基于多传感器信息融合的码垛过程监控与故障诊断的研究
|
摘要
在自动化生产过程中,要用各种不同的传感器来监视或控制生产过程中的各个参数,使设备处于最佳工作状态,保证产品有最好的质量。但目前主要依靠内部传感器,例如各种位移、速度、压力等传感器,很少采用视觉等外部传感器。并且在对系统运行状态进行监控时,单一种传感器信息只能获得设备系统的部分信息段,反映设备系统运行状态的某一个侧面;并且如果传感器本身发生故障,系统并不能自行判断某个传感器失效并发出警报。为了全面地、综合地反映设备的真实运行状态并就故障给出相应的诊断信息,本文提出了将视觉和内部传感器信号感知信息融合监控诊断的方法,从而真实的反映设备运行状态,对设备故障做出科学、正确的分析与决策。
对监控用内外传感器进行了选择和设置。重点对视觉跟踪中的采集信息进行了分析,采用点模式匹配的方法对目标进行跟踪,并设计了卡尔曼滤波器,对被跟踪目标的运动参数进行滤波预测,缩小了搜索区域,提高了系统的实时性和跟踪精度。
分析了单独使用视觉或内部传感器进行系统监控的不足之处,在此基础上,针对码垛机器人的码垛过程、产品的安全性等要求,将多传感器信息融合技术应用到码垛系统中,并以信息融合技术的理论知识为基础,总结得到码垛系统的多传感器信息融合的基本原理图、所采用的结构形式以及信息融合故障检测方法。针对码垛机器人多传感器系统的特性和现场环境特点,总结出系统常见的故障类型,并采用模糊BP神经网络信息融合方法实现对码垛系统的故障监控和诊断。本文通过码垛系统的故障初始训练表形成了故障诊断理论样本,然后就码垛系统中的典型故障模式采用了模糊BP神经网络进行判断,仿真结果表明该方法是行之有效的。
本文深入研究了基于多传感器信息融合的码垛过程故障监控与诊断的问题,当码垛过程发生故障时,该系统可以判断出故障原因并给出诊断结果,从而保证了系统运行的稳定性和可靠性。
During the automatic production, people use all kinds of sensor to monitor and control each parameter. Make equipment the best operating state and ensure product the best quality. But,rely mainly on the internal sensor now, for example, all kinds of displacement, velocity, pressure sensors, and use visual(external sensor) so little. In the monitoring system running state, a single sensor information can only get part of the information systems equipment, reflect one aspect of the a system running state; and if sensor break down itself, the system can not make their own judgment as a sensor failure and the alarm. To reflect roundly, synthetically the reality state of equipment and fault diagnosis is given corresponding information, This paper advanced a monitoring and diagnostic information fusion method on visual and the internal sensor signal, which reflect the real equipment operations, get equipment failure scientific, correct analysis and decision-making.
Choose and set Internal and external sensors for monitor , Thereinto, we placed emphasis on vision sensor. This paper tracked target with the point-pattern matching(PPM)algorithm. And a Kalman filter is designed to filter and predict motion parameter of tracked target. By this method, searching area is reduced (reduced to the area of a matched model), response time and accuracy of tracking are improved.
A analysis get deficiencies of the separate use of visual sensor systems or internal sensor systems. on that basis ,To contrapose palletizing process, the safety of such products, and other requirements of Palletizing Robot, apply multi-sensor information fusion technology to the palletizing system.And based on the Theory knowledge of information fusion technology, we can get the basic Schematic diagram, Structural form and fault diagnosis frame of robot palletizing mufti-sensor information fusion fault diagnosis technology. In view of the characteristic and complex environment of robot palletizing mutisensor system, we can get the common fault type, and uses the fuzzy nerve network information fusion method to carry on the fault monitoring and diagnosis of the robot palletizing multisensor system. This paper formed fault diagnosis theoretics sample, took the fuzzy BP nerve network to judge based on typical fault. The simulation experiment indicated that, this method is very good in fault diagnosis.
This paper has thoroughly studied information fusion fault monitoring and diagnosis question of robot palletizing multisensor system. When palletizing robot breaks down, this system can judge the fault reason and get diagnosis result. which has ensured the stability and the reliability of system.
