摘要
果蔬收获属于一类劳动密集型工作,由于人口老龄化和农业劳动力资源的缺乏,致使人工收获成本在整个果蔬生产成本中所占的比例高达33~50%,大大降低了产品的市场竞争力。研究开发智能型采摘机器人不仅有利于解放劳动力、提高生产率、降低生产成本、保证新鲜果蔬品质,而且还可以促进机械结构、机器人技术、图像处理、智能控制技术、传感器技术等相关领域的深入研究和应用,具有重要的现实意义和战略意义。
本课题在全面分析研究国内外采摘机器人优缺点的基础上,以黄瓜为研究对象,从黄瓜采摘机器人本体结构设计、运动学与动力学分析、轨迹规划与轨迹跟踪、定位精度补偿及控制系统设计等方面入手,对黄瓜采摘机器人的运动规划与控制系统进行了全面而系统的研究。主要研究工作与成果如下:
1.针对黄瓜采摘作业的特点,研制开发了一套四自由度关节型采摘机械手,该机械手结构简单紧凑,重量轻,成本低,动作灵活平稳。为使采摘机械手能收获目标空间的所有果实,提出了一种适用于采摘空间为任意立方体的关节型机械手结构参数优化方法:根据黄瓜目标采摘空间的要求,利用MATLAB优化工具箱对采摘机械手进行了结构参数优化。
2.对采摘机械手进行运动学和动力学分析与仿真。采用D-H法建立采摘机械手正运动学,利用解析法进行逆运动学求解;并根据拉格朗日法建立采摘机械手的简化动力学模型。最后进行了运动学、工作空间及动力学仿真实验与分析,验证了模型的正确性及结构参数优化的合理性。
3.研究了机器人关节空间轨迹规划与轨迹跟踪控制算法。通过对各种轨迹规划算法的分析比较,结合采摘机器人系统自身结构特点与果实采摘作业要求,提出了采用两段摆线运动组合的水平抓取规划方法,实现机器人快速平稳运动。同时,为实现对期望轨迹的精确跟踪,构造了一种快速非奇异终端滑模控制器,解决了普通滑模控制器在线性滑模条件下渐进收敛、传统终端滑模控制的奇异性与抖振问题。仿真实验证明它能够准确跟踪期望轨迹,并能使位置跟踪误差在有限时间内收敛到平衡点,响应时间短,鲁棒性好。
4.对采摘机器人定位误差补偿算法进行了研究。提出了基于LMBP神经网络的直接误差补偿法和预置偏移量补偿法,用于减少由结构参数偏差引入的机器人末端位置误差,并通过仿真实验验证了方法的有效性。
5.基于CAN总线的黄瓜采摘机器人控制系统软硬件设计。提出了基于DSP的上位机运动控制器+CAN总线+ DSP关节控制器的分布式控制方案。搭建了控制系统的硬件平台;并采用模块化设计思想,设计了采摘机器人控制系统初始化、逆运动学运算、轨迹规划、末端执行器控制、CAN模块、关节电机位置采集与控制等程序。
6.对研制的黄瓜采摘机器人进行了性能测试实验。实验结果表明:机器人系统运行稳定可靠,机械手重复性定位精度为2.4mm;误差补偿后,x轴最大误差由补偿前4.3mm减少到3.5mm;y方向最大误差由6.9 mm减少到3.7mm;z轴误差基本不变。采摘成功率约为86%,采摘一根黄瓜平均耗时18s左右。
本课题为果蔬采摘机器人智能化和实用化的进一步研究提供了良好的基础,在今后的工作中,可以进一步从优化机器人机械结构和提高控制精度两方面进行深入地研究。
Fruits and vegetables harvest belongs to labor-intensive task. Because of population aging and lack of agricultural labor resources, the ratio of harvest price of human labor in the whole production price reaches to 33~50%, which greatly recedes their market competition. Therefore, studying and developing intelligent picking robots can not only liberate labor, improve productivity, decrease production price and maintain fresh quality of fruits and vegetables, but also promote further study and application of related techniques, such as mechanics, robotics, image processing, intelligent control and sensor. So it has significant practical meaning and academic value to develop picking robots.
On the comprehensive analysis of advantages and disadvantages of picking robots at home and abroad, taking the cucumber as research object, motion planning and control system of the cucumber picking robot are systematically studied based on techniques of robot mechanics, kinematic and dynamic analysis, trajectory planning and tracking, position error compensation and so on. Research contents and achievements are listed as follows:
1. Aiming at characteristics of cucumber picking operation, an articulated picking manipulator with four degree-of-freedom is developed, and the designed manipulator possesses favorable property of simple and compact structure, light weight, low price, flexible and stable picking motion. In order to harvest all of the fruits and vegetables in the target space, the general structure optimization method of articulated manipulators is proposed, which is applied in random cube of the picking space. Then the manipulator is optimized by MATLAB optimization Toolbox according to the target picking space of the cucumber.
