转台频域性能测试软件开发及控制技术研究
|
摘要
三轴仿真转台用来模拟飞行器在空中的姿态和运动,是半实物仿真系统中的重要设备之一,转台性能的优劣,直接关系到半实物仿真试验结果的精度和置信度。因此,高性能转台对航空航天工业和国防建设的发展具有重要的意义。
本文针对一新型三轴仿真转台在半实物仿真系统中的应用情况,设计了其动态性能测试软件,分析了其对仿真结果的影响,并且对动态神经网络在转台实时控制中的应用进行了研究。
首先,实现了仿真计算机和转台控制计算机之间的以太网络通信和串口通信。程序采用VC++语言设计,模块化结构,通过调用VC++提供的MFC类库实现通信,不受操作系统版本限制,可方便地嵌入到不同的仿真程序中,完成不同对象的半实物仿真试验。
其次,开发了基于串口通信的转台动态性能测试软件,它分为两部分—单一频率信号测试和频率特性分析测试。该软件具有友好的用户界面,该软件系统具有数据存储功能,保存的数据文件为神经网络控制提供了样本数据。
最后,研究了动态神经网络在转台实时控制中的应用。在选择了一种适合转台控制的神经网络的基础上,详细阐述了神经网络前馈控制器的设计过程,然后利用神经网络前馈控制方法对转台进行非线性补偿。并通过仿真试验验证了控制方案的可行性。
Three axixs simulation table,which is used to simulate the pose and movement of the aerocraft in the air, is one of the most important equipments in Hardware-In- The-Loop Simulation System. The performance of the table is directly related to the productions’capability and precision. So the table with high performance is of great social and economical benefit to the development of aviation industry and construction of national defense.
The paper mainly research on the application of the new simulation table in Hardware-In-The-Loop Simulation System. It designs the test software about the dynamic function of the table, and analyses its influence to the results of simulation. It also research on the application of dynamic neural networks in table system.
At first, the paper realizes the communication of ethernet network and serial port between the simulation computer and the table control computer. The model-structure programs designed by VC++ language implement communication by using the usual MFC without any limitation of operation systems. It can insert into different simulation programs conveniently, and complement the Hardware-In -The-Loop Simulation experiment on different objects.
Secondly, the paper develops a kind of software which is used to test the table’s dynamical capability based on the serial port communication. It divides into two parts- the single frequency testing and frequency characteristic testing. The software with friendly users’interface has the function of data storage, and offers the data for the control of neural networks.
Lastly, an application of dynamic neural networks in table system is investigated. Based on the choosing of suitable neural networks for the table system, the design process of the neural network feed-forward controller is described in detail .Then, the table is nonlinearly compensated through the method of neural network feed-forward control and simulation is applied to confirm the feasibility of the controller.
本文针对一新型三轴仿真转台在半实物仿真系统中的应用情况,设计了其动态性能测试软件,分析了其对仿真结果的影响,并且对动态神经网络在转台实时控制中的应用进行了研究。
首先,实现了仿真计算机和转台控制计算机之间的以太网络通信和串口通信。程序采用VC++语言设计,模块化结构,通过调用VC++提供的MFC类库实现通信,不受操作系统版本限制,可方便地嵌入到不同的仿真程序中,完成不同对象的半实物仿真试验。
其次,开发了基于串口通信的转台动态性能测试软件,它分为两部分—单一频率信号测试和频率特性分析测试。该软件具有友好的用户界面,该软件系统具有数据存储功能,保存的数据文件为神经网络控制提供了样本数据。
最后,研究了动态神经网络在转台实时控制中的应用。在选择了一种适合转台控制的神经网络的基础上,详细阐述了神经网络前馈控制器的设计过程,然后利用神经网络前馈控制方法对转台进行非线性补偿。并通过仿真试验验证了控制方案的可行性。
Three axixs simulation table,which is used to simulate the pose and movement of the aerocraft in the air, is one of the most important equipments in Hardware-In- The-Loop Simulation System. The performance of the table is directly related to the productions’capability and precision. So the table with high performance is of great social and economical benefit to the development of aviation industry and construction of national defense.
The paper mainly research on the application of the new simulation table in Hardware-In-The-Loop Simulation System. It designs the test software about the dynamic function of the table, and analyses its influence to the results of simulation. It also research on the application of dynamic neural networks in table system.
