Vincent Beech and Paul Groves, University College London, UK; Paul Wright, Terrafix Limited, UK

View Abstract Sign in for premium content


In emergency and military scenarios, standard positioning and communications infrastructure cannot be relied upon. Ad-hoc mobile networks offer a decentralized alternative, so this paper addresses the challenge of incorporating a ranging capacity to an existing ad-hoc digital communications system. However, signal protocols and signal processing chains that are optimised for efficient communications do not easily support the precise timing that is required for accurate range measurements. For this system, there is the further challenge that it uses narrow bandwidth signals in the 10?20kHz range. For perspective, each data symbol is 2-3 orders of magnitude larger than a chip of a Global Navigation Satellite System (GNSS) signal. Therefore, signal timing measurements must be accurate to a tiny fraction of a symbol if a useful positioning capability is to be achieved. By applying innovative synchronization techniques to a High-Performance Software-Defined Radio (HPSDR) network, ranging data collected under benign conditions exhibits a standard deviation of only 2.61m despite using these very narrowband signals. This paper will describe how such a level of performance was achieved and verified experimentally. This paper begins with an overview of the radio network used in this project and how the conclusion that two-way ranging would be the best choice of position fixing for this network rather than other options available. Then the paper continues with a summary of some preliminary work where range estimates have been computed using the radio’s existing synchronization protocols. These primitive ranging protocols form the foundation upon which more advanced ranging protocols will be implemented. What follows is the body of the paper where the three key components of signal processing that led to the precise ranging achieved are discussed: timing recovery, non-coherent integration and clock drift correction. Finally, a breakdown of the best results achieved by implementing these key components are discussed followed by a brief discussion of further work that could be done.