An Enhanced Taxi Demand Perception System Leveraging Fusion and Automated Sensor Integration

Abstract

Academia has recently focused on taxi demand prediction, seeing its potential in intellectual transference systems. Older methodologies often overlooked nuanced journey conditions, mainly forecasting from origin locations. This approach lacks efficiency, disregarding demand dynamics between origins and destinations. The research introduces taxi origin-destination demand prediction, leveraging mechanical automation. The authors aim to anticipate future demand across all potential area pairings, acknowledging complex location interplay. A crucial challenge is efficiently collecting diverse contextual data for effective analysis. They employ a sophisticated mechanical automation system integrating deep neural networks (DNNs) to classify journey starting and ending points, outperforming traditional methods in accuracy and performance. Through extensive testing on large-scale datasets, the DNN-based system excels in predicting taxi demand. Leveraging advanced technologies like mechanical automation, the authors pave the way for more efficient transportation systems.