Suboptimal Soft Range Estimators with Applications in UWB Sensor Networks

In this paper, we consider the positioning problem in wireless sensor networks from a systems perspective, including aspects of the physical layer, channel, ranging algorithm and positioning algorithm. We argue that gains in positioning accuracy and robustness can be achieved by extending the interface between the physical layer ranging algorithm and the higher layer positioning algorithm to include more information than just a single distance estimate per link. Towards this end, we propose and analyze a "soft" distance estimator. This estimator, differently from traditional distance estimators, outputs a list of likely distances for each link together with a list of weights similar to likelihoods. Hence, we broaden the interface between ranging and positioning algorithm, and provide additional information that the positioning algorithm can exploit. The soft distance estimator we propose is of comparably low complexity, and offers competitive performance while making no assumptions on a priori channel information or network synchronization. The analysis and simulations in this work are based on the channel models adopted by the IEEE 802.15.4a working group, and highlight the benefits and drawbacks of the proposed approach.