The occupancy of a room can often be determined by a number of context clues. Predictably, the more clues to the software's disposal, the great the accuracy of the estimation. The following three categories outline simple by helpful data and how it can be gathered by a building.
ONLINE/OFFLINE
Today, occupants of a building are almost always connected. Data can be collected by considering Wi-Fi connection, instant messaging, and laptop usage. This is relatively easy to monitor, as connectivity is usually controlled by the workplace, laptops and computers are often company owned, and the majority of IM clients offer status options such as "Available," "Busy," or "Away." Furthermore, the potential of Wi-Fi connectivity can be enhanced when considering signal strength; knowing where a strong connection is and determining which devices have the corresponding connection is yet another hint to help determine real-time occupancy.
|
Some other features of a workplace also require the interaction of an occupant to function. Like computer stations, devices such as microwaves, lights, and stoves often are only in use when someone is actively manipulating or monitoring the system. Collecting electricity consumption data from these sources can help increase accuracy [20].
PLANNING DATA
Much of an person's schedule is planned, noted, and recorded in ways easily analyzed by occupancy software. Having access to a person's calender or monitoring how many invitations have been extended for a meeting can hint to potential future occupancy of a space. Software can be designed to use schedules and the time of day as futher context clues.
AROUND THE WORKPLACE
Different spaces around the workplace should not be treated as though they are the same. Offices, hallways, and cubicals all offer their own challenges. Typical human behavior and the intended use of the tested space all offer new clues. Software can analyze the above information to create what one team calls "Situations". Situations such as "At Her Cubical", "Taking a Break", or "Attending a Meeting" suggest different behaviors, and these likely behaviors often correspond to likely locations [21].
ACCURACY POTENTIAL
All of this data, from sensors and opportunistic sources both, helps create a picture of just what's happening in a building. This thermal energy snapshot is getting more and more precise. This energy-saving technique infers, using probability equations and algorithms to guess with high accuracy the number of people in an area. Figures 9 and 10 to below demonstrate how all the different parts come together to create a more accurate picture.
Currently, workplace occupancy detection accuracy can be as high as 80% [21 & 22], and which greater mechanical and data aquisition technology, this number can only grow.
Click on the button to the right for Data Analysis, an explanation of accuracy, an addressing of the concerns of personal data usage, and an idea of just how precise using sensors in conjunction with opportunistic data.
|
Sources:
[20] Kleiminger, Wilhelm, "Occupancy Detection from Electricity Consumtion Data", 2016 December 7th, from: https://www.vs.inf.ethz.ch/publ/papers/wilhelmk-occupan-20131.pdf
[21] Kumar Ghai, Sunil, "Occupancy Detection in Commercial Buildings using Opportunistic Context Sources", 2016 December 7th, from: www-07.ibm.com/in/research/documents/p469-ghai.pdf
[22] Jain, Mohit and Chandan, Vikas, "Softwar-only Occupancy Inference in a Workplace", 2016 December 7th, from: homes.cs.washington.edu/_mohitij/ISGT-2016.pdf
[20] Kleiminger, Wilhelm, "Occupancy Detection from Electricity Consumtion Data", 2016 December 7th, from: https://www.vs.inf.ethz.ch/publ/papers/wilhelmk-occupan-20131.pdf
[21] Kumar Ghai, Sunil, "Occupancy Detection in Commercial Buildings using Opportunistic Context Sources", 2016 December 7th, from: www-07.ibm.com/in/research/documents/p469-ghai.pdf
[22] Jain, Mohit and Chandan, Vikas, "Softwar-only Occupancy Inference in a Workplace", 2016 December 7th, from: homes.cs.washington.edu/_mohitij/ISGT-2016.pdf