Device-free localization (DFL) methods use measured changes in the received signal strength (RSS) between many pairs of RF nodes to provide location estimates of a person inside the wireless network. Fundamental challenges for RSS DFL methods include having a model of RSS measurements as a function of a person's location, and maintaining an accurate model as the environment changes overtime. Current methods rely on either labeled empty-area calibration or labeled fingerprints with a person at each location. Both need to be frequently recalibrated or retrained to stay current with changing environments. Other DFL methods only localize people in motion. In this paper, we address these challenges by, first, introducing a new mixture model for link RSS as a function of a person's location, and second, providing the framework to update model parameters without ever being provided labeled data from either empty-area or known-location classes. We develop two new Bayesian localization methods based on our mixture model and experimentally validate our system at three test sites with seven days of measurements. We demonstrate that our methods localize a person with non-degrading performance in changing environments, and, in addition, reduce localization error by 11 - 51 percent compared to other DFL methods.

We specify and evaluate a new software-defined clock network architecture, Stitch. We use Stitch to derive all subsystem clocks from a single local oscillator (LO) on an embedded platform, and enable efficient radio frequency synchronization (RFS) between two nodes' LOs. RFS uses the complex baseband samples from a low-power low-cost narrowband transceiver to drive the frequency difference between the two devices to less than 3 parts per billion (ppb). Recognizing that the use of a wideband channel to measure clock frequency offset for synchronization purposes is inefficient, we propose to use a separate narrowband radio to provide these measurements. However, existing platforms do not provide the ability to unify the local oscillator across multiple subsystems. We demonstrate Stitch with a reference hardware implementation on a research platform. We show that, with Stitch and RFS, we are able to achieve dramatic efficiency gains in ultra-wideband (UWB) time synchronization and ranging. We demonstrate the same UWB ranging accuracy in state-of-the-art systems but with 59% less utilization of the UWB channel.

This paper addresses the performance of systems which use commercial wireless devices to make bistatic RF channel measurements for non-contact respiration sensing. Published research has typically presented results from short controlled experiments on one system. In this paper, we deploy an extensive real-world comparative human subject study. We observe twenty patients during their overnight sleep (a total of 160 hours), during which contact sensors record ground-truth breathing data, patient position is recorded, and four different RF breathing monitoring systems simultaneously record measurements. We evaluate published methods and algorithms. We find that WiFi channel state information measurements provide the most robust respiratory rate estimates of the four RF systems tested. However, all four RF systems have periods during which RF-based breathing estimates are not reliable.

We present results from a system which uses received signal strength (RSS) measurements to estimate the speed at which a person is walking when they cross the link line. While many RSS-based device-free localization systems can detect a line crossing, this system estimates additionally the speed of crossing, which can provide significant additional information to a tracking system. Further, unlike device-free RF sensors which occupy tens of MHz of bandwidth, this system uses a channel of about 10 kHz. Experiments with a person walking from 0.3 to 1.8 m/s show the system can measure walking speed within 0.05 m/s RMS error.

Detecting and locating outdoor border crossing events is valuable information in curbing drug trafficking, reducing poaching, and protecting high-asset equipment and goods. However, border sensing is notoriously challenging, prone to false alarms and missed detections, with serious consequences. Weather events, like rain and wind, make it even more challenging to maintain a low level of missed detections and false alarms. In this paper, we propose and test an automated system of wireless sensors which uses received signal strength (RSS) measurements to localize where a border crossing occurs. In addition, we develop new RSS-based statistical models and methods that can quickly be initialized and updated by using link RSS statistics to adapt to time-varying RSS changes due to weather events. These models are implemented in two new classifiers that localize border crossings with few missed detections and false alarms. We validate our proposed methods by implementing one of the classifiers in a three month long deployment of a solar-powered, real-time system that captures images of the border for ground truth validation. Furthermore, over 75 hours of RSS measurements are collected with an emphasis on collection during weather events, like rain and wind, during which we expect our classifiers to perform the worst. We demonstrate that the proposed classifiers outperform four other baseline classifiers in terms of false alarm probability by 1 to 4 orders of magnitude, and in terms of the misclassification probability by 1 to 2 orders of magnitude.

Detecting and localizing a person crossing a line segment, i.e., border, is valuable information in security and data analytic applications. To that end, we use the received signal strength (RSS) measured on RF links between nodes deployed linearly along a border as a border crossing detection and localization system. RSS measurements from any single RF link are noisy and prone to variations due to environmental changes (e.g. branches moving in wind). The redundant overlapping nature of the links between pairs of nodes in our proposed system provides an opportunity to mitigate these issues. We propose a hidden Markov model (HMM) which models the RSS on network links as a function of the neighboring nodes between which a person crosses. We demonstrate that the forward-backward solution to this HMM provides a robust and real time border crossing detection and localization system.