Energy-efficient adaptive routing in heterogeneous wireless sensor networks via hybrid PSO and dynamic clustering

This paper presents a deterministic routing framework for static heterogeneous wireless sensor networks (WSNs) that addresses energy imbalance and scalability challenges through three synergistic innovations. First, an Improved Particle Swarm Optimization (IPSO) algorithm dynamically adjusts inertia weights and learning factors via a cosine-modulated mechanism, enabling adaptive transitions between global exploration and local refinement to avoid premature convergence. Second, a hybrid communication strategy integrates single-hop and multi-hop routing, dynamically selecting transmission paths based on real-time energy gradients and node density distributions to mitigate congestion near sinks. Third, a heterogeneity-aware design embeds node-specific attributes—including residual energy, transmission range, and functional roles—into a weighted fitness model, ensuring optimal resource allocation across diverse node types. Experimental evaluations demonstrate significant improvements over conventional protocols: a 32.7% reduction in per-round energy consumption compared to LEACH, a 28% extension in network lifespan (first-node-death metric), and 63% lower energy variance across nodes. The framework’s lightweight implementation achieves sublinear regret bounds in dynamic environments while maintaining compatibility with resource-constrained edge devices, as validated on a Raspberry Pi 4B testbed.