A prototype application was developed using Threshold Based algorithm technique to classify the events that the driver does during driving, the prototype system design consists of a scoring system to assist the driver with the events detected, the user can pick up the events and explore the feedback to improve their driving skills. The user can also navigate to different trips to check improvements, and the progress bars and graph view provide another vision of driving performance; the prototype was guided by the methodologies found during this research. Most of the specific technique was developed using the thresholds established by (Vavouranakis et al., 2017)  research. It establishes a calibration filter for accelerometer signals and processes using conditionals, some of the challenges found are related, to the smartphone positioning that required an extra system for repositioning which was not implemented, calibration process which works well although is very sensible and this varies depending on the DBA schedule task, it is concluded that the smartphone environment is not predictable when is used different mechanisms at the same time. For example, the system is implemented using the accelerometer’s x and y-axis. However, it was tried also to use the gyroscope and magnetometer to use as a threshold for the Pitch and Roll. However, extracting those values at the same time becomes unpredictable in the sampler system, which checks the behaviour events every certain time. Further work will base on extracting events in a controllable environment and analyse them with a machine-learning model.

SCORING DRIVING PROTOTYPE ANDROID APPLICATION

Calibrating preprocessing system Difference between filtered and non filtered:

NO Filter is applied

Filter is applied