Collaborative Filtering Recommender System for Timely Arrival Problem in Road Transport Networks Using Viterbi and the Hidden Markov Algorithms

Abstract

In this study, a timely arrival recommender system (TARS) using Viterbi and hidden Markov Model (HMM) was developed. Ratings from current road users were used as inputs and trained to provide recommendations to prospective road users on the best routes to follow to get to their destinations from any source in time. The system was deployed on Android devices and iPhones with Google map. Real time data on current road conditions were collected from twenty-one (21) bolt drivers in Calabar Metropolis traversing various routes from Unical to Watt Market. The system could calculate arrival time in km/h, generate nearest nodes on each route, detect routes with free or congested traffic flow, and then recommend the best route in real time to users for timely arrival. The application, if adopted, can aid road users to save time, cost, and reduce stress on both humans and the vehicles used.