FALANFA: AN INTELLIGENT DRIVING APPLICATION FOR MONITORING DRIVERS’ DRIVING BEHAVIOURS

ABSTRACT

Driving is a complex and vigorous activity, it involves drivers making quick perceptions of the information associated with driving skills within an instance of time. There have been few deployments of intelligent applications for driving, but because Nigeria is a developing country in Africa there is a need to deploy this system with locally collected data, at least to prove that data extracted from this part of the world is trainable. This paper presents an intelligent driving application that monitors drivers by detecting and predicting the driver’s behaviours while on the road to curb road accidents in Nigeria. The dataset used in this research was locally collected from Nigeria while the driving events adopted for this research are braking, speeding, and safe driving. The object detection/machine learning algorithm that was used to evaluate the model is Single Shot Multibox Detection (SSD) Mobilenet. The results proved that data extracted from Nigeria is trainable. We deployed the model to mobile application, the mobile application when launched activates the camera by default enabling the system to detect and predict. The predicted values were all positive. The three driving events were all detected perfectly in real time while testing without being perverse. The system presents its predicted value in percentage, therefore showing the level of adherence to each of the driving events detected.

Keywords: Machine learning, object detection, SSD Mobilenet, Intelligent System, driving events.