Optimal ALOHA-like Random Access with Heterogeneous QoS Guarantees for Multi-Packet Reception Aided Visible Light Communications

TL;DR

This paper introduces a novel distributed random access algorithm for VLC systems that leverages SIC and MPR capabilities to improve collision management and QoS guarantees, formulated as an optimization problem solved by advanced heuristics.

Contribution

It proposes an EC-based ALOHA-like algorithm tailored for MPR-enabled VLC systems with heterogeneous QoS, using a new conflict graph model and a memetic optimization approach.

Findings

EC expression matches simulation accurately

Algorithm achieves competitive throughput

Supports heterogeneous QoS guarantees

Abstract

There is a paucity of random access protocols designed for alleviating collisions in visible light communication (VLC) systems, where carrier sensing is hard to be achieved due to the directionality of light. To resolve the problem of collisions, we adopt the successive interference cancellation (SIC) algorithm to enable the coordinator to simultaneously communicate with multiple devices, which is referred to as the multi-packet reception (MPR) capability. However, the MPR capability could be fully utilized only when random access algorithms are properly designed. Considering the characteristics of the SIC aided random access VLC system, we propose a novel effective capacity (EC)-based ALOHA-like distributed random access algorithm for MPR-aided uplink VLC systems having heterogeneous quality-of-service (QoS) guarantees. Firstly, we model the VLC network as a conflict graph and derive the EC for each device. Then, we formulate the VLC QoS-guaranteed random access problem as a saturation throughput maximization problem subject to multiple statistical QoS constraints. Finally, the resultant non-concave optimization problem (OP) is solved by a memetic search algorithm relying on invasive weed optimization and differential evolution (IWO-DE). We demonstrate that our derived EC expression matches the Monte Carlo simulation results accurately, and the performance of our proposed algorithms is competitive.