AI-Powered Smart Fuel Level Indicator and Prediction of Mileage for Two-Wheelers
Most two-wheelers in India use analog fuel meters, which lack accuracy. Traditional fleet management relies on fixed conditions for fuel estimation, but real-world mileage varies due to factors like driving style, speed, and road conditions. To address this, our work develops a digital fuel measurement system using sensors and a microcontroller to accurately display the exact fuel inlet. This cost-effective solution ensures transparency and prevents fraud. Along with this it uses machine learning-based regression to predict the distance a two-wheeler can travel with available fuel. Sensors on the vehicle and fuel tank provide real-time data, enhancing accuracy in mileage prediction for efficient fleet management. The study evaluates various fuel sensors, identifies limitations in existing systems, and introduces an improved digital sensing technology for precise fuel measurement. Major Findings: The regression model achieved an R-squared of 0.94, explaining 94% of the variance in mileage, with RMSE 2.72 and MSE of 5.44 indicating high accuracy. Road conditions, traffic, and weather were the most influential factors, and residual plots confirmed model reliability, predicting mileage at 26.65 kmpl for a given input.