mapped to a two-dimensional raster map to build a map.Following completion,the improved rapidly exploring random tree(IRRT)algorithm is proposed by optimizing the traditional RRT algorithm in the search path stage.[Methods]First,the original fixed step size is changed to a dynamic step size,which is defined by judging the distance between the node and the target point.When the node is far from the target point,a larger step size is adopted to expedite the search speed.The step size gradually reduces as the node approaches the goal,improving search accuracy and preventing the node from blindly expanding.Second,the generation range of random sampling points is constrained.Instead of randomly generating sampling points in the entire space,the generation range is set according to the position of the starting and the target points so that the sampling points are more targeted and more likely to fall in the effective area of path search,thereby reducing the search space size.A weight considering both target and random points is set to solve the problem that new points are only determined by random sampling points in the traditional RRT algorithm.When the algorithm is trapped in the minimum value,a random perturbation strategy is selected to escape,which
作者
杨振舰
庄亚楠
陈亚东
YANG Zhenjian;ZHUANG Yanan;Chen Yadong(School of Computer and Information Engineering,Tianjin Chengjian University,Tianjin 300384,China)
出处
《实验技术与管理》
CAS
北大核心
2024年第2期31-42,共12页
Experimental Technology and Management
基金
天津市自然科学基金(19JCTPJC47700)。
关键词
建筑机器人
建筑信息模型(BIM)
路径规划
IRRT算法
动态窗口法
实时避障
construction robot
building information model(BIM)
route planning
IRRT algorithm
dynamic window approach
real-time obstacle avoidance
