Improvement of reliability of environmental recognition on automatic driving system by using data of objects
Naomi Hashimoto, M Omae, K. Komoriya, Hiroshi Shimizu · 2005
This study proposes a method of environmental recognition by using detail position data of objects in the course such as white lines, road shape, light poles, road signs and so on. The proposed method realizes efficient detection of surrounding objects based on object's position data. Moreover, it estimates the vehicle's position and detects obstacles from visionary information and object position data. Experimental results show that the vehicle can continue to run under GPS-based automatic driving, even when the GPS information is not available by means of the proposed method. Also, these results show that the vehicle can detect obstacle with high probability.