对监控用内外传感器进行了选择和设置。重点对视觉跟踪中的采集信息进行了分析,采用点模式匹配的方法对目标进行跟踪,并设计了卡尔曼滤波器,对被跟踪目标的运动参数进行滤波预测,缩小了搜索区域,提高了系统的实时性和跟踪精度。
分析了单独使用视觉或内部传感器进行系统监控的不足之处,在此基础上,针对码垛机器人的码垛过程、产品的安全性等要求,将多传感器信息融合技术应用到码垛系统中,并以信息融合技术的理论知识为基础,总结得到码垛系统的多传感器信息融合的基本原理图、所采用的结构形式以及信息融合故障检测方法。针对码垛机器人多传感器系统的特性和现场环境特点,总结出系统常见的故障类型,并采用模糊BP神经网络信息融合方法实现对码垛系统的故障监控和诊断。本文通过码垛系统的故障初始训练表形成了故障诊断理论样本,然后就码垛系统中的典型故障模式采用了模糊BP神经网络进行判断,仿真结果表明该方法是行之有效的。
本文深入研究了基于多传感器信息融合的码垛过程故障监控与诊断的问题,当码垛过程发生故障时,该系统可以判断出故障原因并给出诊断结果,从而保证了系统运行的稳定性和可靠性。
During the automatic production, people use all kinds of sensor to monitor and control each parameter. Make equipment the best operating state and ensure product the best quality. But,rely mainly on the internal sensor now, for example, all kinds of displacement, velocity, pressure sensors, and use visual(external sensor) so little. In the monitoring system running state, a single sensor information can only get part of the information systems equipment, reflect one aspect of the a system running state; and if sensor break down itself, the system can not make their own judgment as a sensor failure and the alarm. To reflect roundly, synthetically the reality state of equipment and fault diagnosis is given corresponding information, This paper advanced a monitoring and diagnostic information fusion method on visual and the internal sensor signal, which reflect the real equipment operations, get equipment failure scientific, correct analysis and decision-making.
Choose and set Internal and external sensors for monitor , Thereinto, we placed emphasis on vision sensor. This paper tracked target with the point-pattern matching(PPM)algorithm. And a Kalman filter is designed to filter and predict motion parameter of tracked target. By this method, searching area is reduced (reduced to the area of a matched model), response time and accuracy of tracking are improved.
A analysis get deficiencies of the separate use of visual sensor systems or internal sensor systems. on that basis ,To contrapose palletizing process, the safety of such products, and other requirements of Palletizing Robot, apply multi-sensor information fusion technology to the palletizing system.And based on the Theory knowledge of information fusion technology, we can get the basic Schematic diagram, Structural form and fault diagnosis frame of robot palletizing mufti-sensor information fusion fault diagnosis technology. In view of the characteristic and complex environment of robot palletizing mutisensor system, we can get the common fault type, and uses the fuzzy nerve network information fusion method to carry on the fault monitoring and diagnosis of the robot palletizing multisensor system. This paper formed fault diagnosis theoretics sample, took the fuzzy BP nerve network to judge based on typical fault. The simulation experiment indicated that, this method is very good in fault diagnosis.
This paper has thoroughly studied information fusion fault monitoring and diagnosis question of robot palletizing multisensor system. When palletizing robot breaks down, this system can judge the fault reason and get diagnosis result. which has ensured the stability and the reliability of system.
