M.F. Shu, J.X. Zhu

Pages: 127-136

Abstract
When the traditional method is used for vehicle collision early warning, the preset conditions are not comprehensive, resulting in the problems of high false alarm rate and long early warning time, and the early warning effect is poor. Therefore, an optimization method of two vehicle collision early warning time considering driving speed is proposed. Tk103b on-board GPS locator is selected to locate the vehicle track under driving status and obtain the vehicle track information, The adaptive filtering algorithm is used to filter the vehicle positioning trajectory data to improve the positioning accuracy, fully considering different vehicle speeds, the two-way recurrent neural network is used to optimize the collision warning time of two vehicles. The experimental results show that the maximum false alarm rate is only close to 1%, and the minimum collision warning time is 0.3s.
Keywords: driving speed; collision warning; trajectory positioning; trajectory repair; GPS locator; adaptive filtering algorithm