Undergrad Research Project - Real-Time Fine Grained Occupancy Estimation

Spring 2017

Student
Jonathan Appiagyei
Advisor
Anthony Rowe
Project description

Accurate sensing of room occupancy has the potential to improve building energy efficiency, security, and room scheduling. Unfortunately, current methods are either inaccurate, costly, or difficult to install. We propose using depth-sensing technology in tandem with machine-learning algorithms to create a new type of occupancy sensor. Specifically, we will be evaluating the performance of the Texas Instruments’ OPT8241 and Intel’s RealSense R200 in terms of range, power and accuracy. We will design an easy to deploy platform that incorporates an Intel Compute Stick with the depth sensor to perform local processing of depth images. We will benchmark multiple approaches in terms of accuracy when deployed in a variety of real-world scenarios.

Return to project list