Investigating the Complexity-Performance Tradeoff of URA8 Topology for Bluetooth 5.1 HAAT

Abstract

The High Accuracy Asset Tracking (HAAT) of Bluetooth 5.1 Low Energy needs tiny antenna arrays to achieve efficient and very compact solutions. Rectangular Antenna Arrays guarantee good performances, both in azimuth and elevation in the Direction of arrival (DoA) estimation. This paper studies the eight antenna rectangular array topology, called URA8. Historically, MUSIC has been the most used algorithm on this type of array. Unfortunately in HAAT real-time applications, where it is essential to have a fast response from the DoA Locator, MUSIC, due to its high computational complexity in the peak searching over the pseudo-spectrum, may be inadequate for the scope. This paper proposes an innovative approach to reduce the complexity of this peak-searching phase: instead of only one peak-search over the complete azimuth-elevation hemisphere, we propose a two step search, the first with a reduced resolution and the latter with a fine resolution inside a proper confidence region, found in the first step. The proposed method ensures a good trade-off between computational complexity and detection performances: it turns out to approximately 6 times faster than the classic MUSIC, maintaining the same precision.