At first, the paper realizes the communication of ethernet network and serial port between the simulation computer and the table control computer. The model-structure programs designed by VC++ language implement communication by using the usual MFC without any limitation of operation systems. It can insert into different simulation programs conveniently, and complement the Hardware-In -The-Loop Simulation experiment on different objects.
Secondly, the paper develops a kind of software which is used to test the table’s dynamical capability based on the serial port communication. It divides into two parts- the single frequency testing and frequency characteristic testing. The software with friendly users’interface has the function of data storage, and offers the data for the control of neural networks.
Lastly, an application of dynamic neural networks in table system is investigated. Based on the choosing of suitable neural networks for the table system, the design process of the neural network feed-forward controller is described in detail .Then, the table is nonlinearly compensated through the method of neural network feed-forward control and simulation is applied to confirm the feasibility of the controller.
引文
[1] 陈明恩,蓝牙技术在无人机半实物仿真系统中的应用研究,[硕士学位论文],南京,南京航空航天大学,2006
[2] 孟宪强,程瑞琪,蒋兆远,基于以太网的 VC6.0 网络通信编程研究,兰州交通大学学报,2000,23(4):74～77
[3] 雷斌,杨建华,黄超,何斌,Visual C++6.0 网络编程技术,北京,人民邮电出版社,2000:21～22
[4] 何华,用 VC6.0 MFC 实现多线程网络编程,现代计算机,2000,99:62～64
[5] 孙亚飞,测试仪器发展概述,仪器仪表学报,2003.10:78～81
[6] 赵会兵,虚拟仪器技术规范与系统集成,北京,清华大学出版社,2003.5:123～124
[7] 张毅,周绍磊,虚拟仪器技术分析与应用,北京,机械工业出版社,2004.2:92～96
[8] 陈科,傅承毓,虚拟动态信号测试仪的实现与应用,光电工程,1998,25:88～92
[9] 汤宝平,何辉,秦树人等,基于 PC 总线虚拟动态测试分析仪软件的设计,震动、测试与诊断,2000,20(2):82～88
[10] 周舒梅,动态信号分析和仪器,北京,机械工业出版社,1990:45～47
[11] 刘思程,余彬.,MFC 类库实现多线程的调度和处理,计算机时代,2003,8:22～24
[12] Jeff Prosise,MFC Windows 程序设计,北京,清华大学出版社,2001:884～916
[13] 冯美霞,多线程应用程序的同步技术,计算机应用,1999(1):28～32
[14] CHARLES PETZOLD,Windows 程序设计(第五版),北京,北京大学出版社,1999:234～256
[15] 胡祐德,马东升,张莉松,伺服系统原理与设计,北京,北京理工大学出版社,1999:122～123
[16] 李小璐,动态递归网络在转台控制中的应用,[硕士学位论文],南京,南京航空航天大学,2006
[17] 徐丽娜,神经网络控制,哈尔滨,哈尔滨工业大学出版社,1999:65～68
[18] Lin T, Horne BG, Tino P, et al, Learning long term dependencies in NARX recurrent neural networks, IEEE Transactions on Neural Networks, 1996, 7(6):1329～1338
[19] WANG Lei, ZHOU Guoxing, WU Qidi, Hopfield neural network based identification and control of induction motor drive system, Proceeding of 14th IFAC World Congress, 1999, 279～283 and 285～289
[20] R.Yasdi, Prediction of road traffic using a neural network approach, Neural Comput & Applic, 1999, 8:135～142
[21] Tsoi AC, Back AD, Locally recurrent globally feed forward networks: A critical review of architectures, IEEE Transaction on Neural Networks, 1994, 5(2): 229～239
[22] Elman J, Finding structure in time, Cognitive Science, 1990, 14: 179～211
[23] 丛爽,高雪鹏,几种递归神经网络及其在系统辨识中的应用,系统工程与电子技术,2003,25(2):194～197
