Dynamic Path Planning Based on Traffic Flow Prediction and Traffic Light Status

Traffic flow prediction and path planning are crucial components of effective intelligent transportation systems research. The intelligent transportation system can optimize vehicle driving routes by utilizing predicted traffic flow data for each road segment and considering the periodic changes in traffic light patterns at intersections. However, most studies on traffic flow prediction have overlooked the frequency domain information of traffic flow sequences, resulting in a lack of effective modeling of this vital frequency domain information. Furthermore, existing path planning approaches only consider factors such as traffic density and road length in decision-making, neglecting the influence of traffic light status on vehicle travel time. We propose a traffic flow prediction model called mWDN-LSTM-ARIMA to address these issues, incorporating frequency feature extraction and residual testing. Additionally, we present a path planning method that leverages the traffic flow predictions from mWDN-LSTM-ARIMA and takes into account the periodic transformation law of traffic lights at urban intersections. Our experimental results validate the effectiveness of the proposed approach in reducing the average travel time and waiting count of vehicles.