异构低能耗的跟随机器人系统设计

doi:10.16180/j.cnki.issn1007-7820.2022.10.002

摘要/Abstract

摘要：

服务机器人在单任务模式下适用领域较小,智能化程度较低。而在多任务情况下,服务机器人的算力成本较高,导致能耗大幅提升,缩短了续航时间。针对上述问题,文中设计了一种基于ROS平台的低功耗、多任务跟随机器人系统,并使该系统能够在定点巡航和视觉跟随两种任务下实现智能切换。在硬件方面,通过构建低能耗的异构计算平台来保证在输出同等算力的同时降低能耗。在软件方面,根据异构计算平台的性质,设计了跟随机器人系统的算法框架。机器人系统中各模块状态切换采用基于马尔可夫链的状态转移控制方法来实现。实验结果表明,文中提出的机器人系统能够降低能耗75%,在功能上能够实现定点巡航与视觉跟随的实时切换。

关键词: 低能耗, 多任务, 定点巡航, 视觉跟随, 马尔可夫链, 状态转移, 系统设计, ROS

Abstract:

In the single-task mode, the service robot has a small application field and a low degree of intelligence. In the case of multitasking, the computing power cost of service robots is high, which leads to a significant increase in energy consumption and shortens the battery life.In view of the above problems, a low-power, multitask follow-up robot system based on the ROS platform is designed in this study,and the system can realize intelligent switching under two tasks of fixed-point cruise and visual follow-up.In terms of hardware, a low-energy heterogeneous computing platform is established to ensure that the same computing power is output while reducing energy consumption. In terms of software, according to the nature of the heterogeneous computing platform, the algorithm framework of the follow-up robot system is designed. The robot system uses the state transition control method based on Markov chain to realize the state switching of each module.The experimental results show that the proposed robot system can reduce energy consumption by 75%, and can realize the real-time switching between fixed-point cruise and target following in function.

Key words: low energy consumption, multitask, fixed-point cruise, visual following, Markov chain, state transition, system design, ROS

中图分类号:

TN256

汪庆海,杨芳艳. 异构低能耗的跟随机器人系统设计[J]. 电子科技, 2022, 35(10): 8-14.

WANG Qinghai,YANG Fangyan. Design of Heterogeneous and Low-Power Following Robot System[J]. Electronic Science and Technology, 2022, 35(10): 8-14.

