以視覺辨識設計之車輛側向控制系統

A Vision-based Design Approach for Vehicle Lateral Control System

戴凱易(Kai-I Tai)

指導教授 ： 丁振聲

國立虎尾科技大學/電機資訊學院/電機工程研究所/碩士(2014年)

https://doi.org/10.6827/NFU.2014.00240

摘要

本文應用自行車模型建立系統數學模式並以視覺辨識系統作為控制策略。首先運用傳統的PD控制器以超音波做為感測器，使原型車能保持穩定的方向行進，透過Zigbee無線傳輸模組收集系統動態資料後，使用Matlab鑑別系統模式並驗證測試與實作之結果，建立正確的系統數學模型，進行系統分析與控制器設計。獲得車體正確的數學模型後，透過DaVinci DM6437數位視訊開發平台，將CCD獲取的畫面透過Canny邊緣檢測法與座標轉換得到實驗環境中的車道中心偏移量，以視覺辨識取代超音波感測側向偏移量。

在硬體架構中，原型車上控制核心為TMS320F28335，須與視訊開發平台進行系統整合。因此，本文以RS232串列通訊介面作為DM6437與F28335數據傳輸方式，將辨識結果傳送到F28335進行實控，最後由實驗結果比較狀態回授控制與傳統PD之優劣，印證所設計的控制系統其正確性與應用性。

關鍵字

側向控制 ； 自行車模式 ； 系統鑑別 ； 視覺辨識

並列摘要

This thesis presents a vision-based design approach for controlling vehicle lateral dynamics, which is derived in a bicycle-framework model car. To construct the mathematical model, the conventional PD controller is first utilized to maintain a stable driving direction. Then, the system performance is transmitted by Zigbee module to the host computer for collecting data. With the help of the Matlab System Identification toolbox, the obtained mathematic model is verified by a further comparison of the experimental results and computer simulation to guarantee its exactness. The control algorithm is developed based on the conventional pole-placement design. In particular, the vehicle lateral position to the road center line is discerned via the image processing, which is operated on a digital video platform DaVinci DM6437, instead of measurement by an ultrasonic sensor. Based on the CCD image, the deviation from the lane center is acquired by the Canny edge detection method and the coordinate transformation.

In the hardware architecture, A digital signal processor(TMS320F28335) is the main controller of the model car and combined with video development platform to establish the vision-based control system. The data communication between F28335 and DM6437 is performed through the serial transmission interface RS232. With the discerned results, the controller can modulate the right direction of the model car and keep driving safely on the road center. Finally, the correctness of the proposed control scheme is verified by the practical experiment and compared with the PD controller.

並列關鍵字

lateral control ； bicycle model ； system identification ； image processing

參考文獻

[31] Chao-Jung Chen, Hsin-Yuan Peng, Bing-Fei Wu and Ying-Han Chen, “A Real-Time Driving Assistance and Surveillance System” , JOURNAL OF INFORMATION SCIENCE AND ENGINEERING , Vol. 25 , pp. 1501 – 1523, 2009.

[1] Lin Cai, Rad and Wai-Lok , “A Genetic Fuzzy Controller for Vehicle Automatic Steering Control”, IEEE Transactions on Vehicular Technology, vol. 56, vo. 2, March 2007.

[2] R.E. Benton and D. Smith, “A static-output-feedback design procedure for robust emergency lateral control of a highway vehicle”, IEEE Trans. Contr Syst Technol, vol. 13, no. 4, pp. 618-623, 2005.

[3] S.B. Choi, “The design of a look-down feedback adaptive controller for the lateral control of front-wheel-steering autonomous highway vehicles”, IEEE Trans. Vehicular Technol, vol. 49, no.6, pp. 2257-2269, 2000.

[4] D.E. Smith and J.M. Starkey, “The effects of model complexity on the performance of automated vehicle steering controllers: controller development and evaluation” , Vehicle Syst Dyn, vol.23, pp.627-645, 1994.