[24] Moody J E, The effective number of parameters: An Analysis of Generalization andRegularization in Nonlinear Learning System, NIPS 4, 1992, 847～854
[25] Pham D T, Oh S J, Adaptive control of dynamic systems using neural networks, Proc IEEE-SMC Conference, Systems Engineering in the Service of Humans, 1993, 4: 97～102
[26] MD. Minkova, D. Minkov, J.L. Rodgerson, et al, Adaptive neural speed controller of a dc motor, Electric Power Systems Research, 1998, 47: 123～132
[27] Jun Oh Jang, Implementation of indirect neuro-control for a nonlinear two-robot MIMO system, Control Engineering Practice, 2001, 9: 89～95
[28] Refaat S, Ahmed and Kuldip S, Rattan, Real-time tracking control of a DC motor using a neural network, IEEE Trans. On Neural Network, 1995, 593～600
[29] Narendra Kumpati S,Kanan Parthasarathy,Gradient Methods for the Optimization of Dynamical System Containing Neural Networks,IEEE Trans on Neural Networks,1991,2 (2) 252-262
[30] Pham D T, Liu X, Dynamic system modeling using partially recurrent neural networks, Journal of Systems Engineering, 1992, 2: 90～97
[31] 丛爽,神经网络、模糊系统及其在运动控制中的应用,合肥,中国科学技术大学出版社,2001:238～243
[32] 丛爽,戴谊,递归神经网络的结构研究,计算机应用,2004,24(8),18～21
[33] 谢庆国,沈轶,万淑芸,Elman 人工神经网络的收敛性分析,计算机工程与应用,2002,6:61～63
[34] Frank L Lewis, Kai Liu, Aydin Yesildirek, Neural net robot controller with guaranteed tracking performance,IEEE Trans. On Neural Networks, 1995, 6(3): 703～715
[35] Lewis F L, Yegildirek A, Kai Liu, Multilayer neural-net robot controller with guaranteed tracking performance,IEEE Trans. On Neural Networks, 1996, 7(2): 388～399
[36] 扈宏杰,神经网络控制理论的研究及其在高精度仿真转台系统中的应用,[博士学位论文],北京,北京航空航天大学,2001
[37] Cataltepe Z, Abu-mostafa Y S, Magdon-Ismail M, No free lunch for early stopping, Neural Computation, 1999, 11: 995～1009
[38] 李新忠,电液伺服系统的神经网络非线性控制,[博士学位论文],西安,西安交通大学,1998
[39] Kawato M , Gomi H , Neural Network Control for A Closed-Loop System Using Feedback-Error-Learning,Neural Network.1993,6(6):933～946
[40] 盛守照,前向神经网络学习问题研究及其在转台控制中的应用,[博士学位论文],南京,南京航空航天大学,2005
[41] 徐立新,王常虹,庄显义,基于神经网络模型的伺服系统控制与补偿方法研究,宇航学报,1998,19(3):83～92
[2] 孟宪强,程瑞琪,蒋兆远,基于以太网的 VC6.0 网络通信编程研究,兰州交通大学学报,2000,23(4):74～77
[3] 雷斌,杨建华,黄超,何斌,Visual C++6.0 网络编程技术,北京,人民邮电出版社,2000:21～22
[4] 何华,用 VC6.0 MFC 实现多线程网络编程,现代计算机,2000,99:62～64
[5] 孙亚飞,测试仪器发展概述,仪器仪表学报,2003.10:78～81
[6] 赵会兵,虚拟仪器技术规范与系统集成,北京,清华大学出版社,2003.5:123～124
[7] 张毅,周绍磊,虚拟仪器技术分析与应用,北京,机械工业出版社,2004.2:92～96
[8] 陈科,傅承毓,虚拟动态信号测试仪的实现与应用,光电工程,1998,25:88～92
[9] 汤宝平,何辉,秦树人等,基于 PC 总线虚拟动态测试分析仪软件的设计,震动、测试与诊断,2000,20(2):82～88
[10] 周舒梅,动态信号分析和仪器,北京,机械工业出版社,1990:45～47
[11] 刘思程,余彬.,MFC 类库实现多线程的调度和处理,计算机时代,2003,8:22～24
[12] Jeff Prosise,MFC Windows 程序设计,北京,清华大学出版社,2001:884～916
[13] 冯美霞,多线程应用程序的同步技术,计算机应用,1999(1):28～32
[14] CHARLES PETZOLD,Windows 程序设计(第五版),北京,北京大学出版社,1999:234～256
[15] 胡祐德,马东升,张莉松,伺服系统原理与设计,北京,北京理工大学出版社,1999:122～123
[16] 李小璐,动态递归网络在转台控制中的应用,[硕士学位论文],南京,南京航空航天大学,2006
