Hand washing is an effective countermeasure to the spread of many types of infection. Recently, sensing technology has automated the sampling and study of hand hygiene rates. Surprisingly, many questions about the area are unresolved, motivating further exploration based on wrist-worn commodity sensors (accelerometer and MEMS gyroscope). This paper describes initial work on techniques for measuring the duration of washing events and classifying different scrubbing motions. The work compares different sensor types and their fusion, compares sensing from one wrist to measuring both wrists, and explains results of experiments on a range of hand washing motions in a variety of subject populations, some in clinics of a teaching hospital. Machine learning is used to explore such questions: the paper investigates numerous features extracted from sensor data, looking at sampling rates, windowing, and platform details that affect classification. In training and classification experiments, data collection starts on the wrist, activated by a message from a disinfectant dispenser; data is then transferred by radio to a base station for subsequent reduction, analysis and characterization. Results show that hand hygiene motions can be classified with up to 93% accuracy.

Neighbor discovery is a component of communication and access protocols for ad hoc networks. Wireless sensor networks often must operate under a more severe low-power regimen than do traditional ad hoc networks, notably by turning off radio for extended periods. Turning off a radio is problematic for neighbor discovery, and a balance is needed between adequate open communication for discovery and silence to conserve power. This paper surveys recent progress on the problems of neighbor discovery for wireless sensor networks. The basic ideas behind these protocols are explained, which include deterministic schedules of waking and sleeping, randomized schedules, and combinatorial methods to ensure discovery.

We captured 3-dimensional accelerometry data from the wrists of 116 healthcare professionals as they performed hand hygiene (HH). We then used these data to train a k-nearest-neighbors classifier to recognize specific aspects of HH technique (ie, fingertip scrub) and measure the duration of HH events.

This paper will summarize the development and teaching of a multidisciplinary, project-based, pilot course on the Internet of Things using strategies inspired by the Lean Startup movement. The course was taught at Rose-Hulman Institute of Technology, a small teaching institution in the Midwest with an emphasis on engineering education. Eight students from four different majors: electrical engineering, computer science, computer engineering and mathematics, were selected to enroll in the course. Our basic approach was to first inspire student learning and then manage the learning. We used a just-in-time strategy to introduce core IoT concepts and principles. The course revolved around a project consisting of deploying sensors on treadmills in our university's sports and recreation center. The lean startup strategy we used was largely successful, dramatically reduced the lead time needed to develop and offer the course to only a few weeks. The multidisciplinary team of instructors greatly expanded the range of expertise available to the students and reduced the teaching burden on any individual faculty member. Students, however, significantly underestimated the amount of time and effort the faculty expected them to spend on the technical work and the final technical report that documented their achievements. This paper will present details and lessons learned from this pilot IoT course.

No abstract