针对单自由度机械臂的系统中存在多个区间不确定性参数,并在工作时受到外部随机扰动的影响,提出了一种基于深度学习神经网络模型的方法来估计单自由度机械臂的定位可靠性。对系统进行不确定性分析,利用区间数描述系统动作过程中的不确定性参数,建立单自由度机械臂的不确定模型;区间不确定参数在一次动作过程中可视为定值,通过实验数据结合参数辨识的方法辨识出几组实验的参数,并验证单自由度机械臂仿真模型的准确性;对区间不确定参数进行拉丁超立方采样代入仿真模型,机械臂在外部随机扰动中实现定位,得到训练神经网络模型的样本数据,构造基于Levenberg-Marquardt
(levenberg-marquardt,LM)算法的神经网络模型,进行了Monte-Carlo仿真分析,估计了单自由度机械臂的定位可靠性为84.12%。最后通过多组实验数据的分析,表明提出的方法具有高效性和有效性,可为其他非线性复杂系统的定位可靠性估算提供了新的思路。
Abstract
Here, aiming at multiple parameters with interval uncertainty existing in a SDOF manipulator system and the system being subjected to effects of external random disturbance during operation, a method based on deep learning neural network model was proposed to estimate positioning reliability of the manipulator. Firstly, the uncertainty analysis was performed for the system, and uncertain parameters in the system operation process were described using interval number to establish the uncertain model of the SDOF manipulator. Interval uncertain parameters could be regarded as fixed values in process of one action. Parameters of several sets of experiments were identified through combining experiment data and the parametric identification method, and the correctness of the simulation model of SDOF manipulator was verified. Secondly, Latin hypercube sampling was done for interval uncertain parameters and the sampled data were brought into the simulation model. The manipulator could realize positioning in external random disturbance to obtain sample data for training the neural network model, construct the neural network model based on levenberg marquardt (LM) algorithm, and perform Monte Carlo simulation analysis. It was shown that the positioning reliability of the SDOF manipulator is estimated as 84.12%. Finally, analyzing several sets of experimental data showed that the proposed method can have high efficiency and effectiveness, and provide a new idea for positioning reliability estimation of other nonlinear complex systems.
关键词
机械臂 /
深度学习 /
定位可靠性 /
参数辨识 /
随机扰动
{{custom_keyword}} /
Key words
manipulator /
deep learning /
positioning reliability /
parametric identification /
random disturbance
{{custom_keyword}} /
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1]RAO S S, BHATTI P K. Probabilistic approach to manipulator kinematics and dynamics[J]. Reliability Engineering and System Safety, 2001, 72(1): 47-58.
[2]KANG Z, LUO Y J. Reliability-based structural optimization with probability and convex set hybrid models [J]. Structural and Multidisciplinary Optimization, 2010, 42(1):89-102.
[3]姜潮,刘丽新,龙湘云.一种概率-区间混合结构可靠性的高效计算方法[J].计算力学学报,2013,30(5):605-609.
JIANG Chao, LIU Lixin, LONG Xiangyun. An efficient method for calculating the reliability of probabilistic-interval hybrid structures[J]. Chinese Journal of Computational Mechanics,2013, 30(5):605-609.
[4]孟广伟,魏彤辉,周立明,等.基于泰勒展开法的结构混合可靠性分析[J].兵工学报,2018,39(7):1404-1410.
MENG Guangwei, WEI Tonghui, ZHOU Liming, et al. Hy-brid reliability analysis of structures based on taylor expan-sion method[J]. Acta Armamentarii, 2018, 39(7):1404-1410.
[5]邱涛,张建国,邱继伟,等.基于二参数寻优设计点的混合结构可靠性分析算法[J].兵工学报,2019,40(4):869-873.
QIU Tao, ZHANG Jianguo, QIU Jiwei, et al. Two-parameter optimization design point-based reliability analysis algorithm for structures with mixed uncertainty[J]. Acta Armamentarii,2019, 40(4):865-873.
[6]LI M, WU H, WANG Y, et al. Modified levenberg-marquardt algorithm for backpropagation neural network training in dynamic model identification of mechanical systems[J]. Journal of Dynamic Systems Measurement and Control-Transactions of the ASME, 2017, 139(3): 031012.
[7]朱坚民,周亚南,何丹丹,等. 基于神经网络建模的机床滑动结合面动态特性参数识别[J]. 振动与冲击, 2018, 37(7): 109-115.
ZHU Jianmin, ZHOU Yanan, HE Dandan, et al. Dynamic characteristic parameters identification for machine tool sliding joints based on neural network modeling[J]. Journal of Vibration and Shock, 2018, 37(7): 109-115.
[8]高学星,侯保林,孙华刚.弹药协调器动作可靠性估计[J].弹道学报,2015,27(4):84-90.
GAO Xuexing, HOU Baolin, SUN Huagang. Action reliability estimation of shell transfer arm[J]. Journal of Ballistics, 2015, 27(4):84-90.
[9]周志华. 机器学习[M]. 北京: 清华大学出版社,2018.
[10]赵抢抢,侯保林.火炮弹药协调器区间不确定参数辨识[J].兵工学报,2017,38(1):35-42.
ZHAO Qiangqiang, HOU Baolin. Identification of interval uncertainty parameters of a howitzer shell transfer arm[J]. Acta Armamentarii, 2017,38(1):35-42.
[11]MOREL M, ACHARD C, KULPA R, et al. Time-series averaging using constrained dynamic time warping with tolerance[J]. Pattern Recognition, 2018, 74:77-89.
[12]常军,刘大山. 基于量子粒子群算法的结构模态参数识别 [J]. 振动与冲击, 2014, 33(14): 72-76.
CHANG Jun, LIU Dashan. Approach on structural modal parameter identification based on quantum-behaved particle swarm optimization[J]. Journal of Vibration and Shock, 2014, 33(14): 72-76.
[13]GOODFELLOW I, BENGIO Y, COURVILLE A. Deep learning[M]. MIT Press, 2016.
[14]INTRATOR O, INTRATOR N. Interpreting neural-network results: a simulation study[J]. Computational Statistics and Data Analysis,2001,37(3):373-393.
[15]SMITH J S, WU B, WILAMOWSKI B M. Neural network training with levenberg-marquardt and adaptable weight compression[J]. IEEE Transactions on Neural Networks and Learning Systems, 2019, 30(2): 580-587.
[16]TRA V, KIM J, KHAN S A, et al. Bearing fault diagnosis under variable speed using convolutional neural networks and the Stochastic Diagonal Levenberg-Marquardt Algorithm[J]. Sensors (Basel, Switzerland), 2017, 17(12):2834.
{{custom_fnGroup.title_cn}}
脚注
{{custom_fn.content}}