引文
[1] 陈前. 关于工程大系统的状态监测与故障诊断. 振动测试与诊断,2002, (03):163-172
[2] 张菡英. 现代设备管理方法之一——设备状态监测与故障诊断简介. 山东化工,2001,(06) :44-47
[3] 刘玉才,储江伟,邹本存,等.机械设备状态监测及其故障诊断.林业机械与木工设备,2000,(02): 23-25
[4] 王志鹏.基于信息融合技术的故障诊断方法的研究及应用:[博士学位论文].大连:大连理工大学,2001.6
[5] Farina A, Studer F A. Rader date processing. Advanced topic and applications, New York,1986
[6] 刘同明,曾富贵,刘伟.数据融合与数据挖掘相集成的自动目标识别系统.计算机工程与应用, 2002,(03):92-95
[7] Chong chee-yee, MoriS, Barker WH, et a. Architectures and algorithms for track association and fusion. IEEE Aerospace and Electronic System Magazine 2000 IEEE Piscataway NJUSA 5-130885-8985 IESMEA
[8] 陈永光. 复合干扰对多级雷达系统的对抗效能研究. 现代雷达, 1999, (04):1-6
[9] Nejatali A, Ciric I R. Novel image fusion methodology using fuzzy set theory. Optical Engineering 1998, 37(02): 485-491
[10] Solaiman B, Debon R, et al. Information fusion:Application to data and model fusion for Ul trasound image segmentation. IEEE Transactions on Biomedical Engineering. 1999, 46(10):1171-1175.
[11] 曲晓慧, 安钢. 数据融合方法综述及展望.舰船电子工程, 2003, (02):2-4
[12] 周杰, 朱允民, 宋恩彬. 传感器网络中信息融合的几个基本问题. 信息与控制, 2006,(02)
[13] 张明路, 戈新良, 唐智强,等. 多传感器信息融合技术研究现状和发展趋势. 河北工业大学学报, 2003,(02):30-34
[14] 张雨, 温熙森.设备故障信息融合问题思考.长沙交通学院学报, 1999, 15(02):22-29
[15] 何友等. 多传感器信息融合及其应用. 电子工业出版社, 2000:24-36
[16] Valet.L, Mauris.G, et al. A statistical overview of recent literature in information fusion[G]. ISIF2000: MoC3-22-Moc3-29
[17] Wald L. An european proposal for terms of reference in data fusion. International Archives of Photogrammetry and Re-mote Sensing, 1998.XXXⅡ(07):651-654
[18] Waltz E, Linas J. Multisensor data fusion[M]. Boston: ArtechHouse, 1990
[19] Bar-Shalom Y, Fortmann T E.Tracking and Data Application. Norwood: Artech House, 1986
[20] 案例分析之四 机器人码垛单元. 机器人技术与应用, 2000, (04):21-22
[21] 李刚.工业用码垛机器人. 现代制造, 2005, (24):40-41
[22] Devdas Shetty, Richard A.Kolk(张树生译).机电一体化设计. 北京: 机械工业出版社.2006.1
[23] 权太范. 信息融合神经网络—模糊推理理论与应用. 北京: 国防工业出版社, 2002
[24] Special issue on data fusion. Proceedings of The IEEE,1997,85(1)
[25] Thomopoulos S C A.Optimal decision fusion in multiple sensor systems. IEEE Trans. AES, 1987, 23(05): 644-653
[26] Llinas J, Waltz E. Multisensor data fusion. Boston: Artech House,1990
[27] Tao L.Optimal multiple level decision fusion with distributed sensors.IEEE Trans.AES,1993,29(04):1251-1259
[28] Bar-Shalom Y, Li X. Multitarget multisensor tracking: Principles and techniques. New York: YBS 1995
[29] 覃祖旭, 张洪锥. 信息融合的一般过程及其在故障诊断中的应用. 测控技术, 1995, 14(03):5-7
[30] 李瑛. 复杂背景下动目标图像跟踪方法的研究: [硕士学位论文]. 太原:中北大学, 2006.3
[31] Barron J, Fleet D. Beauchem in S. Performance of optical flow techniques[J]. International Journal of Computer Vision, 1994, 12(01):42-47
[32] T.Gautama, M.Mvan Hulle. A Phase-based approach to the estimation of the optical flow field using spatial filter, IEEE Trans. Neural Networks, 13(5):1127-1136.
