Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Traffic Accident Severity Prediction Model using AI
All vehicle drivers are prone to traffic accidents no matter their age, years of experience, type of vehicle driven, etc. It is estimated that 1.35 million people are killed yearly in crashes involving vehicles such as cars, buses, motorcycles, bicycles, trucks, etc. or even while being bare pedestrians. Such accidents often have a critical economic and social impact on the lives of the deceased's family. Whilst some may die at the instant moment of the accident and some may survive the impact to further meet death later, the severity of the accident itself plays a huge role in the prior. Hence, prediction of accident severity using various features such as time of day, weather conditions, ages of drivers, etc. may aid in the readiness of the dedicated emergency response and may further then increase the chances of surviving the after-impact. This paper discusses the prediction of traffic accident severity using Artificial Neural Networks (ANNs) whilst having a total of 10 input features such as the number of vehicles involved in the accident, day of the week, time of day, road surface condition, street lighting condition, weather condition, type of vehicles involved in the accident, type of causal vehicle, gender of causal driver and their age. The prediction model was implemented using the deep learning toolbox on MATLAB. The model was able to achieve an accuracy of 80% approximately.
Traffic Accident Severity Prediction Model using AI
All vehicle drivers are prone to traffic accidents no matter their age, years of experience, type of vehicle driven, etc. It is estimated that 1.35 million people are killed yearly in crashes involving vehicles such as cars, buses, motorcycles, bicycles, trucks, etc. or even while being bare pedestrians. Such accidents often have a critical economic and social impact on the lives of the deceased's family. Whilst some may die at the instant moment of the accident and some may survive the impact to further meet death later, the severity of the accident itself plays a huge role in the prior. Hence, prediction of accident severity using various features such as time of day, weather conditions, ages of drivers, etc. may aid in the readiness of the dedicated emergency response and may further then increase the chances of surviving the after-impact. This paper discusses the prediction of traffic accident severity using Artificial Neural Networks (ANNs) whilst having a total of 10 input features such as the number of vehicles involved in the accident, day of the week, time of day, road surface condition, street lighting condition, weather condition, type of vehicles involved in the accident, type of causal vehicle, gender of causal driver and their age. The prediction model was implemented using the deep learning toolbox on MATLAB. The model was able to achieve an accuracy of 80% approximately.
Traffic Accident Severity Prediction Model using AI
Hamdan, Saeed M. S. (Autor:in) / Barakat, Samer (Autor:in) / Mahfouz, Khaled Hossam (Autor:in) / Ghuzlan, Khalid A. (Autor:in)
20.02.2023
692986 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Accident prediction models for traffic signals
| British Library Conference Proceedings | 2008