基于视觉辅助与自适应粒子滤波的机器人自定位  

Robot Self-localization Based on Visual-aided and Adaptive Particle Filter

作　　者：刘超凡 于文涛[1] 曲枫 张祺 徐浩 LIU Chaofan;YU Wentao;QU Feng;ZHANG Qi;XU Hao(Department of Computer and Information Engineering,Central South University of Forestry and Technology,Changsha 410004,China)

机构地区：[1]中南林业科技大学计算机与信息工程学院,湖南长沙410004

出　　处：《郑州大学学报（理学版）》2023年第5期73-80,共8页Journal of Zhengzhou University:Natural Science Edition

基　　金：国家自然科学基金项目(61602529);智慧物流技术湖南省重点实验室项目(2019TP1015)。

摘　　要：为解决机器人自定位粒子退化问题,提出一种基于视觉辅助与自适应粒子滤波的机器人自定位方法。首先采用改进萤火虫优化算法调整基于双里程运动模型的建议分布,然后使用融合视觉标签验证信息的子群互助重采样算法,在缓解粒子退化的同时提高粒子多样性。从位姿跟踪误差和粒子多样性两个方面对算法进行评价,结果表明,所提算法在仿真和真实环境下都具有较高的定位精度和稳定性。A robot self-localization method based on visual-aided and adaptive particle filter was proposed to solve the problem of particle degradation in robot self-localization.Firstly,an improved firefly optimization algorithm was adopted to adjust the proposed distribution based on the dual-odometry motion model.Then,a subgroup mutual aid resampling algorithm incorporating visual tags verification information was used to alleviate the particle degeneracy and improve particle diversity.The algorithm was evaluated from the pose tracking error and particle diversity,and the results showed that the proposed algorithm had a high localization accuracy and stability in both simulation and the real environment.

关 键 词：自适应粒子滤波 萤火虫算法 视觉辅助定位 互助重采样 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]