An AI-powered traffic congestion prediction & route recommendation in Uganda using recurrent neural networks

Abstract

This study was about An AI-powered Traffic Congestion Prediction & Route Recommendation in Uganda using Recurrent Neural Networks that will help Kampala and other cities in Uganda, our main aim was to develop a mobile application powered by an AI model to predict traffic congestion with high accuracy and suggest alternative routes in real-time for Ugandan roads using RNNs, we also evaluated the effectiveness of the proposed application in terms of prediction accuracy and route recommendation efficiency, testing and validating the AI-powered system for predicting traffic congestion was also done. The methods we used for data collection and analysis were interviews, Questionnaires, observation Python script, Qualitative and quantitative analysis. For application design we used a use case diagram and user interface design. Google sheets, Visual studio code, google sheets, Draw.io and colabs were used during application implementation, the language used were Python, Dart flutter. Based on the development and testing of the traffic congestion prediction app, the following recommendations are made: Regularly update the AI model with new data to improve prediction accuracy, Implement a feedback mechanism to gather user input and incorporate suggestions for future updates, Explore incorporating additional data sources, such as weather information, public transportation data, and social media feeds, to enhance prediction accuracy and Develop features to detect and alert users about traffic incidents, such as accidents or road closures. Potential areas for future research can handle Integration of advanced machine learning techniques, such as deep learning and reinforcement learning, Development of real-time traffic incident detection and notification systems, Integration with connected vehicle technology to enhance data collection and prediction accuracy, Incorporating additional data sources, such as weather information, public transportation data, and social media feeds, to enhance prediction accuracy.