A Supervisory Learning Based Two-Wheeler Drive Pattern 2015-01-0221
The life of a two-wheeler and its parts depend much on its usage during its years of running. The quality of its parts determine the life and efficiency; however the handling of the two-wheeler also plays a major role in estimating it's life and other performance parameters. Hence, it is beneficial to have an efficient system which enhances the life of a two-wheeler and also gives better mileage.
This paper constitutes an efficient drive pattern system which addresses the above. This system consists of two main parts: the data collection system and an Android-based mobile application which runs on a mobile phone. The data collection system collects data from various sensors on the vehicle and then the data is processed and sent to the mobile phone of the rider during the run time of the two-wheeler. The application uses this data to depict useful information like drive pattern and various indicators.
In addition to this, a novel communication protocol was developed to transfer the real-time data from vehicle to the mobile phone application. A machine learning  technique called supervised learning   was used along with a hard coded data set to generate suggestions useful to the rider. These suggestions would help in improving the drive pattern of the rider. These suggestions are displayed as results in the mobile application. The above described system was developed as a proto-type and tested on a two-wheeler.