[33] 朱明旱, 罗大庸, 曹倩霞. 帧间差分与背景差分相融合的运动目标检测算法. 计算机测量与控制. 2005, 13(03):215-217
[34] 魏志强,纪筱鹏,冯业伟.基于自适应背景图像更新的运动目标检测方法.电子学报, 2005, 33(12):2261-2264
[35] 何斌, 马天予, 王运坚等. Visual C++数字图像处理(第二版). 北京: 人民邮电出版社,2002:459-471
[36] 侯志强, 韩崇昭. 视觉跟踪技术综述. 自动化学报, 2006, 32(04):603-617
[37] Aschwanden.P, Guggenbuhl.W. Experimental results from a comparative study on correlation-type registration algorithms. In Robust Computer Vision(Forstner and Ruwiedel, Eds.), Wickmann, 1993 ,pp: 268-289
[38] 许波, 李正明. 一种新的基于自适应模板的相关跟踪算法. 光学与光电技术, 2004.2(4):62-64
[39] 武林伟, 马宝峰. 相关跟踪算法的频域研究和改进. 光电系统, 2005.12
[40] 张有为. 维纳与卡尔曼滤波理论导论. 北京: 人民教育出版社, 1980
[41] Kalman,R.E. A New approach to linear filtering and prediction problems. Transaction of the ASME--Journal of Basic Engineering. 1960:35-45
[42] Welch.G, Bishop.G. An introduction to the Kalman Filter. 2001, http://www.cs.unc. edu/~welch/kalman/
[43] 徐剑, 段哲民. 卡尔曼预测器在目标相关跟踪中的应用. 计算机仿真, 2005,22(11):120-122
[44] 何凯, 安如, 周绍光, 金夏玲. 一种快速角点探测算子研究.测绘学报, 2005, 34(03):223-227
[45] Perez,M.M., Dennis. An adaptive implementation of the SUSAN method for image edge and feature detection[A]. Image Processing[C].1997,International Conference on Volume 2:26-29, 394-397
[46] 胡社教, 葛西旺, 陈宗海, 基于角点特征的 KLT 跟踪全景图像拼接算法. 系统仿真学报, 2007, 19(08):1742-1744
[47] 刘博, 仲思东. 一种基于自适应阈值的 SUSAN 角点提取方法. 红外技术, 2006, 28(06):331-333
[48] SMITH SM, BRADY M. SUSAN—A new approach to low level image processing . International Jaurnal of Computer Vision, 1997, 23(01):45-78.
[49] Smith S M, Brady J M. SUSAN-A new approach to low level image processing [R]. Internal Technical Report TR95SMS1, Defence Research Agency, Chobham Lane Chertsey Surry UK, 1995.
[50] 张立华, 徐文立. 点模式匹配. 计算机学报, 1999, 22(07):740-745
[51] Sprinzak J, Werman M. Affine point matching. Pattern Recognition Letters, 1994, 15:337-339
[52] 徐文立, 张立华. 一种基于几何推理的点模式匹配算法.中国科学 E 辑, 2002, 32(04):553-560
[53] Wison G B. Some aspects of data fusion. 1985 Intelligence Conference in Advances in C3I. Proceedings of IEEE conference, United Kingdom, 1985:99-105
[54] 吴浩中, 王开文. 模糊神经网络在摆式列车作动系统故障诊断中的应用. 中国铁道科学, 2002, (03): 33-36
[55] 潘兵, 熊静琪.多传感器信息融合在液压系统智能故障诊断中的应用.机床与液压. 2006, (05):190-192
[56] 王浩, 庄钊文. 模糊可靠性分析中的隶属函数确定. 电子产品可靠性与环境试验, 2000, (04):2-7
[57] Zhu Diqi, Yu Shenglin.Neural networks data fusion a lgorithmof electronic equipment fault diagnosis[C]. Proceedings of the 5th world congress on intelligent control and Automation, June5-19, 2004
[58] Wen Jin wei, Zhao Jiali, Luo Siwe, etal. The improvements of BP neural network learning algorithm[C]. Prodeedings of ICSP, 2000
[59] 丛爽. 面向 MATLAB 工具箱的神经网络理论与应用. 合肥:中国科技技术大学出版社, 1998
[60] 飞思科技产品研发中心. 神经网络理论与 MATLAB7 实现. 北京:电子工业出版社, 2005.3
[2] 张菡英. 现代设备管理方法之一——设备状态监测与故障诊断简介. 山东化工,2001,(06) :44-47
[3] 刘玉才,储江伟,邹本存,等.机械设备状态监测及其故障诊断.林业机械与木工设备,2000,(02): 23-25
[4] 王志鹏.基于信息融合技术的故障诊断方法的研究及应用:[博士学位论文].大连:大连理工大学,2001.6
[5] Farina A, Studer F A. Rader date processing. Advanced topic and applications, New York,1986
[6] 刘同明,曾富贵,刘伟.数据融合与数据挖掘相集成的自动目标识别系统.计算机工程与应用, 2002,(03):92-95
[7] Chong chee-yee, MoriS, Barker WH, et a. Architectures and algorithms for track association and fusion. IEEE Aerospace and Electronic System Magazine 2000 IEEE Piscataway NJUSA 5-130885-8985 IESMEA
[8] 陈永光. 复合干扰对多级雷达系统的对抗效能研究. 现代雷达, 1999, (04):1-6
[9] Nejatali A, Ciric I R. Novel image fusion methodology using fuzzy set theory. Optical Engineering 1998, 37(02): 485-491
[10] Solaiman B, Debon R, et al. Information fusion:Application to data and model fusion for Ul trasound image segmentation. IEEE Transactions on Biomedical Engineering. 1999, 46(10):1171-1175.
