Willecke, Alexander (Technische Universität Braunschweig), Stahl, Roman (Technische Universität Braunschweig), Wolf, Lars (Technische Universität Braunschweig)

Object Fusion for Collective Perception: Will It Scale?

Scheduled for presentation during the Regular Session "Collective perception and localization" (ThAT4), Thursday, September 26, 2024, 10:30−10:50, Salon 7

2024 IEEE 27th International Conference on Intelligent Transportation Systems (ITSC), September 24- 27, 2024, Edmonton, Canada

This information is tentative and subject to change. Compiled on December 26, 2024

Abstract

With Collective Perception, stations exchange information from their local perception with each other via broadcast communication. The information about perceived objects can enhance their awareness, allowing perception beyond their line of sight and increasing traffic safety. Therefore, received object information must be processed by an Object Fusion system that creates a holistic environment model. Solving this fusion problem is not trivial and computationally expensive. However, the computing power required for a single station that implements an Object Fusion system depends on the amount of object data that must be processed. The amount of data will rise with a progressing market introduction as more stations share their perception. For analyzing the scalability of fusion, the Artery V2X Simulation framework was extended with an Object Fusion architecture. The results show how much object data must be processed and stored for three realistic large-scale traffic scenarios at different Market Penetration Rates.