2. Analysis and simulation on kinematics and dynamics of the designed manipulator are implemented. Forward kinematics is constructed based on the D-H method and the inverse kinematics is solved by analytical mehod; meanwhile, the simplified dynamics are constructed by utilizing Lagarange method. Finally, simulation experiments of kinematics, workspace, and dynamics are carried out, results verified the validity of the models and resonability of the structure parameters optimization method.
3. Algorithms of trajectory planning and tracking of the picking robot are studied. The trajectory planning algorithm based on combination of two cycloidal motions with grasping horizontally is proposed by comparing various planned algorithms and combined with the structure and picking operation characteristics of the cucumber picking robot. Moreover, in order to achieve precise tracking of the desired trajectory, a fast and non-singular terminal sliding mode controller is constructed, it can solve the problems of asymptotic convergence of the linear sliding mode controller, and the singularity and chattering of the conventional terminal sliding mode controller. Simulation experimental results show that the desired trajectory can be tracked precisely, and the postion tracking error can converge to equilibrium point in finite time, faster and high-precision tracking performance is obtained by using the presented algorithm.
4. Position error compensation algorithms of the picking robot are studied. The direct error compensation method based on LMBP network and the preseting position offsets are proposed, which are used to compensate for position error caused by structure parameter deviation, the validity of the proposed algorithms has verified by simulation experiments.
5. Control system of the cucumber picking robot is designed based on CAN bus communication. The distributed control mode of supervisory motion controller based on DSP+CAN bus+joint controllers based on DSP is presented. Hardware platform of control system is constructed; and programs for realizing the control system are designed by using modular principle, such as initiation, inverse kinematic computation, trajectory planning, the end effector control, CAN module, collection and control of the joint position and so on.
6. The performance test experiments of the cucumber picking robot are carried out. Experimental results show that the robot system can run stably and reliably, the repetitious position precision is about 2.4mm; after compensation, the maximum position errors in x and y axis have expectively declined from 4.3mm to 3.5mm, 6.9mm to 3.7mm; and position errors in z axis are nearly no changing. The ratio success of the robot’s picking is approximately 86%, and the average time consumed for picking a cucumber is about 18s.
This thesis provides favorable base for further research on intelligence and practicality of picking robots, and in the future, optimization of the mechanic structure and control accuracy should be studied more deeply.
引文
[1]沈明霞,姬长英.农业机器人的开发背景及技术动向[J].农机化研究, 2000, (5): 31-35.
[2]汤修映,张铁中.果蔬收获机器人研究综述[J].机器人, 2005, 27(1): 90-96.
[3]赵匀,武传宇,胡旭东等.农业机器人的研究进展及存在的问题[J].农业工程学报, 2003, 19(1): 20~24.
[4] Kondo N, Monta M. Basic study on chrysanthemum cutting sticking robot[C]. In proceedings of the international symposium on agricultural mechanization and automation, 1997, (1): 93-98.
[5] Kondo N, Monta M, Fujiura T. Fruit harvesting robots in Japan[J]. Adv.Space Res, 1996, 18 (1): 181-184.
[6] Kondo N, Monta M, Shibano Y. Agricultural robot in grape production system[C]. IEEE international conference on robotics and automation, 1995, 2504-2509.
[7] Kondo N, Ninomiya Kazunori, Hayashi Shigehiko et al. A new challenge of robot for harvesting strawberry grown on table top culture[C]. ASAE annual international meeting, 2005, 1-9.
[8] Shiigi Tomowo, Kondo N, Kurita Mitsutaka etc. Strawberry harvesting robot for fruits grown on table top culture [C]. ASABE annual international meeting, 2008, 1-9.
[9]周増产, Bontsema J, Van Kollenburg Crisan L.荷兰黄瓜收获机器人的研究开发[J].农业工程学报, 2001, 17(6): 77-80.
[10] Van Henten E J, Van Tuijl B J A, Hemminget J et al. An autonomous robot for harvesting cucumbers in greenhouses[J]. Autonomous robots, 2002, 13 (3): 241-258.
[11] Van Henten E J, Hemming J, Van Tuijl B J A et al. Collision-free motion planning for a cucumber picking robot[J]. Biosystems Engineering, 2003, 86(2): 135-144.
