Greg Pottie

Areas of Research Interest

My students are presently pursuing systems research on a variety of projects related to wireless health. Human activities and motions are important in a large number of medical applications including rehabilitative therapy, chronic disease management, and health and wellness. Currently, the only really reliable method to assess how well activities are performed has been direct human observation. Our goal is to enable low cost quality assessment of activities in the home and other settings outside the clinic, given realistic constraints, at very large scale in the general population. Research includes experimental methods, multi-layered classification algorithms, creation of large and easily used datasets, and effective feedback to end-users.

A limited set of topics in wireless communications related to optimal resource allocation in cellular networks is also being pursued.

How to Get Involved in Research

Undergraduate: UCLA students can become involved in research either through course EE 199 (see the courses page) or via the NSF Site REU listed below.

Graduate: M.S. students can pursue a project course EE 299 (see the courses page) or discuss research interests during office hours. A 299 course is a good way to explore potential Ph.D. research topics.


Wireless Health Institute (WHI)

Wireless Health Institute

The UCLA Wireless Health Institute is a collaboration among the Schools of Engineering, Medicine and Nursing. The goal is to enable personalized and lower cost care through the application of modern sensing and information technologies to the practice of medicine. The research is distinguished by considering the complete end to end system, including such important issues as ease of use, deployment logistics and integration into existing practice.

The research pursued in Prof. Pottie’s group focuses on robust inference of human motions using low cost devices, including inertial sensors worn on the body and short-range depth camera systems such as the Kinect. At the lowest level of decision granularity, we would like to know whether a person is active or not. The next level is to determine the type of activity being pursued, and finally, we would like to characterize the quality of the activity. With such information, rehabilitation patients or athletes could receive feedback on whether exercises are being performed properly, and clinicians or coaches could track progress and change training regimens appropriately for faster learning of physical skills. All of this could be determined if we knew the trajectories of the limbs. Unfortunately, while in principle it is possible to determine trajectories of particular limbs through integration of accelerometer and gyroscope data, in practice noise and sensor drift quickly make the estimates of position inaccurate with low-cost devices. Thus, models of the motion must be constructed in order to correct for these errors, using an iteration between multiple decision levels. Moreover, members of the general public will not consistently place sensors on the body in the same position and orientation, causing further errors in motion reconstruction. We can use a combination of motion models, classification of opportunistic motions such as walking, and motions captured in front of depth cameras to determine the unknown sensor pose.

For more information on the WHI, click here.

NSF Site Research Experience for Undergraduates: Wireless Health Research and Education

Site REU: Wireless Health Research and Education

Sponsor: National Science Foundation

This program supports 10 undergraduate researchers each summer to pursue a variety of projects related to wireless health. The program is open to US citizens and permanent residents pursuing undergraduate degrees in engineering or computer science. Both a stipend and housing allowance are supplied. Students who are selected for this program work in teams of two, under the supervision of one or two graduate student mentors. The program includes enrichment activities such as how to apply to graduate schools, technical writing, and research seminars. A significant fraction of the projects result in conference publications, while also supporting development of materials for senior design projects. Thus, summer research can have both professional and educational impact.

For information on how to apply, see

Reliable Inference with Missing, Masking, Malfunctioning or Malicious Sensors

Reliable Inference with Missing, Masking, Malfunctioning or Malicious Sensors

Sponsor: IARPA

This project is concerned with how to reliably make inferences concerning physical phenomena given practical sensor impairments. The goal is construction of a general theoretical framework that can apply to a variety of domains. The initial applications are cognitive radios, wireless health, and utility monitoring.<

The focus of Prof. Pottie’s group is on the wireless health application, with the main impairments being noise and drift, improper sensor placement, and missing sensors. One research thrust is to easily go back and forth between data from visual motion capture systems and inertial sensors, since the former are costly but accurate, while the latter are portable and low cost in themselves but as noted above it is difficult to establish ground truth. If carefully modeled, it would be possible to simulate a variety of sensor positions on the body, in order to develop more robust algorithms without the need for costly experiments to acquire large quantities of ground truth. Along similar lines, the motions themselves do not usually need all the observations to be recorded in order for motion trajectories to be reconstructed. Thus motion models can also lead to data compression.

These experimental activities support the main goal of understanding how to best construct a multi-level modeling framework on Bayesian principles so that the best quality of inference can be made, given the observations available. Abstraction levels in this case include context of the motions (where, when), activities, sensor pose and placement, and motion trajectories for particular limbs. Particular algorithms are associated for making inferences at each of these levels, and also for iterating among decisions at different levels to improve the inference quality.

For more information on this project, see