·929· 伍明，等：纯方位角目标跟踪及移动平台可观性控制方法 第5期 [4]伍明，李琳琳，孙继银.基于概率数据关联交互多模滤 -2.5 1.5 波的移动机器人未知环境下动态目标跟踪).机器人， -2.0 1.0 0.5 2012,346):668-679 2.5 -3.0 WU Ming,LI Linlin,SUN Jiyin.PDA-IMM based mov- -3.5 -1.0 -1.5 ing object tracking with mobile robots in unknown envir- 0.500.51.01.52.02.5 -2.0-1.001.02.0 onments[J】.Robot,.2012,34(6):668-679. /m m [5] MUNARO M.LEWIS C.CHAMBERS D,et al.RGB-D (b)机器人跟踪结果 (c)目标跟踪结果 human detection and tracking for industrial environ- 图20实体机器人跟踪结果 ments[M]//Intelligent Autonomous Systems 13.Cham: Fig.20 Physical robot tracking results Springer International Publishing,2015:1655-1668. 由图20(b)可见，由于机器人和目标在相同平 [6] ZHANG Rui.WANG Zhaokui,ZHANG Yulin.Astro- 面运动，为了产生观测视差，机器人形成曲线运 naut visual tracking of flying assistant robot in space sta- 动轨迹，同时估计轨迹和实际轨迹一致。目标估 tion based on deep learning and probabilistic model[J]. 计轨迹和实际轨迹开始一致性较差，之后逐步趋 International journal of aerospace engineering,2018, 2018:1-17. 于一致，其原因在于初始阶段机器人以追随为主 [7]CASTLER O.MURRAYDW.Obiect recognition and 并未产生足够视差，当机器人接近目标后开始曲 localization while tracking and mapping[C]//Proceed- 线运动，足够的视差使目标跟踪精度得以提高。 ings of 8th IEEE International Symposium on Mixed and- 7结论 Augmented Reality.Orlando,IEEE,2009:179-180. [8] KLEIN G,MURRAY D.Parallel tracking and mapping 针对未知环境下基于单目视觉的移动机器人 for small AR workspaces[Cl//Proceedings of6th IEEE 目标跟踪问题展开研究，提出了一种纯方位角观 and ACM International Symposium on Mixed and Aug- 测条件下的机器人同时定位、地图构建与目标跟 mented Reality.Nara,IEEE,2007:225-234. 踪及控制方法。该方法在机器人单目视觉SLAM [9] DE OLIVEIRA D CB.DE SOUZA DA SILVA R L 基础上，设计了扩展式卡尔曼滤波框架的目标跟 Moving object tracking for SLAM-based augmented real- 踪算法，并将目标状态估计和机器人运动控制作 ity[J].Journalof mobile multimedia,2021,17(4): 577-602. 为耦合问题进行处理，提出了基于目标协方差阵 更新最大化的机器人优化控制方法。仿真和实体 [10]DAI W.ZHANG Y,ZHENG Y,et al.RGB-D SLAM 实验验证了方法的有效性和准确性，通过对实验 with moving object tracking in dynamic environments [J].IET cyber-systems and robotics,2021,3(4):11. 数据的分析可知：目标跟踪准确性与机器人定位 [11]LI Guihai,CHEN Songlin.Visual slam in dynamic scenes 精度以及机器人机动能力大小成正相关性。接下 based on object tracking and static points detection[J]. 来将针对障碍物条件下的估计和控制方法展开进 Journal of intelligent robotic systems,2022,104(2): 一步研究。 1-10 参考文献： [12] CIVERA J,DAVISON A J,MONTIEL J MM.Inverse depth parametrization for monocular SLAM[J].IEEE [1]蔡自兴，邹小兵.移动机器人环境认知理论与技术的研 transactions on robotics,2008,24(5):932-945. 究)机器人，2004,26(1)：87-91. [13]DISSANAYAKE M W MG,NEWMAN P,CLARK S. CAI Zixing,ZOU Xiaobing.Research on environmental et al.A solution to the simultaneous localization and map cognition theory and methodology for mobile robots[J]. building (SLAM)problem[J].IEEE transactions on robot- Robot,.2004.261:87-91 ics and automation,2001,17(3):229-241. [2] YANG Nan,WANG Rui,GAO Xiang,et al.Challenges [14]WANGSIRIPITAK S,MURRAY D W.Avoiding mov- in monocular visual odometry:photometric calibration, ing outliers in visual SLAM by tracking moving motion bias,and rolling shutter effect[J].IEEE robotics objects[C]//Proceedings of IEEE International Confer- and automation letters,2018,3(4):2878-2885. ence on Robotics and Automation.Kobe,IEEE,2009: [3]WANG C C,THORPE C,THRUN S,et al.Simultan- 375-380 eous localization,mapping and moving object tracking[J]. [15]MIGLIORE D,RIGAMONTI R.MARZORATI D,et al. The international journal of robotics research,2007, Use a single camera for simultaneous localization and 26(9):889-916. mapping with mobile object tracking in dynamic environ-(a) 总体跟踪结果 机器人 终点 机器人 终点 目标 终点 目标 终点 环境特 征估计 目标和机器 人实际轨迹 目标和机器 人估计轨迹 环境特征 估计 Y/m X/m −1 −2 −3 −4 −6 −4 −2 (b) 机器人跟踪结果 (c) 目标跟踪结果 −2.0 −2.5 1.5 1.0 0.5 0 −0.5 −1.0 −1.5 −2.5 −3.0 −3.5 −0.5 −2.0 −1.0 0 0.5 0 1.0 2.0 Y/m Y/m X/m X/m 1.0 1.5 2.0 2.5 图 20 实体机器人跟踪结果 Fig. 20 Physical robot tracking results 由图 20(b) 可见，由于机器人和目标在相同平 面运动，为了产生观测视差，机器人形成曲线运 动轨迹，同时估计轨迹和实际轨迹一致。