Road Traffic Prediction and Optimal Alternate Path Selection Using HBI-LSTM and HV-ABC

Abstract

Objectives: The objective of this work is to monitor and manage the traffic flow, so an Intelligent Transportation System (ITS) is developed that comprises the fundamental information of the real-time traffic flow. Methods: For reducing road Traffic Congestion (TC), this paper proffers an efficient traffic prediction framework and the optimal alternate route selection. The conversion of videos (from surveillance camera) into frames is done, and then pre-processing occurs. Then, for recognizing the traffic on the roadways, the background elimination utilizing Gaussian Mixture Model (GMM) is performed. Next, for identifying the vehicle motion, Motion Estimation (ME) utilizing the Virtual loop-based Lucas-Kanade (VLK) technique is performed. Utilizing the You Only Look Once (YOLO) technique, the frames are segmented centered on the estimated motion for identifying the type of objects on the road. Then, for classifying the traffic centered on the number of objects in the segmented frames, the H-detach optimized Bidirectional Long Short Term Memory (HBI-LSTM) is utilized. The traffic is classified by the classifier as heavy traffic, medium traffic, and low traffic. Findings: Utilizing the Horizontal Vertical cross-search appended Artificial Bee Colony (HV-ABC) optimization algorithm, the optimal alternate paths are chosen from different other routes if the traffic is heavy or medium. Novelty: The experimental outcomes demonstrate that the other top-notch models are outperformed by the proposed framework. Keywords: Traffic Congestion, Surveillance Videos, Noise Removal, Motion Estimation, Path Selection, Artificial Bee Colony (ABC) Optimization