基于改进布谷鸟搜索优化神经网络的AMT换挡电机控制
|
摘要
针对神经网络在训练过程中出现权值和阈值更新不及时和误差较大等问题,提出一种基于改进布谷鸟搜索(CS)算法优化神经网络训练权值和阈值的方法,该算法模拟布谷鸟繁殖策略能够对电控机械式自动变速器(AMT)换挡电机进行位置和转矩实时跟踪控制以提高AMT换挡精度。通过对AMT换挡电机进行建模和仿真,研究经CS算法优化后的神经网络在控制电机时负载转矩的瞬态响应性能和抗负载扰动能力。通过与常规神经网络训练方法对比,结果表明,改进的CS算法优化神经网络控制策略在负载扰动、跟踪精度、控制响应等方面具有良好的动态性能和鲁棒性,其中响应时间和达到稳态时间分别减少了0. 02 s和0. 08 s,因此所提出的改进CS优化神经网络控制策略能够有效改善网络训练方法提高AMT的换挡品质。
In view of the problems that the weight and threshold updating of neural network is not timely and the errors are large during training, a method based on improved Cuckoo Search( CS) algorithm to optimize the weights and thresholds of neural network training was proposed. In the algorithm, the cuckoo breeding strategy was simulated to improve the AMT( Automatic Manual Transmission) shifting accuracy by performing position and torque real-time tracking control on the electronically controlled AMT shifting motor. By modeling and simulating the AMT shifter motor, the transient response performance and anti-load perturbation ability of the load torque of the neural network optimized by the CS algorithm when controlling the motor were studied. Compared with the conventional neural network training method, the results show that the improved CS neural network control strategy has good dynamic performance and robustness in terms of load disturbance,tracking accuracy, control response, etc., in which the response time and reach the steady-state time reduced by 0. 02 s and0. 08 s respectively. Therefore, the proposed improved CS optimization neural network control strategy can effectively improve the AMT shift quality.
In view of the problems that the weight and threshold updating of neural network is not timely and the errors are large during training, a method based on improved Cuckoo Search( CS) algorithm to optimize the weights and thresholds of neural network training was proposed. In the algorithm, the cuckoo breeding strategy was simulated to improve the AMT( Automatic Manual Transmission) shifting accuracy by performing position and torque real-time tracking control on the electronically controlled AMT shifting motor. By modeling and simulating the AMT shifter motor, the transient response performance and anti-load perturbation ability of the load torque of the neural network optimized by the CS algorithm when controlling the motor were studied. Compared with the conventional neural network training method, the results show that the improved CS neural network control strategy has good dynamic performance and robustness in terms of load disturbance,tracking accuracy, control response, etc., in which the response time and reach the steady-state time reduced by 0. 02 s and0. 08 s respectively. Therefore, the proposed improved CS optimization neural network control strategy can effectively improve the AMT shift quality.
引文
[1]张炳力,汤波,周德明,等.基于AMT的自动换挡协调控制策略研究[J].合肥工业大学学报(自然科学版),2016,39(9):1178-1183.
[2]张宽,吴训成. AMT换挡电机调速系统模糊PID控制[J].上海工程技术大学学报,2013,27(1):24-28.
[3]申业,吴光强,罗先银. AMT换挡电机精确跟踪控制[J].汽车技术, 2014(1):24-28.
[4]黄旭.基于SPWM的AMT选换挡电机控制系统研究[D].长春:吉林大学,2011:7-27.
[5]高智,钟再敏,孙泽昌.电动AMT选换挡电机执行机构位置最优控制[J].汽车工程,2011,33(2):133-137.
[6]丁然然.基于模糊规则的AMT换挡执行机构位置的高精度控制[D].上海:上海交通大学,2008:25-39.
[7] YANG X S, DEB S, FONG S,et al. Swarm intelligence:today and tomorrow[C]//Proceedings of the 2016 3rd International Conference on Soft Computing&Machine Intelligence. Washington, DC:IEEE Computer Society, 2016:219-223.