[12] Van Henten E J, Van Tuijl B J A , Hemming J et al. Field test of an autonomous cucumber picking robot[J]. Biosystems Engineering, 2003, 86(3): 305-313.
[13] Mario M Foglia, Giulio Reina. Agricultural robot for radicchio harvesting[J]. Journal of filed robotics, 2006, 23(6): 363-377.
[14] Reed J N, Miles S J, Butler J et al. Automatic mushroom harvester development[J]. J. agric. Engang Res, 2001, 78(1): 15-23.
[15]陆怀民.林木球果采集机器人设计与试验[J].农业机械学报, 2001, 32(6): 52-58.
[16]宋健,孙学岩,张铁中等.开放式茄子采摘机器人设计与试验[J].农业机械学报, 2009, 40(1): 143-147.
[17]高锐.草莓收获机器人初步研究[D].北京:中国农业大学,2004.
[18]陈利兵.草莓收获机器人采摘系统研究[D].北京:中国农业大学,2005.
[19]徐丽明.草莓收获机器人系统的研究[D].北京:中国农业大学,2006.
[20]姜丽萍.番茄力学特性及其在采摘机器人执行器设计中的应用[D].镇江:江苏大学,2006.
[21]戈志勇.机械手运动学仿真与避障[D].镇江:江苏大学,2007.
[22]王沈辉.机器人采摘番茄中的双目定位技术研究[D].镇江:江苏大学,2006.
[23]梁喜凤,王永维.番茄收获机械手奇异性分析与处理[J].农业工程学报, 2006, 22(1): 85-88.
[24]赵金英.基于三维视觉的西红柿采摘机器人技术研究[D].北京:中国农业大学,2006.
[25]赵金英,张铁中,杨丽.西红柿采摘机器人视觉系统的目标提取[J].农业机械学报, 2006, 37(10): 200-203.
[26]汤修映,张铁中.黄瓜收获机器人避碰轨迹规划[C]中国农业机械学会2008年学术年会论文集, 2008, 63-68.
[27]汤修映.果蔬收获机器人系统的研究[D].北京:中国农业大学,2006.
[28]田素博,邱立春,秦军伟等.国内外采摘机器人机械手结构比较的研究[J].农机化研究, 2007, (3): 195-197.
[29]方建军.移动式采摘机器人研究现状与进展[J].农业工程学报, 2004, 20(2): 273-278.
[30]殷际英,何广平.关节型机器人[M].北京:化学工业出版社, 2003.
[31]马香峰.工业机器人的操作机设计[M].北京:冶金工业出版社,1996.
[32]张家杰.国内外碳纤维生产现状及发展趋势[J].化工技术经济, 2005, (4): 12-15.
[33]白丽平.机器人的动力学与电机选择计算[D].沈阳:中国科学院沈阳自动化研究所,2000.
[34]宋健.开放式茄子采摘机器人关键技术研究[D].北京:中国农业大学,2006.
[35] Yang Qinghua, Zhang Libin, Bao Guanjun etc. Research on Novel Flexible Pneumatic Actuator FPA, IEEE RAM'2004, 2004, 12: 385-389.
[36]孙杏香.关节型机器人主连杆(手臂)参数的优化设计[J].北京航空航天大学学报, 1995, 22(4): 509-512.
[37]赵才军.机器人工作空间及关节运动可靠性规划[D].南京:南京理工大学,2003.
[38]龚水明,詹小刚.基于MATLAB优化工具的机械优化设计[J].机械工程师, 2008, (10): 92-94.
[39] Magrab Edward B, Balachandran Balakumar, Herold Keith E et al. MATLAB原理与工程应用[M].北京:电子工业出版社,2006.
[40]蔡自兴.机器人学[M].北京:清华大学出版社, 2000.
[41]付京逊,冈萨雷斯R C,李C S G.机器人学[M].北京:中国科学技术出版社,1989.
[42]王平,杨沿平,邓晓.基于模具表面抛光机器人系统的运动控制研究[J].中国机械工程, 2007, 18(20): 2422-2425.
[43]王洪斌,宋佐时,王洪瑞等.基于模糊神经网络的机器人逆运动学问题[J].系统仿真学报, 2002, 14(7): 852-858.
[44]付西光,商雨青,苏中义等.基于RBF网络的冗余扫查机器人逆运动学研究[J].中国机械工程, 2006, 17(17): 1765-1768.
[45] Rasit Koker. Reliability-based approach to the inverse kinematics solution of robots using Elman’s networks[J]. INTELLIGENCE, 2005, 18: 685-693.
[46]刘成良,张凯,付庄等. RV12L-6R焊接机器人动力学分析及计算机仿真研究[J].应用科学学报, 2002, 20(3): 222-223.
