Curvature-Aware Sparse Array Synthesis for Edge-Aware UWB Localization in Multipath-Rich IoT Environments

Ultrawideband (UWB) technology is rapidly gaining adoption across Internet of Things (IoT) deployments such as warehouse automation and asset tracking, where Global Navigation Satellite System (GNSS) signals are often unavailable or unreliable. In such settings, metallic infrastructure, reflective surfaces, and narrow aisles produce delayed secondary paths that distort time-of-arrival (TOA) estimation and undermine system robustness. This article proposes a hardware-level multipath-suppression strategy for UWB sensing tailored to IoT edge nodes. It leverages a conformal sparse array (CSA) optimized via a Penrose-Tessellation-Inspired Invasive Weed Optimization (PT-IWO) algorithm. The design explicitly models surface curvature to accommodate real device form factors and introduces an omnidirectional sidelobe-balancing criterion for spatially uniform interference rejection. Penrose-based quasi-periodic layouts enhance spatial diversity, while adaptive nonlinear modulation ensures stable optimization. The framework is passive and computation-neutral, enabling drop-in integration without added processing load or power draw. A CSA-integrated UWB localization testbed is evaluated in representative IoT environments with strong multipath. Results show up to a 267% increase in first-path amplitude, up to 380% higher signal-to-multipath ratio, and a 20% extension in effective range. These gains translate to higher valid-fix rates and fewer outages, validating a robust, cost-effective physical-layer enhancement for resilient, edge-aware UWB localization at IoT scale.