[17] 徐丽娜,神经网络控制,哈尔滨,哈尔滨工业大学出版社,1999:65～68
[18] Lin T, Horne BG, Tino P, et al, Learning long term dependencies in NARX recurrent neural networks, IEEE Transactions on Neural Networks, 1996, 7(6):1329～1338
[19] WANG Lei, ZHOU Guoxing, WU Qidi, Hopfield neural network based identification and control of induction motor drive system, Proceeding of 14th IFAC World Congress, 1999, 279～283 and 285～289
[20] R.Yasdi, Prediction of road traffic using a neural network approach, Neural Comput & Applic, 1999, 8:135～142
[21] Tsoi AC, Back AD, Locally recurrent globally feed forward networks: A critical review of architectures, IEEE Transaction on Neural Networks, 1994, 5(2): 229～239
[22] Elman J, Finding structure in time, Cognitive Science, 1990, 14: 179～211
[23] 丛爽,高雪鹏,几种递归神经网络及其在系统辨识中的应用,系统工程与电子技术,2003,25(2):194～197
[24] Moody J E, The effective number of parameters: An Analysis of Generalization andRegularization in Nonlinear Learning System, NIPS 4, 1992, 847～854
[25] Pham D T, Oh S J, Adaptive control of dynamic systems using neural networks, Proc IEEE-SMC Conference, Systems Engineering in the Service of Humans, 1993, 4: 97～102
[26] MD. Minkova, D. Minkov, J.L. Rodgerson, et al, Adaptive neural speed controller of a dc motor, Electric Power Systems Research, 1998, 47: 123～132
[27] Jun Oh Jang, Implementation of indirect neuro-control for a nonlinear two-robot MIMO system, Control Engineering Practice, 2001, 9: 89～95
[28] Refaat S, Ahmed and Kuldip S, Rattan, Real-time tracking control of a DC motor using a neural network, IEEE Trans. On Neural Network, 1995, 593～600
[29] Narendra Kumpati S,Kanan Parthasarathy,Gradient Methods for the Optimization of Dynamical System Containing Neural Networks,IEEE Trans on Neural Networks,1991,2 (2) 252-262
[30] Pham D T, Liu X, Dynamic system modeling using partially recurrent neural networks, Journal of Systems Engineering, 1992, 2: 90～97
[31] 丛爽,神经网络、模糊系统及其在运动控制中的应用,合肥,中国科学技术大学出版社,2001:238～243
[32] 丛爽,戴谊,递归神经网络的结构研究,计算机应用,2004,24(8),18～21
[33] 谢庆国,沈轶,万淑芸,Elman 人工神经网络的收敛性分析,计算机工程与应用,2002,6:61～63
[34] Frank L Lewis, Kai Liu, Aydin Yesildirek, Neural net robot controller with guaranteed tracking performance,IEEE Trans. On Neural Networks, 1995, 6(3): 703～715
[35] Lewis F L, Yegildirek A, Kai Liu, Multilayer neural-net robot controller with guaranteed tracking performance,IEEE Trans. On Neural Networks, 1996, 7(2): 388～399
[36] 扈宏杰,神经网络控制理论的研究及其在高精度仿真转台系统中的应用,[博士学位论文],北京,北京航空航天大学,2001
[37] Cataltepe Z, Abu-mostafa Y S, Magdon-Ismail M, No free lunch for early stopping, Neural Computation, 1999, 11: 995～1009
[38] 李新忠,电液伺服系统的神经网络非线性控制,[博士学位论文],西安,西安交通大学,1998
[39] Kawato M , Gomi H , Neural Network Control for A Closed-Loop System Using Feedback-Error-Learning,Neural Network.1993,6(6):933～946
[40] 盛守照,前向神经网络学习问题研究及其在转台控制中的应用,[博士学位论文],南京,南京航空航天大学,2005
[41] 徐立新,王常虹,庄显义,基于神经网络模型的伺服系统控制与补偿方法研究,宇航学报,1998,19(3):83～92