EmbedUWB: Low Power Embedded High-Precision and Low Latency UWB Localization.

Positioning based on Ultra-wideband (UWB) wireless communication promises multipath immunity, high resolution, and strong wall penetration. Hence, many academic and industrial researchers are proposing UWB positioning systems especially to track and locate objects and person in indoor environments. UWB indoor localization, based on stationary "anchors" tracking mobile "tags", is already a reality in many application scenarios. However, most of the available systems either perform tracking on servers which collect data from tags and anchors or delegate to the anchors to calculate the actual position of the tags. Both approaches imply long latency and limited throughput in terms of position fixes per unit time. In this paper, we present the design and the implementation of a highly accurate and low latency asymmetric double-sided two-way ranging algorithm that runs at very modest energy budget on low-power tags based on ARM Cortex-M4F microcontrollers (MCUs). We implemented a complete tag platform, which hosts the MCU and the UWB module, and we evaluated the algorithm in the field. Experimental results demonstrate high accuracy (4.3cm) and position fix rate (20Hz), the energy cost of only 4.9mJ for a single fix, and a cold-start first position fix in only 81ms.