I have been searching for literature that helps me to answer the questions, I found numerous research focusing on different smartphone sensors and different techniques to use them. Some read the sensor and use a general algorithm to score the driving style, for example, Driver behaviour profiling using smartphone sensory data in a V2I environment paper show a basic algorithm to profile drivers and this is analysed for the vehicle-to-infrastructure environment. The events generated can be clustered according to their geolocation to indicate possible black. This data can be used locally in a vehicle-to-infrastructure environment, where location-based advanced warning messages can be broadcast to vehicles close to the black spot areas. Therefore, any drivers who are within a certain distance of the black spots will be warned, and their awareness level should be raised to be more cautious. I found this useful as an integrated system.

Other applications that use complex machine learning techniques among the most common are Artificial Neural Network (ANN), Support Vector Machine (SVM), Random Forest (RF), and Bayesian Network (BN) where ANN showed high accuracy. This was indicated in the article, A comprehensive review of driver behaviour analysis utilising smartphones. Those applications use different methods to read the driver behaviour including visual features, in the last papers I found several applications that use the most common smart sensor but include the camera to read eye behaviour to determine if the driver is Drowsiness. This application can be invasive, and I think hard to implement, but it is an advance to increase road safety.

According to Statista, there are currently 6.648 billion smartphone users worldwide, which corresponds to 83.37% of the world’s population and future growth is anticipated. The vehicle and navigation industries now have new ways to collect data, which benefits drivers, vehicle owners, and society at large. This is made possible by the continually increasing smartphone penetration around the world. The enormous amount of recently completed projects, both in academia and in business, demonstrate the enormous potential of smartphone-based data capabilities and have established a solid foundation for upcoming mass market applications. Smartphone-based solutions are typically scalable, upgradeable, and affordable as a result of the unprecedented increase in smartphone demand. Indicated in the article Smartphone-based vehicle telematics: A ten-year anniversary.