Position prediction based method for detecting the track center of the Smart car

Track center detection is a key step of the Smart car traveling along the track, and its accurate detection provides necessary information for direction and speed controlling. This paper proposes a simple and effective position prediction based method for detecting the track center of the Smart car, which consists of the following four steps: (1) Initialization; (2) Position prediction; (3) Position refining; (4) Gradient determining. Experiments prove that the tract center detecting method proposed in this paper has a high detecting accuracy and is of high efficiency which can meet the time request in real-time application.