Study on real-time collision avoidance decision-making under 5G environment

Aiming at the problems of real-time collision prevention and untimely early warning, an intelligence infrastructure which based on 5G and Beidou is constructed. The real-time data collection of Beidou and AIS is used for preprocessing and filling, which improves the accuracy and usability of Beidou and AIS data, and provides a reliable data source for the real-time collision risk assessment of ships. By establishing distance to closest point of approach (DCPA) model and Time to closest point of approach (TCPA) model, the collision risk of ships in space and in time domain are obtained, and then real-time navigation collision risk index (CRI) are obtained, the safety warning of ships with different risk levels is realized by CRI. Taking Nanjing Banqiao ferry as a demonstration point, the DCPA model and TCPA model were applied to the ferry safety management system, multiple real-time warnings are carried out, and the early warning accuracy rate was up to 96.15%, which was unanimously praised by the management department. Finally, the factors to be considered for subsequent optimization of the CRI model are pointed out.