OTONOM ARAÇ TASARIMI PROJESİ

 Click on figures for VIDEO Presentations

- VIDEO SUNUMLAR ICIN RESIMLERE TIKLAYINIZ -

Abstract—In this paper we describe the design and prototype of a low cost and efficient self operating vehicle. In this implementation we realized an image processing algorithm, a decision-making algorithm, a user interface and the necessary hardware designed to the control the vehicle. This system can be enhanced and used as a driver’s assistance, or as a standalone automated vehicle control system.

Index Terms— Automated vehicle control system, Image processing, Intelligent vehicle, Lane detection.

A screenshot in the image processing is given in Fig. 5(a). We can see from Fig. 5(b) that our scanning region is determined quite nicely, since there are no edges visible other than lane sides.

Fig. 5 (a). Original Image Coming From Capture card

Fig. 5 (b).First Processed Image, All edges on the scanning region is visible.

Fig. 5 (c).Resulting Image

In Fig. 5(c). the final image is illustrated where lane sides are extracted further by averaging the edges. In addition, 3 points of the road pattern is found in mid-points of the lane in Y axis.

The terminal program is written in C++. The user interface is described in Fig. 7. A picture of the prototype vehicle is illustrated in Fig. 8.

Fig. 7. Screenshot of the Terminal Program

Fig. 8. The Prototype Vehicle with the camera mounted on top.

In this work, we explain the framework of a self operating vehicle which can find the road information from a camera attached and guide itself in a lane. The system relies on the brightness levels of the improved edge detected road information and efficiently extracts this information.

In our experiments we used a Celeron 1,7GHz CPU, 512 MB DDR RAM for the image processing computations. This CPU completes the processing of a frame approximately in 10ms, with a video frame rate of 29,97 fps. In future, some improvements like lane changing, pedestrian recognition and obstacle avoidance will be added to algorithm. The wireless system is used to relinquish the processing power to a more advanced unit with an added user interface. With this processing capability, the wireless environment also provides an excellent development platform for our future improvements.

The image processing performance has been improved by enhancing the edge detection algorithm. In image processing, our algorithm is not yet capable of adjusting the ymin, ymax values automatically and user intervention is sometimes necessary. In future works we will enhance the algorithm to make the y axis adjustments automatically, and more robust to external conditions such as shadows on the road, broken lanes etc. Also according to the shape of the road, the vehicle can be made to adjust its speed automatically in future versions.

For a real self operating car this system can be realized by implementing the processing units in a powerful embedded system. This system can be enhanced and used as an assistance system for a driver, or can be used as a standalone automated vehicle control system.

References

[1] D. Özgünüer et al., “The OSU demo ’97 vehicle”, Proc. IEEE Conf. Intelligent Transportation Systems, 1997, pp. 502–507.

[2] C. Thorpe, T. Jochem, and D. Pomerleau, “The 1997 automated highway free agent demonstration,” in Proc. IEEE Conf. Intelligent Transportation Systems, 1997, pp. 496–501.

[3]Richard W. Wall; Jerry Bennett; Greg Eis; Kevin Lichy; Elizabeth Owings, “Creating A Low-Cost Autonomous Vehicle”, IECON 02 [Industrial Electronics Society, IEEE 2002 28th Annual Conference], Vol. 4, pp. 3112 – 3116, 5-8 Nov. 2002.

[4]     A. Kutami; Y. Maruya; H. Takahashi, A. Okuno, “Visual Navigation of Autonomous On-Road Vehicle”, Proceedings Intelligent Robots and Systems '90. 'Towards a New Frontier of Applications', IROS '90, Vol. 1, pp. 175 - 180, 3-6 July 1990.

[5]     Yuan Shu; Zhen Tan; “Vision Based Lane Detection in Autonomous Vehicle”, Intelligent Control and Automation, 2004. WCICA 2004, Vol. 6, pp. 5258-5260, 15-19 June 2004.

[6]     I. Masaki; S.Decker; A.Gupta; B.K.P. Horn;, H-S. Lee; D. A. Martin; C.G. Sodini; J. K. Whilte;  J. L. Wyatt, Jr. “Cost-Effective Vision Systems For Intelligent Vehicles”, Intelligent Vehicles '94 Symposium, pp. 39-43, 24-26 Oct. 1994.

[7]     Kuo-Yu Chiu; Sheng-Fuu Lin; “Lane Detection using Color-Based Segmentation”, Intelligent Vehicles Symposium, 2005, pp. 706-711, 6-8 June 2005.

[8]     Keith A. Redmill; Srinivasa Upadhya; Ashok Krishnamurthy; Ümit Özgüner; “A Lane Tracking System for Intelligent Vehicle Applications”, IEEE Intelligent Transportation Systems, 2001, pp. 273-279, 25-29 Aug. 2001.

[9]     Natalia Kazakova; Martin Margala, Nelson G. Durdle; “Sobel Edge Detection Processor For a Real-time Volume Rendering System”, Circuits and Systems, 2004. ISCAS '04. Proceedings of the 2004 International Symposium on Circuits & Systems, Vol. 2, pp. II - 913-916, 23-26 May 2004.