关键词: 麦克风阵列, MVDR算法, 卷积神经网络, 智能控制, 隧道照明节能

Abstract:

: The radar, video, induction coil, etc. are generally used for vehicle perception of conventional tunnel lighting energy-saving technology, which cannot meet the requirements of low cost, high precision and easy installation and maintenance. Therefore, a method and system with precision lighting for energy-saving of highway tunnel lighting based on microphone array vehicle detection is proposed. The specific schemes include: (1) The existing contact sensing technology is replaced by low-cost and non-buried vehicle detection device based on microphone array. (2) The improved MVDR algorithm is used to de-noise and de-reverberate the signals picked up by microphone array in audio devices, and the accurate and sectioned vehicle sense in tunnel is realized based on the convolutional neural networkbased traffic incident recognition method. (3) A multi-information cooperative control method based on environmental sensing is proposed, which intelligently adjusts the illumination brightness and time of the tunnel lights by multiinformation linkage of stepless dimming controller according to the environment and vehicle driving information in the tunnel. The experimental analysis shows that the proposed microphone array audio vehicle detection technology has a detection accuracy of more than 98% for single event and a detection accuracy of more than 95% for mixed events. The engineering practice shows that the comprehensive energy consumption of system mentioned-above is 20% less than that of conventional LED stepless dimming technology, it is worth popularizing.

Key words: microphone array, MVDR algorithm, convolutional neural network, intelligent control, energy saving of tunnel lighting