基于RBPF-SLAM算法的学前儿童机器人路径规划模型研究  

Research on Path Planning Model of Preschool Children’s Robot Based on RBPF-SLAM Algorithm

作　　者：李洁[1] 曹艳龙 JIE Li;CAO Yanlong(Xianyang Normal University,Xianyang Shanxi 712000,China;Xidian University,Xi’an 710071,China)

机构地区：[1]咸阳师范学院,陕西咸阳712000 [2]西安电子科技大学,西安710071

出　　处：《自动化与仪器仪表》2024年第6期189-193,共5页Automation & Instrumentation

基　　金：陕西省教育厅专项科研计划项目(人文社科)《陕西省0~3岁早期教育师资培养现状及策略研究》(20JK0427)。

摘　　要：在学前教育领域,机器人技术的应用已经成为一种引人注目的研究方向。为了提升学前儿童机器人的路径寻优能力,此次研究采用基于粒子滤波的同时定位与地图创建算法将雷达观测模型和运动模型相结合,并采用退火参数和重采样避免采样效率降低,构建机器人运动模型和地图。最后采用改进的A-Star算法构建混合路径规划模型。结果显示,基于粒子滤波的改进SLAM算法误差平均值为0.340。改进后的基于粒子滤波的SLAM算法在粒子数为50、125、200时,运行时间分别为0.215 s、0.225 s、0.268 s,估计误差分别为0.314、0.282、0.291。基于粒子滤波的SLAM算法生成的地图整体精度明显高于改进前的算法,地图上的线条整齐且清晰。在仿真和真实两种实验环境下,改进A*算法分别经过18次和50次迭代后趋于收敛,最小路径长度为12 m和32 m。结果验证了混合路径规划模型在不同环境下路径寻优的性能较好。该方法实现了精确的机器人定位,提高了路径规划的准确性和可靠性。In the field of preschool education,the application of robotics technology has become a remarkable research direction.In order to enhance the path optimization ability of preschool children's robots,this study uses the simultaneous localization and mapping algorithm based on particle filtering to combine radar observation models with motion models,and using annealing parameters and resampling to avoid a decrease in sampling efficiency,a robot motion model and map are constructed.Finally,a hybrid path planning model was constructed using the improved A-Star algorithm and the time elastic band algorithm.The experimental results show that the average error of the improved particle filter based SLAM algorithm is 0.340.The improved SLAM algorithm based on particle filtering has run times of 0.215 s,0.225 s,and 0.268 s with particle numbers of 50,125,and 200,and estimation errors of 0.314,0.282,and 0.291,respectively.The overall accuracy of the map generated by the SLAM algorithm based on particle filtering is significantly higher than the improved algorithm,and the lines on the map are neat and clear.In both simulated and real experimental environments,the improved A∗algorithm converged after 18 and 50 iterations,respectively,with a minimum path length of 12 m and 32 m.The results verified that the hybrid path planning model performs well in path optimization in different environments.This method achieves precise robot positioning and improves the accuracy and reliability of path planning.

关 键 词：学前儿童机器人 路径规划 RBPF-SLAM A-STAR算法 精度 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]