Low-Power Routing Algorithms for WSNs in Agricultural IoT Systems

Abstract

Agricultural Internet-of-Things (IoT) deployments rely on Wireless Sensor Networks (WSNs) to monitor soil moisture, temperature, humidity, leaf wetness, and micro-climate conditions across large, heterogeneous fields. Because nodes are battery-powered and often difficult to access, routing must maximize network lifetime while preserving data fidelity for decisions such as irrigation scheduling and disease prevention. This manuscript evaluates low-power routing strategies for agricultural WSNs and proposes AELR (Adaptive Energy & Link-quality Routing), a lightweight cross-layer approach that blends residual-energy awareness, link-quality metrics, geographic progress, and application-level sampling dynamics. 

Using a first-order radio energy model and realistic traffic patterns (periodic telemetry with event-driven bursts tied to soil-moisture thresholds), we simulate baseline protocols (LEACH, HEED, and RPL-OF0) against AELR on 100–300 nodes over a 500×500 m farm plot. AELR introduces (i) score-based next-hop selection with hysteresis to prevent route flapping, (ii) adaptive cluster rotation driven by energy skew, (iii) duty-cycling coordinated with sampling phases, and (iv) in-network delta aggregation to compress slowly varying signals. Results show improvements in time-to-first-node-death (FND), packet delivery ratio (PDR), and energy-per-useful-bit, with statistically significant gains over baselines under both uniform and hot-spot traffic. The study demonstrates that modest cross-layer cues—particularly coupling routing with application thresholds—yield meaningful lifetime extensions without expensive computation or GPS hardware. We discuss complexity, implementation notes for Contiki-NG/RIOT, limitations (e.g., mobility and sparse anchors), and directions for field trials with mobile sinks and energy harvesting.