ANN based traffic congestion analysis applied to parking recommendation system for electric three-wheelers

Abstract The increasing adoption of Electric Vehicles (EVs) is transforming transportation systems with an emphasis on sustainability. In the Asia-Pacific region, Electric Three Wheelers (E3W) have gained significant market share due to their affordability and eco-friendly nature. However, the rise in E3W usage has also led to road congestion and parking challenges, particularly in suburban and rural areas. This paper presents a smart roadside parking recommendation system that utilizes a proposed Artificial Neural Network (ANN) to forecast traffic status and recommend nearby parking options. The system uses vehicle dynamics data such as voltage, current, and timestamps collected via an IoT-based data logger integrated with a cloud-based and map-based visualization interface. By analyzing the collected data, the system predicts real-time traffic status for parking recommendations, enabling an efficient parking solution that saves both time and energy. The experimental result demonstrates that the proposed ANN configuration gives the most optimal result in terms of accuracy, loss, MAE, MSE, RMSE, and R Square on purely unseen data in real-time scenarios. Moreover, the trained ANN model effectively categorizes the parking recommendation as ‘Recommended,’ ‘Waiting State,’ or ‘Not Recommended’ for specific parking areas. To validate the model performance, experiments are conducted in two distinct regions with unique routes and traffic patterns. The result is visualized through map API and demonstrates optimal parking recommendations as compared to real-world traffic scenarios, thereby improving decision-making for E3W drivers.