A Lightweight SparkLink-Based Framework for Indoor Vehicle Sensing via Channel-Spliced Estimation and Compact RF Architecture

Recent advances in wireless communication have intensified the need for high-precision, low-latency indoor location-awareness applications in fields such as autonomous navigation and smart manufacturing. To address the limitations of existing RF-based technologies in bandwidth and latency, this paper investigates the SparkLink protocol and presents a novel wireless sensing framework. First, we employ frequency-hopping and multi-channel splicing to construct a wideband channel for improved channel parameters estimation. To address synchronization-induced phase errors, a bidirectional communication approach is introduced to effectively calibrate and eliminate inter-channel phase offsets. Building upon the spliced multi-channel, atomic norm minimization is applied for high-precision time-of-flight estimation. Furthermore, we develop a receiver architecture based on a single radio frequency chain integrated with a four-element uniform planar array, leveraging a tailored time-switching strategy to sequentially acquire spatial measurements. By incorporating the iterative adaptive approach, the system enables high-resolution two-dimensional direction-of-arrival estimation within the constraints of compact hardware, thereby supporting precise spatial sensing in complex indoor environments. Experimental evaluations in a real-world underground parking facility demonstrate that the proposed system achieves a median localization accuracy of 1.13 m for stationary targets and 1.32 m for dynamic tracking, delivering performance on par with state-of-the-art wireless sensing solutions.