An Evolvable Fuzzy Logic System for handoff management in heterogeneous Wireless Networks

Hossein Fayyazi, Mohammad SabokrouarXiv 2016 · arXiv

Introduction

In Next-Generation Wireless Networks (NGWNs), mobile terminals (MTs) frequently traverse heterogeneous environments, switching between diverse connectivity standards like Wi-Fi, LTE/5G, and WiMax. This process, known as a handoff (or handover), is critical for maintaining service continuity. However, traditional handoff decisions often rely on rigid Received Signal Strength (RSS) thresholds, leading to the "ping-pong effect"—unnecessary, rapid switching between base stations (BS) that wastes energy and degrades user experience.

Comparison of Hard and Soft Handoff mechanisms during the connection transition.

Methodology: The Genetic Fuzzy Logic System (GFLS)

This paper proposes the Genetic Fuzzy Logic System (GFLS), a hybrid approach designed to handle the inherent uncertainty of wireless environments while optimizing decision-making through evolution.

1. Fuzzy Logic Module

Instead of binary thresholds, the system employs fuzzy logic to process three primary inputs:

  • MT Velocity: The speed of the mobile device.
  • Distance: Proximity to the current base station boundary.
  • Free Channels: The available capacity of the target base station.

By incorporating "Free Channels" as a fuzzy input, the system ensures capacity-aware switching, avoiding handoffs to congested towers that might otherwise offer a strong signal.

Membership functions for the 'FreeChannels' input variable, ranging from Low to High capacity.

2. Evolutionary Module (Genetic Algorithm)

To avoid the manual tuning of the 27 fuzzy rules required for this system, the authors utilize a Genetic Algorithm (GA). The GA treats rule consequences as "chromosomes" and evolves them over multiple generations. This "auto-tuning" process selects for rule sets that minimize both unnecessary handoffs and dropped connections, adapting the logic to specific network topologies.

Simulation and Results

The GFLS was evaluated in a custom 6000m × 6000m simulation environment featuring 7 base stations with varying radii and capacities. The performance was compared against a standard Fuzzy Logic-based Adaptive Handoff (FLAH) and a GA-enhanced version (GFLAH).

  • Reduced Signaling Overhead: GFLS significantly outperformed the baseline, averaging only 23 handoffs compared to 58 for the standard FLAH.
  • Enhanced Connectivity: The system maintained active connections for 29.38% of the simulation time, a notable improvement over the 27% achieved by FLAH.
  • System Stability: Even in worst-case scenarios, the GFLS recorded a minimum of 14 handoffs, whereas the baseline's best-case scenario remained at 30.

Conclusion and Future Directions

The GFLS demonstrates that combining fuzzy logic with evolutionary algorithms can create a robust, capacity-aware handoff mechanism. While the system showed a slight increase in energy wastage (16.76% vs 16.17% for GFLAH), the gains in connection stability and reduced signaling overhead are substantial. Future work aims to incorporate 3D environmental modeling, handoff latency, and network cost-efficiency to further refine the system for commercial deployment.

Summary updated

BibTeX
@misc{fayyazi2016evolvable,
  author = {Hossein Fayyazi and Mohammad Sabokrou},
  title = {{An Evolvable Fuzzy Logic System for handoff management in heterogeneous Wireless Networks}},
  year = {2016},
  eprint = {1603.02419},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/1603.02419},
}
Figure