Fuel Detection and Distance Prediction Using Multi-dimension Regression Method

Previous Next

At present, the mileage (distance coverage) for two wheeler claimed by the manufacturer is based on the fixed laboratory testing conditions. Many factors such as individual driving pattern, traffic condition, road condition, load, speed condition, petrol level, and terrain will affect the mileage of the vehicle to a greater extent. We have proposed a machine learning model to predict the mileage considering the above factors. Develop a multi-feature regression model on the past experimental driving factor data (Petrol, place, traffic conditions, load, speed, and road nature). Acquire driving factor data in real time through sensors. Apply the sensors values on a regression model to predict the mileage at real time. Machine learning is the biggest revolution of this 5th generation. The most important aspect that one expects from his very own vehicle is its mileage even when affected by many factors like road condition, the way of driving etc. But, when a machine is trained to bridge the promised mileage and the on-road value, it becomes handy and useful for a person.

Such an idea is proposed to practically identify the value of the mileage that a vehicle could possibly be able to give. This would enable the person to understand his vehicle’s performance better. There is no proper system to detect the mileage in real time condition. In the existing model, only fuel present in the fuel tank is displayed which won’t be accurate. In this project, we propose a multi-feature regression model on the past experiment driving factor. Predict the mileage in real time by fitting sensor data on a developed regression model. Linear regression is one of the fundamental models in statistics used to determine the relationship between dependent and independent variables. An extension of this model, namely multiple linear regression, is used to represent the relationship between a dependent variable and several independent variables. Developed regression model for the mileage calculation Y=13.2128+50.2500*(x1)+1.2144*(x2)+1.3155*(x3)+1.4744*(x4)+0.1335*(x5) - 0.5913(x6) . Where, as X1-X2 values will the 6 factors considered for mileage prediction. Thus a multi-feature regression model on the past experimental driving data factors is developed. Thus driving factor data in real time through sensors is acquired. The sensor values are applied on the regression model and the mileage is predicted in real time.


Voting is closed!


  • Name:
    Mahendaran Elangovan
  • Type of entry:
    Team members:
    Mahendaran E Akshay Geedan Nubail Ahamed
  • Software used for this entry:
  • Patent status: