Optimizing End-to-End Throughput in Network-Sliced 5G Systems for Real-Time Collision Avoidance

Abstract

Network slicing in fifth-generation (5G) wireless systems enables flexible resource allocation and service customization across diverse use cases. Real-time collision avoidance demands high end-to-end throughput and deterministic latency guarantees, since vehicles, drones, or autonomous machinery rely on rapid data exchange and decision-making logic. Traditional static configurations struggle to account for variable traffic patterns, dynamic channel conditions, and stringent reliability requirements. Adaptive slicing frameworks can integrate cross-layer optimization to ensure that capacity and latency constraints are jointly satisfied, while also maximizing network utilization under heterogeneous workloads. Resource blocks, slicing policies, and interference mitigation strategies must be dynamically orchestrated to balance throughput against latency, jitter, and reliability. End-to-end throughput, from the local edge domain to the core network, influences collision detection algorithms, alert forwarding, and actuator control for timely evasive actions. Emerging approaches leverage advanced scheduling, machine learning-based prediction, and real-time analytics to handle bursts in traffic and topological changes. Ongoing standardization and industrial collaborations focus on refining these solutions, emphasizing robust, scalable, and secure network orchestration. This work explores a comprehensive framework for throughput optimization and real-time collision avoidance in 5G-sliced systems, discussing the principal design challenges and highlighting how multi-dimensional resource management schemes can support mission-critical, latency-sensitive operations.