News About the mPerf Project

mPerf Project deploys 1,000 mobile sensors to study everyday workplace performance

(Dec. 11, 2017) In six months, the mPerf project has procured, assembled, and deployed more than 1,000 mobile and/or wearable sensors in field and lab studies at three large metropolitan locations across the country for the study of workplace behavior. In the mPerf field study, led by Dr. Deniz Ones from the University of Minnesota, more than 350 participants have already been recruited from numerous organizations (6 companies are registered partners of the mPerf field study), who are contributing approximately 200 gigabytes of sensor and phone usage data each day during the 10-week study.

The mPerf project grew out of the Center of Excellence for Mobile Sensor Data-to-Knowledge (MD2K), and, like MD2K, is headquartered at the University of Memphis. The mPerf project is collecting high frequency mobile sensor data to identify sensor-based markers of workplace behavior. mPerf’s goal is to discover sensor-based indicators that can measure and predict job performance. Such sensor-based assessments can provide organizations valuable insights into employees’ performance and work behaviors, but also aid in the prediction of these behaviors of interest based on reliable individual differences. In addition, data from sensor suites can help individuals improve their work performance through the use of unobtrusive and objective monitoring of psychological and performance states.

At a recent meeting, the mPerf team, which includes researchers from six universities, shared its progress with the Intelligence Advanced Research Projects Activity (IARPA), sponsors of the mPerf project. IARPA is under the office of the Director of National Intelligence.

The sensor data collected will be used to identify markers, including those associated with tasks (e.g., typing), stress, activity, alertness, routine, as well as individual differences in personality attributes. These markers will then be used to develop models that can identify, gauge, and predict work behavior without subjective measures such as interviews and questionnaires.

The mPerf project includes Dr. Santosh Kumar (lead PI) and Dr. Eugene Buder from the University of Memphis, Dr. Deniz Ones and Dr. Mustafa al’Absi from the University of Minnesota, Dr. Emre Ertin from The Ohio State University, Dr. Tanzeem Choudhury from Cornell University; Dr. Mani Srivastava from the University of California, Los Angeles, and Dr. Deepak Ganesan and Dr. Benjamin Marlin from the University of Massachusetts Amherst. The mPerf team also includes 25 engineers and staff, 5 postdocs, and 16 students.