[47]石炜,郗安民,张玉宝.基于凯恩方法的机器人动力学建模与仿真[J].微计算机信息, 2008, 24(10): 222-223.
[48] Tanner H G, Kyriakopoulos K J, Krilelis N I. Advanced agricultural robots: kinematics and dynamics of multiple mobile manipulators handling non-rigid material [J]. Computers and electronics in agriculture, 2001, 31: 91-105.
[49]王卫忠,赵杰,高永生等.基于螺旋理论的可重构机器人动力学分析[J].机械工程学报, 2008, 44(11): 99-104.
[50]霍伟.机器人动力学与控制[M].北京:高等教育出版社,2005.
[51]罗家佳,胡国清.基于MATLAB的机器人运动仿真研究[J].厦门大学学报, 2005, 44(5): 640-644.
[52]曹毅,王树新,李群智.基于随机概率的机器人工作空间及其面积求解[J].制造业自动化, 2005, 27(2): 24-29.
[53] Snyman J A, Plessis L J. An optimization approach to the determination of the boundaries of manipulator workspace[J]. ASME Journal of Mechanical Design, 2000, 122: 447-456.
[54]崔玉洁,张祖立,范磊.基于蒙特卡洛方法的采摘机械手工作空间分析[J].农机化研究, 2007, 12: 62-63.
[55] Neelam R P, Kamal T S. Intelligent planning of trajectories for pick-and-place operations[C]. In Proceedings of International Conference on System, Man and Cybernetics. 2000, 55-60.
[56] Gasparetto A, Zanotto V. A new method for smooth trajectory planning of robot manipulators[J]. Mechanism and Machine Theory, 2007, (42): 455-471.
[57]徐向荣,马春峰.机器人运动轨迹规划与算法[J].机器人, 1988, 2(6): 18-25.
[58]韩军,郝立.机器人关节空间的轨迹规划及仿真[J].机器人, 2000, 24(6): 540-543.
[59] Xu Xiangrong. Trajectory planning of robot manipulators by using spline function approach[C]. Proceedings of the 3rd World Congress on Intelligent Control and Automation, 2000, 1215-1219.
[60]何平,刘宏,金明河.基于样条函数的机器人轨迹规划方法[J].机器人, 2003, 25(7): 614-618.
[61]张兴国,殷爱华.基于抛物线过渡法的环保压缩机装配机器人轨迹规划[J].机床与液压, 2005, (4): 52-53.
[62]方菲,刘平安.抛物线过渡的并联机器人线性轨迹规划[J].华东交通大学学报, 2005, 22(4): 113-116.
[63]王幼民.机械臂关节空间B样条曲线轨迹规划[J].安徽机电学院学报, 2000, 15(2): 21-26.
[64]王幼民,徐蔚鸿.机器人关节空间B样条轨迹优化设计[J].机电工程, 2000, 17(4): 57-60.
[65]王幼民,徐蔚鸿.机器人连续轨迹控制中的B样条轨迹优化[J].机械设计, 2000, 10(10): 33-35.
[66] Jorge Angeles,宋伟刚.机器人机械系统原理[M].北京:机械工业出版社,2004.
[67]庄鹏,姚正秋.基于摆线运动规律的悬索并联机器人轨迹规划[J].机械设计, 2006, 23(9): 21-24.
[68]田西勇,刘晓平,庄未.基于组合正弦函数的机器人轨迹规划方法[J].机械工程与自动化, 2008, (1): 136-138.
[69]田西勇.机器人轨迹规划方法研究[D].北京:北京邮电大学,2008.
[70]姚立健,丁为民,陈玉仑等.茄子收获机器人机械臂避障路径规划[J].农业机械学报, 2008, 39(11): 94-97.
[71] Fei Y Q, Ding F Q, Zhao X F. Collision-free motion planning of dual-arm reconfigurable robots[J]. Robotics and Computer-Integrated Manufacturing, 2004, 20(4): 351-357.
[72]王娟娟,曹凯.基于栅格法的机器人路径规划[J].农业装备与车辆工程, 2009, 213(4): 14-17.
[73]于红斌,李孝安.基于栅格法的机器人快速路径规划[J].微电子学与计算机, 2005, 22(6): 98-100.
[74] Erdinc S C. Path planning using potential fields for highly redundant manipulator[J]. Robotics and Autonomous Systems, 2005, 52(2): 209-228.
[75]王萌,王晓荣,李春贵.改进人工势场法的移动机器人路径规划研究[J].计算机工程与设计, 2008, 29(6): 1504-1506.
[76] Francesco A, Marco A, Carlo C. Finite-time stabilization via dynamic output feedback[J]. Automatica, 2006, (42): 337-342.
