Blog

How Motion Trails Improve Commercial Space

May 11, 2016
Enlighted Inc.

Part of what makes the IoT so interesting, from a data science perspective, is that we can capture time-synchronized data from a variety of measurements such as temperature, motion events, or energy usage, and then use statistical techniques to discover important connections between them.

Because the built environment is so complex, we often take for granted behaviors we observe that are difficult to express in the code that drives our products and ultimately makes them more intelligent. Although we still make site visits or do manual checks to verify that our results are robust, we no longer have to rely on qualitative observations, such as surveyors taking hand-written notes, as a baseline.

That method was challenging for a number of reasons. First, it was expensive and time consuming. Also, it only provided a snapshot of the building at a single point in time. If we wanted to see how behavior was changing over time, we needed to return to the building and do more observations at intervals. This meant additional time and expense, and also introduced delays. It was a far cry from anything resembling real-time monitoring.

One benefit of distributed sensors, specifically Enlighted’s sensors, which are installed in the lights and are gathering anonymous data around the clock, is that we get a continuous spatial “map” of a floor. However, as a data scientist, I find it’s important to address the common misconception that (since we can perform spatial analyses) this means we know precise data, at any given time.

Think about the “Doppler Radar” on the Weather Channel. The colorful map covers the whole area, but really it reflects a grid of known meteorological stations and the relationships between those that are next to one another.

The difference between the Internet of Meteorological Stations and the Internet of Things is that we’re operating in a constrained space, and because many of Enlighted’s sensors are incorporated into the lights, our grid is relatively constant. This means we can perform real-time monitoring and analytics with a level of precision impossible in the natural world. This rich and more continuous data is very exciting!

We’ve developed an application that allows building operators to see a snapshot of how a building is being used at any point in time, and also to see how space utilization changes over a period of time, with the click of a mouse.

Figure 1. View of sensor locations in an Enlighted space

Figure 1. View of sensor locations in an Enlighted space

When doing this kind of space monitoring and optimization using IoT, one thing that makes Enlighted different from other building monitoring systems is our use of motion trails, rather than just heat maps.

The main difference between the two is that a heat map shows aggregation of occupancy measurements over a given period of time, while motion trails show the aggregate movement of people through that space over time.

In other words, heat maps are tied to places — specifically, the places where we have lighting fixtures — while motion trails are tied to events, these events being the occupancy change between neighboring lighting fixtures within walkable distance and time of one another. Motion trails represent what people actually do in a built environment, and this is a critical distinction as we think about business analytics and the benefits and ramifications of space usage.

Figure 2. Heat maps show where people spend time, but not how they move through space over time

Figure 2. Heat maps show where people spend time, but not how they move through space over time

In our system, motion trails use the same data that the sensors use to compute occupancy in a specific space. For example, when a worker returns to their desk, the motion detection algorithm recognizes re-occupancy and activates the personal lighting profile for that workspace.

The way this works is that the sensor has a knowledge of the “background” containing the desk and other stationary objects, and when the person moves into that space, it detects a difference and alerts the EM to change the profile. Our sensors communicate so often and so rapidly that this communication is seamless.

The motion trails view in Enlighted chains together a series of these individual occupancy detection events as they occur, and then uses a sophisticated algorithm to draw a path that shows where and when a body moved through that specific space.

The simplest example is if an occupant is detected in a long straight hallway with one sensor every 10 feet, and the hallway is otherwise empty, each sensor will see 1 occupant one after another subsequently at the pace the occupant is walking. That is a path over which we can draw a motion trail.

By mapping all of the paths of all of the people traveling through the space over time, we can see how people navigate and use a space, and make space planning decisions that optimize the space for these people. This saves time and money for building owners and operators.

Figure 3. Motion trails show how people move through a space over time

Figure 3. Motion trails show how people move through a space over time

Here are some specific examples:

Consider a room with several sensors and people seated at desks in rows versus one with several sensors and desks in a circular formation. Imagine that some of these desks are pushed up to the wall on one side. You might find that like in a movie theatre, people are less likely to occupy the seats that are more difficult to walk to, but in a circular configuration, people would use all the seats. A heat map, which shows a static view of a room’s usage and no motion, might show people only using half the room, but wouldn’t explain why, which means it’s not very helpful; the motion trails would indicate that people only moved between sensors in a lateral direction along the desk rows, and that a rearrangement might be in order for the space.

Consider a second example. Several small rooms are next to one another, with varying levels of use. They are labelled as rooms A to G. If the furthest off room has very high use for some reason, it will distort a heat map because of it’s weight. If there’s a 20 person conference in Room A and no one in rooms B-F and then 4 people in Room G, the heat map will still show a gradient of use between rooms A and G, as if people were there, even if no sensors were triggered. In a motion trails image, we might see that there are no trails between rooms A and G, rather that A has (for example) an external entry-way and that really no one is passing between it and the main corridor.

Lastly, consider a space like an elevator terminal on the 15th floor of an office building. We might see many motion trails emanating from the terminal, going to other parts of a floor. If it were just a heat map, we could assume people love to spend most of their time in the elevator waiting area, since it would almost always be occupied and very well utilized. But motion trails will show this as an origin point for many paths going forward, indicating that it’s not a place where people remain, but a place of entry and exit.

Motion trails are a standard view in the Enlighted Space App, which we help configure for our customers. Not only can you see the motion trails as shown above for a full day or a full space, but we can apply different styles and filters to make the trails specifically applicable to individual situations.

Customers can also customize the view to show different color schemes, times of day, and more. For example, a customer can look at motion trails only for specific time periods, or specific spaces that are in use, like conference rooms.

By overlaying motion trails on the floor plan, a savvy customer might also think about things like when to change the carpet, schedule maintenance, or promote corporate happiness by engaging in certain activities based on the level of use.

Already this type of analysis has been performed by customers looking at, in one example, carpet replacement, another way of saving money and energy down the line.

How would you use this technology?

Share This Page
TAGS