[11] 曲晓慧, 安钢. 数据融合方法综述及展望.舰船电子工程, 2003, (02):2-4
[12] 周杰, 朱允民, 宋恩彬. 传感器网络中信息融合的几个基本问题. 信息与控制, 2006,(02)
[13] 张明路, 戈新良, 唐智强,等. 多传感器信息融合技术研究现状和发展趋势. 河北工业大学学报, 2003,(02):30-34
[14] 张雨, 温熙森.设备故障信息融合问题思考.长沙交通学院学报, 1999, 15(02):22-29
[15] 何友等. 多传感器信息融合及其应用. 电子工业出版社, 2000:24-36
[16] Valet.L, Mauris.G, et al. A statistical overview of recent literature in information fusion[G]. ISIF2000: MoC3-22-Moc3-29
[17] Wald L. An european proposal for terms of reference in data fusion. International Archives of Photogrammetry and Re-mote Sensing, 1998.XXXⅡ(07):651-654
[18] Waltz E, Linas J. Multisensor data fusion[M]. Boston: ArtechHouse, 1990
[19] Bar-Shalom Y, Fortmann T E.Tracking and Data Application. Norwood: Artech House, 1986
[20] 案例分析之四 机器人码垛单元. 机器人技术与应用, 2000, (04):21-22
[21] 李刚.工业用码垛机器人. 现代制造, 2005, (24):40-41
[22] Devdas Shetty, Richard A.Kolk(张树生译).机电一体化设计. 北京: 机械工业出版社.2006.1
[23] 权太范. 信息融合神经网络—模糊推理理论与应用. 北京: 国防工业出版社, 2002
[24] Special issue on data fusion. Proceedings of The IEEE,1997,85(1)
[25] Thomopoulos S C A.Optimal decision fusion in multiple sensor systems. IEEE Trans. AES, 1987, 23(05): 644-653
[26] Llinas J, Waltz E. Multisensor data fusion. Boston: Artech House,1990
[27] Tao L.Optimal multiple level decision fusion with distributed sensors.IEEE Trans.AES,1993,29(04):1251-1259
[28] Bar-Shalom Y, Li X. Multitarget multisensor tracking: Principles and techniques. New York: YBS 1995
[29] 覃祖旭, 张洪锥. 信息融合的一般过程及其在故障诊断中的应用. 测控技术, 1995, 14(03):5-7
[30] 李瑛. 复杂背景下动目标图像跟踪方法的研究: [硕士学位论文]. 太原:中北大学, 2006.3
[31] Barron J, Fleet D. Beauchem in S. Performance of optical flow techniques[J]. International Journal of Computer Vision, 1994, 12(01):42-47
[32] T.Gautama, M.Mvan Hulle. A Phase-based approach to the estimation of the optical flow field using spatial filter, IEEE Trans. Neural Networks, 13(5):1127-1136.