[8] KRISHNAN K.永磁无刷直流电机及其驱动技术[M].柴凤,等译.北京:机械工业出版社,2015:374-377.
[9] YANG X S, DEB S. Cuckoo search via Lévy flights[C]//Proceeings of the 2009 World Congress on Nature&Biologically Inspired Computing. Piscataway:IEEE, 2009:210-214.
[10] YANG X S, DEB S. Engineering optimisation by cuckoo search[J]. International Journal of Mathematical Modelling and Numerical Optimisation, 2010, 1(4):330-343.
[11] YANG X S, DEB S, KARAMANOGLU M, et al. Cuckoo search for business optimization applications[C]//Proceedings of the2012 National Conference on Computing and Communication Systems. Piscataway:IEEE,2012:1-5.
[12] YANG X S, DEB S. Multi-objective cuckoo search for design optimization[J]. Computers and Operations Research, 2013,40(6):1616-1624.
[13] EL-SAMAHYA A A, SHAMSELDIN M A. Brushless DC motor tracking control using self-tuning fuzzy PID control and model reference adaptive control[J]. Ain Shams Engineering Journal,2018, 9(3):341-352.
[14] CHEKROUN M, ZERIKAT A, MECHERNENE N. Novel observer scheme of fuzzy-MRAS sensorless speed control of induction motor drive[J]. Journal of Physics:Conference Series, 2017, 783(1):012045.
[15]付光杰,林雨晴,牟海维.神经网络预测在无刷直流电机调速中的应用[J].自动化仪表,2017,38(4):9-12.
[2]张宽,吴训成. AMT换挡电机调速系统模糊PID控制[J].上海工程技术大学学报,2013,27(1):24-28.
[3]申业,吴光强,罗先银. AMT换挡电机精确跟踪控制[J].汽车技术, 2014(1):24-28.
[4]黄旭.基于SPWM的AMT选换挡电机控制系统研究[D].长春:吉林大学,2011:7-27.
[5]高智,钟再敏,孙泽昌.电动AMT选换挡电机执行机构位置最优控制[J].汽车工程,2011,33(2):133-137.
[6]丁然然.基于模糊规则的AMT换挡执行机构位置的高精度控制[D].上海:上海交通大学,2008:25-39.
[7] YANG X S, DEB S, FONG S,et al. Swarm intelligence:today and tomorrow[C]//Proceedings of the 2016 3rd International Conference on Soft Computing&Machine Intelligence. Washington, DC:IEEE Computer Society, 2016:219-223.
[8] KRISHNAN K.永磁无刷直流电机及其驱动技术[M].柴凤,等译.北京:机械工业出版社,2015:374-377.
[9] YANG X S, DEB S. Cuckoo search via Lévy flights[C]//Proceeings of the 2009 World Congress on Nature&Biologically Inspired Computing. Piscataway:IEEE, 2009:210-214.
[10] YANG X S, DEB S. Engineering optimisation by cuckoo search[J]. International Journal of Mathematical Modelling and Numerical Optimisation, 2010, 1(4):330-343.
[11] YANG X S, DEB S, KARAMANOGLU M, et al. Cuckoo search for business optimization applications[C]//Proceedings of the2012 National Conference on Computing and Communication Systems. Piscataway:IEEE,2012:1-5.
[12] YANG X S, DEB S. Multi-objective cuckoo search for design optimization[J]. Computers and Operations Research, 2013,40(6):1616-1624.
[13] EL-SAMAHYA A A, SHAMSELDIN M A. Brushless DC motor tracking control using self-tuning fuzzy PID control and model reference adaptive control[J]. Ain Shams Engineering Journal,2018, 9(3):341-352.
[14] CHEKROUN M, ZERIKAT A, MECHERNENE N. Novel observer scheme of fuzzy-MRAS sensorless speed control of induction motor drive[J]. Journal of Physics:Conference Series, 2017, 783(1):012045.
[15]付光杰,林雨晴,牟海维.神经网络预测在无刷直流电机调速中的应用[J].自动化仪表,2017,38(4):9-12.