Agile gravitational search algorithm for cyber-physical path-loss modelling in 5G connected autonomous vehicular network

Okafor, KC ORCID: https://orcid.org/0000-0002-9243-6789, Adebisi, B ORCID: https://orcid.org/0000-0001-9071-9120, Akande, AO and Anoh, K (2024) Agile gravitational search algorithm for cyber-physical path-loss modelling in 5G connected autonomous vehicular network. Vehicular Communications, 45. 100685. ISSN 2214-2096

Preview

Published Version Available under License Creative Commons Attribution. Download (6MB) | Preview

Abstract

Based on the characteristics of the 5 G standard defined in Release 17 by 3GPP and that of the emerging Beyond 5 G (or the so-called 6 G) network, cyber-physical systems (CPSs) used in smart transport network infrastructures, such as connected autonomous vehicles (CAV), will significantly depend on the cellular networks. The 5 G and Beyond 5 G (or 6 G) will operate over millimetre-wave (mmWave) bands. These network standards require suitable path loss (PL) models to guarantee effective communication over the network standards of CAV. The existing PL models suffer heavy signal losses and interferences at mmWave bands and may not be suitable for cyber-physical (CP) signal propagation. This paper develops an Agile Gravitational Search Algorithm (AGSA) that mitigates the PL and signal interference problems in the 5G–NR network for CAV. On top of that, a modified Okumura-Hata model (OHM) suitable for deployment in CP terrestrial mobile networks is derived for the CAV-CPS application. These models are tested on the real-world 5 G infrastructure. Results from the simulated models are compared with measured data for the modified, enhanced model and four other existing models. The comparative evaluation shows that the modified OHM and AGSA performed better than existing OHM, COST, and ECC-33 models by 90%. Also, the modified OHM demonstrated reduced signal interference compared to the existing models. In terms of optimisation validation, the AGSA scheme outperforms the Genetic algorithm, Particle Swarm Optimisation, and OHM models by at least 57.43%. On top of that, the enhanced AGSA outperformed existing PL (i.e., Okumura, Egli, Ericson 999, and ECC-33 models) by at least 67%, thus presenting the potential for efficient service provisioning in 5G-NR driverless car applications.