| Summary: | Vehicle accidents are still happening daily even with the existing technological support
provided. This can be due to limitations of the technology and/or human error. Using a newer
technology, the vehicle-to-everything communication, the aim is to use machine learning
techniques to make predictions on GPS data, in order to provide an early collision warning
system. With such a system in place, drivers would be alerted if a collision might happen
several seconds prior and be mentally prepared for the potential threat. The algorithms
explored in this study is the multi-layered perceptron classifier, random forest and Tabnet.
|
|---|