Building ventilation systems are not responsive to occupants, operating according to predetermined schedules to satisfy the maximum number of people that could be in a space. It's comparable to a car that only runs in the highest gear: ineffective, inflexible, and inefficient. Vast amounts of energy are wasted while making people too cold or hot in their offices. Based on a study conducted by Pacific Northwest National Labs there is a potential for 16% energy savings through incorporation of high resolution real-time occupancy data in building automation systems that control heating, ventilation and air conditioning (HVAC) equating to roughly $2.7 billion annually. A survey conducted by Center for Built Environment at UC Berkeley showed that merely 11% of buildings fulfill standards for thermal comfort, and 26% meet air quality requirements, severely impairing occupant satisfaction and productivity. The small array of occupancy sensing products currently on the market are cumbersome, too costly, and tend to focus on narrow aspects of building operation. As a result they've achieved adoption of less than 1% of the total nonresidential real estate market. This I-Corps team is working on a passive sensing system to address this industry-wide blind spot with a cost-effective and easy to deploy system in order to increase building energy efficiency and improve occupant comfort.
This team is working on a distributed sensing solution that will enable building HVAC control systems to respond to and anticipate building occupancy. The complete initial version of the system will include machine learning and computer vision algorithms embedded in the small fully wireless sensors processing data from a low-cost RGB camera and passive infrared sensor that will be able to detect both stationary and moving people across a coverage area of approximately 600 square feet in an open space. By the end of the program the team intends to have completed a software prototype for the image processing portion capable of delivering real-time occupancy counting and prediction using sample building image streams. More importantly, the team plans to have validated the demand in the market for the proposed product through the interviews, investigated the potential privacy concerns from prospective clients, explored value propositions outside of energy efficiency and comfort from advanced occupancy data, and identified the types of real estate clients that we should focus on. The proposed product has the potential to greatly reduce building energy consumption and improve occupant comfort throughout the US, and more broadly make the built environment far more responsive and data-driven. The team's interviews over the course of I-Corps program will be vital in shaping the business model to bring this application of computer vision and machine learning research and development to market.
