Markov Process Based IoT Model for Road Traffic Prediction

IoT based network models play a vital role in estimating and predicting the behavior of the movement of vehicles. Drastic changes in economics lead to high usage of personal vehicles by the common people. This in turn lead to traffic jams, road blockages and make areas accident prone. It becomes a hectic task to control the system and handle the system smoothly. In this paper, the authors proposed IoT-based network model for predicting traffic in the busy areas by using binary Markovian events. Initially, a hidden Markov models (HMMs) forward algorithm is proposed to schedule and utilize available resources to the devices with extreme possible activation probabilities. During the process of estimating performance, a regret matrix is initiated to check how many transmission slots are being wasted. Secondly, a model has been proposed to optimize the Age of Information (AoI) by maintaining the regret matrix as least as possible. Finally, recapitulate algorithm is projected to assess activation probabilities for online-learning predictions to the real-time traffic problems. To outperform the proposed model simulation results are presented to show the efficiency of the algorithm.