Optimization of a preceding vehicle collision risk prediction model based on SSA-LSTM

Yipei Cai; Tao Zhang; Junlang Cui

doi:10.1117/12.2689471

7 September 2023 Optimization of a preceding vehicle collision risk prediction model based on SSA-LSTM

Yipei Cai, Tao Zhang, Junlang Cui

Proceedings Volume 12790, Eighth International Conference on Electromechanical Control Technology and Transportation (ICECTT 2023); 127906N (2023) https://doi.org/10.1117/12.2689471
Event: 8th International Conference on Electromechanical Control Technology and Transportation (ICECTT 2023), 2023, Hangzhou, China

Abstract

According to statistics, there are approximately 1.24 million deaths due to traffic accidents worldwide every year, and road traffic safety is expected to become the fifth leading cause of death globally by 2030. Therefore, collision avoidance systems have been proposed to predict collision risks and prevent accidents. Existing vehicle collision risk prediction systems based on Long Short-Term Memory (LSTM) networks do not consider the optimization of model hyperparameters. In this paper, we integrate LSTM and Sparrow Search Algorithm (SSA) to optimize four hyperparameters of the LSTM model. This algorithm predicts collision risks by considering both vehicle and driver factors. The proposed model is evaluated using the Next Generation Simulation (NGSIM) dataset, and the experimental results indicate that the vehicle collision warning model based on SSA-LSTM has achieved an accuracy of 97.0%, which is a 2.1% improvement over the LSTM model. Moreover, the model has made significant improvements in reducing both false alarm rate and missed detection rate.

Citation Download Citation

Yipei Cai, Tao Zhang, and Junlang Cui "Optimization of a preceding vehicle collision risk prediction model based on SSA-LSTM", Proc. SPIE 12790, Eighth International Conference on Electromechanical Control Technology and Transportation (ICECTT 2023), 127906N (7 September 2023); https://doi.org/10.1117/12.2689471

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available