图/表 17

图1

图2

图3

图4

图5

图6

表1

表2

图7

表3

图8

图9

表4

表5

图10

图11

图12

参考文献 19

[1]	代兴冉. 自主跟随机器人的研究与设计[D]. 曲阜: 曲阜师范大学, 2020.
	Dai Xingran. Research and design of autonomous following robot[D]. Qufu: Qufu Normal University, 2020.
[2]	Calik S K. UAV path planning with multiagent Ant Colony system approach[C]. Zonguldak: Proceedings of the Twenty-fourth Signal Processing and Communication Application Conference, 2016.
[3]	Hart P E, Nilsson N J, Raphael B. A formal basis for the heuristic determination of minimum cost paths[J]. IEEE Transactions on Systems Science and Cybernetics, 1968, 4(2):100-107.
[4]	Rösmann C, Hoffmann F, Bertram T. Integrated online trajectory planning and optimization in distinctive topologies[J]. Robotics and Autonomous Systems, 2017, 7(4):142-153.
[5]	刘永建, 曾国辉, 黄勃, 等. 改进蚁群算法的机器人路径规划研究[J]. 电子科技, 2020, 33(1):13-18.
	Liu Yongjian, Zeng Guohui, Huang Bo, et al. Research on robot path planning based on improved ant colony algorithm[J]. Electronic Science and Technology, 2020, 33(1):13-18.
[6]	Feyrer S, Zell A. Detection, tracking, and pursuit of humans with an autonomous mobile robot[C]. Kyongju: Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems, 1999.
[7]	Hirai N, Mizoguchi H. Visual tracking of human back and shoulder for person following robot[C]. Kobe: Proceedings of the IEEE/ASM E International Conference on Advanced Intelligent Mechatronics, 2003.
[8]	Hassan M S, Khan A F, Khan M W, et al. A computationally lowcost vision based tracking algorithm for human following robot[C]. Hong Kong: Proceedings of the Second International Conference on Control, Automation and Robotics, 2016.
[9]	余铎, 王耀南, 毛建旭, 等. 基于视觉的移动机器人目标跟踪方法[J]. 仪器仪表学报, 2019, 40(1):227-235.
	Yu Duo, Wang Yaonan, Mao Jianxu, et al. Vision-based object tracking method of mobile robot[J]. Chinese Journal of Scientific Instrument, 2019, 40(1):227-235.
[10]	李建伏, 王思博, 宋国平. 基于出行偏好和路径长度的路径规划方法[J]. 吉林大学学报(理学版), 2021, 59(1):107-114.
	Li Jianfu, Wang Sibo, Song Guoping. Path planning method based on routing preference and path length[J]. Journal of Jilin University(Science Edition), 2021, 59(1):107-114.
[11]	吴国威. 智能除草机器人的视觉导航方法研究[D]. 哈尔滨: 哈尔滨工业大学, 2020.
	Wu Guowei. The visual navigation method of intelligent weeding robot[D]. Harbin: Harbin Institute of Technology, 2020.
[12]	唐小煜, 黄进波, 冯洁文, 等. 基于U-net和YOLOv4的绝缘子图像分割与缺陷检测[J]. 华南师范大学学报(自然科学版), 2020, 52(6):15-21.
	Tang Xiaoyu, Huang Jinbo, Feng Jiewen, et al. Image segmentation and defect detection of insulators based on U-net and YOLOv4[J]. Journal of South China Normal University (Natural Science Edition), 2020, 52(6):15-21.
[13]	Wang C Y, Liao H Y M, Wu Y H, et al. CSPNet: A new backbone that can enhance learning capability of CNN[C]. Seattle: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2020.
[14]	He K, Zhang X, Ren S. et al. Spatial pyramid pooling in deep convolutional networks for visual recognition[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2015, 37(9):1904-1916. doi: 10.1109/TPAMI.2015.2389824 pmid: 26353135
[15]	Liu S, Qi L, Qin H F, et al. Path aggregation network for instance segmentation[C]. Salt Lake City: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018.
[16]	黎阳, 沈烨, 刘敏, 等. 融合运动信息与表观信息的多目标跟踪算法[J]. 电子科技, 2020, 33(9):21-24.
	Li Yang, Shen Ye, Liu Min, et al. Multi-target tracking algorithm by combining motion information and apparent information[J]. Electronic Science and Technology, 2020, 33(9):21-24.
[17]	Dalal N, Triggs B, Schmid C. Human detection using oriented histograms of flow and appearance[C]. Berlin: Proceedings of the European Conference on Computer Vision, 2006.
[18]	Luo H, Gu Y, Liao X, et al. Bag of tricks and a strong baseline for deep person reidentification[C]. Long Beach: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2019.
[19]	Pan X, Luo P, Shi J, et al. Two at once: Enhancing learning and generalization capacities via ibn-net[C]. Munich: Proceedings of the European Conference on Computer Vision, 2018.

模型	AP/%
YOLOv4-coco	0.716 0
YOLOv4-our	0.893 1
yolov4	0.856 1

模型	mAP/%	Rank-1/%
ResNet50	49.8	66.2
ResNet50 IBN-a	65.9	75.4

硬件	型号
机器人底盘	Scout mini
计算平台	IntelNUC+Nvidia Xavier
激光雷达	Velodyne16
相机	Realsense D435
交换机	TP_Link工业交换机
路由器	华为4G路由器
电池	48 V电池供电

计算平台	CPU	GPU	CUDA	cuDNN
Nvidia Xavier	8-Core ARM v8.2 64-Bit CPU,8 MB L2+4 MB L3	512-Core Volta GPU with Tensor Cores	10.0	7.5
Intel NUC	i5-8260U	无	无	无
Dell Precision 7530	i7-8850H	Quadro P1000	9.0	7.6.5

计算平台	电压/V	电流/A	功率/W	CPU使用情况/%	GPU使用情况/%
Nvidia Xavier	18.99	0.82	15.57	54.00	99
Intel NUC	19.00	1.13	21.47	50.53	无
Dell Precision 7530	223.60	0.69	154.00	69.26	97