[33] 朱明旱, 罗大庸, 曹倩霞. 帧间差分与背景差分相融合的运动目标检测算法. 计算机测量与控制. 2005, 13(03):215-217
[34] 魏志强,纪筱鹏,冯业伟.基于自适应背景图像更新的运动目标检测方法.电子学报, 2005, 33(12):2261-2264
[35] 何斌, 马天予, 王运坚等. Visual C++数字图像处理(第二版). 北京: 人民邮电出版社,2002:459-471
[36] 侯志强, 韩崇昭. 视觉跟踪技术综述. 自动化学报, 2006, 32(04):603-617
[37] Aschwanden.P, Guggenbuhl.W. Experimental results from a comparative study on correlation-type registration algorithms. In Robust Computer Vision(Forstner and Ruwiedel, Eds.), Wickmann, 1993 ,pp: 268-289
[38] 许波, 李正明. 一种新的基于自适应模板的相关跟踪算法. 光学与光电技术, 2004.2(4):62-64
[39] 武林伟, 马宝峰. 相关跟踪算法的频域研究和改进. 光电系统, 2005.12
[40] 张有为. 维纳与卡尔曼滤波理论导论. 北京: 人民教育出版社, 1980
[41] Kalman,R.E. A New approach to linear filtering and prediction problems. Transaction of the ASME--Journal of Basic Engineering. 1960:35-45
[42] Welch.G, Bishop.G. An introduction to the Kalman Filter. 2001, http://www.cs.unc. edu/~welch/kalman/
[43] 徐剑, 段哲民. 卡尔曼预测器在目标相关跟踪中的应用. 计算机仿真, 2005,22(11):120-122
[44] 何凯, 安如, 周绍光, 金夏玲. 一种快速角点探测算子研究.测绘学报, 2005, 34(03):223-227
[45] Perez,M.M., Dennis. An adaptive implementation of the SUSAN method for image edge and feature detection[A]. Image Processing[C].1997,International Conference on Volume 2:26-29, 394-397
[46] 胡社教, 葛西旺, 陈宗海, 基于角点特征的 KLT 跟踪全景图像拼接算法. 系统仿真学报, 2007, 19(08):1742-1744
[47] 刘博, 仲思东. 一种基于自适应阈值的 SUSAN 角点提取方法. 红外技术, 2006, 28(06):331-333
[48] SMITH SM, BRADY M. SUSAN—A new approach to low level image processing . International Jaurnal of Computer Vision, 1997, 23(01):45-78.
[49] Smith S M, Brady J M. SUSAN-A new approach to low level image processing [R]. Internal Technical Report TR95SMS1, Defence Research Agency, Chobham Lane Chertsey Surry UK, 1995.
[50] 张立华, 徐文立. 点模式匹配. 计算机学报, 1999, 22(07):740-745
[51] Sprinzak J, Werman M. Affine point matching. Pattern Recognition Letters, 1994, 15:337-339
[52] 徐文立, 张立华. 一种基于几何推理的点模式匹配算法.中国科学 E 辑, 2002, 32(04):553-560
[53] Wison G B. Some aspects of data fusion. 1985 Intelligence Conference in Advances in C3I. Proceedings of IEEE conference, United Kingdom, 1985:99-105
[54] 吴浩中, 王开文. 模糊神经网络在摆式列车作动系统故障诊断中的应用. 中国铁道科学, 2002, (03): 33-36
[55] 潘兵, 熊静琪.多传感器信息融合在液压系统智能故障诊断中的应用.机床与液压. 2006, (05):190-192
[56] 王浩, 庄钊文. 模糊可靠性分析中的隶属函数确定. 电子产品可靠性与环境试验, 2000, (04):2-7
[57] Zhu Diqi, Yu Shenglin.Neural networks data fusion a lgorithmof electronic equipment fault diagnosis[C]. Proceedings of the 5th world congress on intelligent control and Automation, June5-19, 2004
[58] Wen Jin wei, Zhao Jiali, Luo Siwe, etal. The improvements of BP neural network learning algorithm[C]. Prodeedings of ICSP, 2000
[59] 丛爽. 面向 MATLAB 工具箱的神经网络理论与应用. 合肥:中国科技技术大学出版社, 1998
[60] 飞思科技产品研发中心. 神经网络理论与 MATLAB7 实现. 北京:电子工业出版社, 2005.3