Scholarlink

<p>Reliable communication is essential during disasters, yet infrastructure-based networks often collapse due to power outages and base station failures. Autonomous Device-to-Device (D2D) clustering offers an infrastructure-independent solution. However, the stability and efficiency of such networks strongly depend on parameter choices such as fuzzy rules, PSO weight distributions, and cluster head (CH) ratios that govern cluster formation and reselection. If these parameters are set heuristically, the network may suffer from unbalanced energy consumption, premature cluster head failures, and rapid connectivity degradation, leading to unstable and non-reproducible deployment outcomes. This paper proposes a two-stage robust tuning framework for disaster-oriented D2D clustering by integrating One-at-a-Time (O-T-A) sensitivity analysis with a Taguchi L25 orthogonal array design. The O-T-A stage defined meaningful parameter ranges, and the Taguchi method then examined how these factors interacted, measuring robustness through signal-to-noise ratios to determine the most stable configuration. Simulations with 1000 nodes in a 500 × 500&#xa0;m² disaster field revealed that the optimum configuration (Balanced Fuzzy Inference System, Distance-Focused PSO, Small Adaptive CH ratio, High Maximum CH ratio) delayed the First Node Dead by 19,375%, reduced energy consumption by 50.17%, and improved survivability, while maintaining latency within acceptable limits. Although throughput decreased, this trade-off is acceptable in disaster scenarios where coverage continuity and resilience take precedence. The findings highlight adaptive CH ratio tuning as the dominant lever for robustness, with PSO weights and fuzzy variants providing secondary refinements. This work contributes a statistically validated and reproducible pathway for parameter optimisation, strengthening the reliability of autonomous D2D communication under disaster conditions.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A two-stage O-T-A and Taguchi approach for resilient D2D clustering in disaster communication

  • 期刊论文

  • 发表日期：

    2026年5月11日

查看全文
  • Norhisham Mansor,
  • Wahidah Md Shah,
  • Najwan Khambari,
  • Aslinda Hassan,
  • Affero Ismail

摘要

Reliable communication is essential during disasters, yet infrastructure-based networks often collapse due to power outages and base station failures. Autonomous Device-to-Device (D2D) clustering offers an infrastructure-independent solution. However, the stability and efficiency of such networks strongly depend on parameter choices such as fuzzy rules, PSO weight distributions, and cluster head (CH) ratios that govern cluster formation and reselection. If these parameters are set heuristically, the network may suffer from unbalanced energy consumption, premature cluster head failures, and rapid connectivity degradation, leading to unstable and non-reproducible deployment outcomes. This paper proposes a two-stage robust tuning framework for disaster-oriented D2D clustering by integrating One-at-a-Time (O-T-A) sensitivity analysis with a Taguchi L25 orthogonal array design. The O-T-A stage defined meaningful parameter ranges, and the Taguchi method then examined how these factors interacted, measuring robustness through signal-to-noise ratios to determine the most stable configuration. Simulations with 1000 nodes in a 500 × 500 m² disaster field revealed that the optimum configuration (Balanced Fuzzy Inference System, Distance-Focused PSO, Small Adaptive CH ratio, High Maximum CH ratio) delayed the First Node Dead by 19,375%, reduced energy consumption by 50.17%, and improved survivability, while maintaining latency within acceptable limits. Although throughput decreased, this trade-off is acceptable in disaster scenarios where coverage continuity and resilience take precedence. The findings highlight adaptive CH ratio tuning as the dominant lever for robustness, with PSO weights and fuzzy variants providing secondary refinements. This work contributes a statistically validated and reproducible pathway for parameter optimisation, strengthening the reliability of autonomous D2D communication under disaster conditions.