目标估 计轨迹和实际轨迹开始一致性较差，之后逐步趋 于一致，其原因在于初始阶段机器人以追随为主 并未产生足够视差，当机器人接近目标后开始曲 线运动，足够的视差使目标跟踪精度得以提高。 7 结论 针对未知环境下基于单目视觉的移动机器人 目标跟踪问题展开研究，提出了一种纯方位角观 测条件下的机器人同时定位、地图构建与目标跟 踪及控制方法。该方法在机器人单目视觉 SLAM 基础上，设计了扩展式卡尔曼滤波框架的目标跟 踪算法，并将目标状态估计和机器人运动控制作 为耦合问题进行处理，提出了基于目标协方差阵 更新最大化的机器人优化控制方法。仿真和实体 实验验证了方法的有效性和准确性，通过对实验 数据的分析可知：目标跟踪准确性与机器人定位 精度以及机器人机动能力大小成正相关性。接下 来将针对障碍物条件下的估计和控制方法展开进 一步研究。 参考文献： 蔡自兴, 邹小兵. 移动机器人环境认知理论与技术的研 究 [J]. 机器人, 2004, 26(1): 87–91. CAI Zixing, ZOU Xiaobing. Research on environmental cognition theory and methodology for mobile robots[J]. Robot, 2004, 26(1): 87–91. [1] YANG Nan, WANG Rui, GAO Xiang, et al. Challenges in monocular visual odometry: photometric calibration, motion bias, and rolling shutter effect[J]. IEEE robotics and automation letters, 2018, 3(4): 2878–2885. [2] WANG C C, THORPE C, THRUN S, et al. Simultan￾eous localization, mapping and moving object tracking[J]. The international journal of robotics research, 2007, 26(9): 889–916. [3] 伍明, 李琳琳, 孙继银. 基于概率数据关联交互多模滤 波的移动机器人未知环境下动态目标跟踪 [J]. 机器人, 2012, 34(6): 668–679. WU Ming, LI Linlin, SUN Jiyin. PDA-IMM based mov￾ing object tracking with mobile robots in unknown envir￾onments[J]. Robot, 2012, 34(6): 668–679. [4] MUNARO M, LEWIS C, CHAMBERS D, et al. RGB-D human detection and tracking for industrial environ￾ments[M]//Intelligent Autonomous Systems 13. Cham: Springer International Publishing, 2015: 1655−1668. [5] ZHANG Rui, WANG Zhaokui, ZHANG Yulin. Astro￾naut visual tracking of flying assistant robot in space sta￾tion based on deep learning and probabilistic model[J]. International journal of aerospace engineering, 2018, 2018: 1–17. [6] CASTLER O, MURRAYDW. Object recognition and localization while tracking and mapping[C]// Proceed￾ings of 8th IEEE International Symposium on Mixed and￾Augmented Reality. Orlando, IEEE, 2009: 179−180. [7] KLEIN G, MURRAY D. Parallel tracking and mapping for small AR workspaces[C]// Proceedings of6th IEEE and ACM International Symposium on Mixed and Aug￾mented Reality. Nara, IEEE, 2007: 225−234. [8] DE OLIVEIRA D CB, DE SOUZA DA SILVA R L. Moving object tracking for SLAM-based augmented real￾ity[J]. Journalof mobile multimedia, 2021, 17(4): 577–602. [9] DAI W, ZHANG Y, ZHENG Y, et al. RGB-D SLAM with moving object tracking in dynamic environments [J]. IET cyber-systems and robotics, 2021, 3(4): 11. [10] LI Guihai, CHEN Songlin. Visual slam in dynamic scenes based on object tracking and static points detection[J]. Journal of intelligent & robotic systems, 2022, 104(2): 1–10. [11] CIVERA J, DAVISON A J, MONTIEL J M M. Inverse depth parametrization for monocular SLAM[J]. IEEE transactions on robotics, 2008, 24(5): 932–945. [12] DISSANAYAKE M W M G, NEWMAN P, CLARK S, et al. A solution to the simultaneous localization and map building (SLAM) problem[J]. IEEE transactions on robot￾ics and automation, 2001, 17(3): 229–241. [13] WANGSIRIPITAK S, MURRAY D W. Avoiding mov￾ing outliers in visual SLAM by tracking moving objects[C]// Proceedings of IEEE International Confer￾ence on Robotics and Automation. Kobe, IEEE, 2009: 375−380. [14] MIGLIORE D, RIGAMONTI R, MARZORATI D, et al. Use a single camera for simultaneous localization and mapping with mobile object tracking in dynamic environ- [15] ·929· 伍明，等：纯方位角目标跟踪及移动平台可观性控制方法 第 5 期