Metaheuristic Algorithm for Indoor 5G Femtocell Deployment in Multi-Floor Buildings

Abstract

The placement of femtocells in indoor 5G networks within multi-story buildings faces various challenges, including traffic congestion, interference from overlapping coverage, increased energy consumption and network costs, and the potential emergence of coverage blank spots. These challenges become more complex due to multi-wall propagation, differences in floor characteristics, and the requirement for uniform distribution of service quality. This research aims to design the placement of Femtocell Access Points (FAPs) in a multi-floor indoor 5G network at the Classroom Building of the Faculty of Engineering, Hasanuddin University, ensuring that each floor meets the minimum coverage target with an efficient number of FAPs. The proposed method integrates indoor environmental modeling using the COST-231 Multiwall propagation model with metaheuristic algorithms, namely the Flower Pollination Algorithm (FPA) and the Slime Mold Algorithm (SMA), and uses Particle Swarm Optimization (PSO) as a benchmark. Simulations were conducted using MATLAB and the Radiowave Propagation Simulator (RPS) based on ray-tracing techniques. Performance evaluation includes per-floor and global coverage, Reference Signal Received Power (RSRP), Signal-to-Interference-plus-Noise Ratio (SINR), convergence characteristics, the number of active FAPs, and computational time. The results show that SMA achieves the highest global coverage of 93.19% with the fastest computational time of 30,398.74 seconds, and reduces the poor-SINR areas from 9.93% to 2.61% in RPS simulations. Meanwhile, FPA provides the optimal RSRP distribution, reducing the percentage of poor-signal areas from 15.46% to 2.14%. These findings confirm that metaheuristic algorithms integrated with the COST-231 Multiwall model are effective for indoor 5G femtocell network planning in multi-story buildings.

Access to Document

Other files and links

• Link to publication in Scopus
• Open Access Version Available

Fingerprint