ANN modelling for service quality assessment of two-wheelers at signalised intersections

Motorised two-wheelers (MTWs) are very popular in Asian countries due to their affordability, compact size and easy manoeuvrability. Seepage of MTWs is a prevalent phenomenon in diverse traffic environments, in which MTWs use the gap left by larger vehicles to obtain a position at the front of the queue. In the current study, the influence of seepage flow on the delays experienced at a signalised intersection was investigated. The existing service quality assessment methods primarily consider cars as the representative mode of transportation. However, understanding the importance of MTWs in developing countries, the present study focuses on modelling the service quality evaluation of MTWs using artificial intelligence (AI) techniques, such as the artificial neural network (ANN), at signalised intersections. A MTW level of service (MLOS) model was developed using this estimated delay of MTWs, along with other traffic flow parameters, intersection geometry and built-in environment parameters. The predicted MLOS scores found using the best-fitted ANN-based MTW model by optimising the sum of square error are classified into six service classes (A–F) using the DIANA clustering technique. The selected MLOS model shows a high precision score of 0.813. Finally, a sensitivity analysis was performed to prioritise the key strategies for MLOS improvement.