Multimodal transport model: Enhancing collaboration among mobility sharing schemes by identifying an optimal transit station

In order to return urban places to people and create more livable cities, establishing a sustainable, environment-friendly transport system is the need of the hour. Mobility sharing schemes, particularly via public transport systems, are known to reduce traffic congestion by removing a large set of social vehicles occupying the road. The impact of mobility sharing can be enhanced by enabling effective collaboration among different schemes. However, for seamless integration of public transport vehicles into urban transport fabric, dissemination of real-time information concerning vehicle arrivals and their future availability is essential. In this paper, a new model is proposed to facilitate effective collaboration among mobility schemes involving public and private vehicles by forming a multi-modal transport network. The proposed approach identifies a single transit station along the passenger's trajectory which could be used to switch to a public transport vehicle going towards the desired destination. Machine learning algorithms like SVR and ANN are employed to predict arrival times, availability and associated waiting times of buses at each transit station along the route aiding this selection. The proposed model was applied on a comprehensive archived traffic data set and its performance in predicting various parameters was evaluated by comparing the estimated values against the actual ground-truth data.