[77] Ouyang P R, Zhang W J, Gupta Madan M. An adaptive switching learning control method for trajectory tracking of robot manipulators[J]. Mechatronics, 2006, (16): 51-61.
[78] Chiou K C, Huang S J. An adaptive fuzzy controller for robot manipulator[J]. Mechatronics, 2005, (15): 151-177.
[79] Man Z H. A robust MIMO terminal sliding mode control scheme for rigid robotic manipulator. IEEE Trans on Automatic control, 2003, 39(8): 2462-2469.
[80] Slotine J J E. The robust control of robotic manipulators. International Journal of Robot and Research, 2003, 12(4): 49-54.
[81] Barambones O, Etxebarria V. Robust neural control for robotic manipulators[J]. Automatic, 2002, (38): 235-242.
[82] Wai R J. Tracking control based on neural network strategy of robot manipulator[J]. Neurocomptring, 2003, (51): 425-445.
[83]张昌凡,何静.滑模变结构的智能控制理论与应用研究[M].北京:科学出版社,2005.
[84]李运峰.滑模变结构理论在机器人中的控制应用[D].秦皇岛:燕山大学,2000.
[85]胡剑波,时满宏,庄开宇等.一类非线性系统的Terminal滑模控制[J].控制理论与应用, 2005, 22(3): 495-502.
[86]刘金琨.滑模变结构控制Matlab仿真[M].西安:西安电子科技大学出版社,2007.
[87]冯旭刚.基于模糊滑模方法的不确定机器人神经网络控制[J].冶金动力, 2006, (5): 79-80.
[88] Yu S H. Continuous finite time control for robotic manipulators with terminal sliding mode[J]. Automatica, 2005, (41): 1957-1964.
[89]冯勇,鲍晟,余星火.非奇异总段控制控制系统的设计方法[J].控制与决策, 2002, 17(2): 194-197.
[90]焦国太,冯永和,王峰等.多因素影响下的机器人综合位姿误差分析方法[J].应用基础与工程科学学报, 2004, 12(4): 435-442.
[91] Heisel U, Richter F, Wurst K H. Thermal behaviour of industrial robots and possibilities for error compensation[J]. CIRP Annals - Manufacturing Technology, 1997, 46(1): 283-286.
[92]叶东,黄庆成,车仁生.多关节坐标测量机的误差模型[J].光学精密工程, 1999, 7(2): 91-96.
[93]焦国太,李庆,冯永和等.机器人位姿误差的综合补偿[J].华北工学院学报, 2003, 24(2): 104-107.
[94]祝建礼.多自由度检测机器人控制技术开发与位姿误差补偿[D].浙江大学, 2007.
[95]张德丰. MATLAB神经网络应用设计[M].机械工业出版社, 2009.
[96] Rasit K, Cemil O, Tarik C et al. A study of nerual network based inverse kinematics solution for a three-joint robot[J]. Robotics and Autonomous systems, 2004, (49): 227-234.
[97]韩力群.人工神经网络理论、设计及应用[M].北京:化学工业出版社,2007.
[98]李炯城,黄汉雄.神经网络中LMBP算法收敛速度改进的研究[J].计算机工程与应用, 2006, 42(16): 46-49.
[99]王海鸣,孔凡让,赵晓伟等.基于BP神经网络的机器人逆运动学新算法[J].机电一体化, 2009, (1): 46-49.
[100]鞠儒生,王学宁,刘宝宏等.一种改进型神经网络算法NN-LMBP[J].系统仿真学报, 2007, 19(21): 4857-4859.
[101]武传宇.基于PC+DSP模式的开放式机器人控制系统及其应用研究[D].杭州:浙江大学, 2002.
[102]聂磊.基于CAN总线的机器人嵌入式控制系统设计[D].哈尔滨:哈尔滨工程大学, 2006.
[103] dsPIC30F4011/4012 datasheet. dsPIC30F Family Reference Manual, 2007.
[104]何礼高. dsPIC30F电机与电源系列数字信号控制器原理与应用[M].北京:北京航天航空大学出版社, 2007.
[105] MCP2551, High Speed CAN Transceiver. Microchip Technology Inc, 2005.
[106] Austriamicro systems. AS5045 12 bit programmable magnetic rotary encoder datasheet, 2005.
[107]阳宪惠.工业数据通信与控制网络[M].北京:清华大学出版社, 2003.
[108] Hofstee J W, Goense D. Simulation of a CAN-based tractor implement field bus according to DIN 9684[J]. Journal Agricultural Engineering Research, 1999, 73(4): 383-394.