Analysis on the Bus Arrival Time Prediction Model for Human-Centric Services Using Data Mining Techniques.

Area of Science:

• Intelligent Transportation Systems
• Human-Computer Interaction
• Machine Learning Applications

Background:

• Public bus transportation in India faces challenges with real-time information dissemination for passengers.
• Lack of transparency regarding bus numbers and Estimated Time of Arrival (ETA) leads to passenger inconvenience.
• Bus delays due to traffic, weather, and breakdowns exacerbate the information gap.

Purpose of the Study:

• To leverage Human-Computer Interaction (HCI) principles to develop an Intelligent Transportation System (ITS) for Indian bus services.
• To enhance passenger experience by providing real-time bus tracking and accurate ETA predictions.
• To improve bus timing transparency and operational efficiency for passengers, drivers, and managers.

Main Methods:

• Development of an HCI-based web/mobile application for real-time bus tracking.
• Implementation of a real-time bus management system using Raspberry Pi, GPS, and GSM modules.
• Utilizing machine learning, specifically Support Vector Regression, for ETA prediction incorporating traffic and weather data.

Main Results:

• The developed system provides real-time bus location tracking and ETA information to passengers.
• The ETA prediction model, using Support Vector Regression, achieved a Root Mean Square Error (RMSE) of 27 seconds.
• The model demonstrated strong performance compared to other predictive algorithms.

Conclusions:

• The HCI-based system effectively addresses the lack of real-time information in public bus transport.
• Accurate ETA forecasting using machine learning significantly improves passenger satisfaction and system efficiency.
• The real-time bus management system offers a comprehensive solution for stakeholders in the transportation sector.