Multi-Object Tracking for Collision Avoidance Using Multiple Cameras in Open RAN Networks

TL;DR

The paper investigates collision avoidance for ITS by enabling Multi-Object Multi-Camera Tracking (MOMCT) over a realistic Open RAN 6G network. It couples CARLA-based perception from distributed cameras with an edge service that fuses tracklets and performs trajectory prediction using Particle Filters, where the PF state is $\mathbf{x}^i = [x_{lat}, dx_{lat}, x_{lon}, dx_{lon}, w^i]^T$, the trajectory estimate is $\hat{\mathbf{x}} = \frac{1}{N} \sum_{i=1}^N w^i \mathbf{x}^i$, and the dynamics follow $\mathbf{x}^i_t = \mathbf{M}_t \mathbf{x}^i_{t-1} + \mathbf{R}$ with likelihood $\mathcal{L}_{xi} = \mathcal{N}_{pdf}(\| \mathbf{x}^i - \mathbf{y}_t \|, 0, k)$. The contributions include a full Open RAN deployment (SRS RAN + Open5GS) and a PF-based, multi-camera MOMCT demonstration within CARLA, highlighting end-to-end feasibility and practical impact for real-world 6G ITS deployments.

Abstract

This paper deals with the multi-object detection and tracking problem, within the scope of open Radio Access Network (RAN), for collision avoidance in vehicular scenarios. To this end, a set of distributed intelligent agents collocated with cameras are considered. The fusion of detected objects is done at an edge service, considering Open RAN connectivity. Then, the edge service predicts the objects trajectories for collision avoidance. Compared to the related work a more realistic Open RAN network is implemented and multiple cameras are used.

Figures (3)

• Figure 1: Architecture of the MOMCT testbed in Open RAN
• Figure 2: Object detection and tracking in CARLA clients
• Figure 3: MOMCT edge service predicting trajectories