AI-based method for determining vehicles with illegal muffler from pass-by noise

In Japan, illegal noisy vehicle is controlled by the result of a stationary noise test. The stationary noise test used does not have a simple correlation with pass-by noise due to differences in vehicle conditions. If illegal vehicles can be identified by pass-by noise, it will lead to more effective roadside inspections. Therefore, this study investigated whether it is possible to identify vehicles that exceed the limit value of the stationary noise test limit from pass-by noise using AI. First, we investigated the method using a single sound level meter, and then considered the use of a microphone array to make the method more effective in road traffic flow.