結合類神經網路與反應曲面法於多軸轉向聯結車轉向控制

本文之目的在研究多軸轉向聯結車轉向控制器，針對車輛於載重變化、高低速轉彎及其他參數變化下，對車輛穩定性之影響進行分析。採用的全聯結車為曳引車及拖曳車皆具可轉向之前輪軸，與傳統全聯結車轉向方法不同。依四輪可轉向轎車研究結果，首先推導多軸轉向全聯結車數學模式，接著進行最佳控制法則的推導；當前述控制因素改變或參數發生變動時，藉由訓練類神經網路使得控制器參數在強健穩定範圍調整，確保車輛控制器之強健穩定性，並結合反應曲面法求控制器參數最佳解。最後利用MATLAB軟體進行模擬分析，研究結果顯示，此自調強健控制器可在短時間內縮減因變數所造成的誤差，且調整至符合系統規範要求。將此控制器與最佳控制法則控制簡化的系統比較後，系統偏差率從9.7589到8.7375，有效改善10.5%。

關鍵字

多軸轉向聯結車；類神經網路；反應曲面法

並列摘要

The purpose of this content was studying The controller of a multi-axle-steering tractor and trailer that analyzed effects of a vehicle’s roadability which aimed the vehicle’s changes at loading, high and low speedy turning and other parameters. The tractor-full trailer adopted was formed by a tractor and a trailer what both had the front axles that were able to make turning different form the way of traditional tractor-full trailer. According to the consequence of investigating a four-wheel-turning compact, deriving firstly the multi-axle-steering model, and then proceeding the derivation of the optimal control method which used the trained artificial neural network to make the parameter of the controller adjust within the range of the robust stability for assuring the robust stability of a vehicle’s controller and combining the response surface methodology to request the best solution of the controller when the control factors or parameters stated above changed. At last, applying MATLAB software to proceed simulation analysis. The initial studying result showed that the automatic adjusted robust controller could shorten in short time the error caused by variables, and it adjusted to qualify the request from the system. Comparing this controller with the system of simplified control of the optimal control method, the error of the system became 8.7375 from 9.7589, it improved 10.5%.

並列關鍵字

Tractor-Full Trailer ； Artificial neural network ； Response surface Method

參考文獻

7. Der-Ho Wu, “A Theoretical Study for Yaw/Roll Motions of Multiple Steering Articulated Vehicle”.

8. D. de Bruin and P.P.J van den Bosch, “Modeling and control of a double articulated vehicle with four steerable axes”, Proceedings of the American Control Conference San. Diego. California, June 1999.

9. Lu Qiang and Wang Huiyi, “Identification and Control of Four-Wheel-Steering Vehicles Based on Neural Network”, IEEE.

10. C. Y. Seong and B. Widrow, “Neural dynamic optimization for control systems- Part I: Background,” IEEE Trans. Syst., Man, Cybern. B, vol. 31, pp. 482-489, Aug. 2001.

11. C. Y. Seong and B. Widrow, “Neural dynamic optimization for control systems- Part II: Theory,” IEEE Trans. Syst., Man, Cybern. B, vol. 31, pp. 490-501, Aug. 2001.

國際替代計量

結合類神經網路與反應曲面法於多軸轉向聯結車轉向控制

全文下載

主題瀏覽