Wireless indoor positioning system for customer tracking

Back in 2013, a young team of enthusiastic engineers approached our MULTILUX team in search for a solution that would allow shopping cart tracking. They heard rumours about us having developed our own Local Indoor Positioning System - LIPS.

At that time, LIPS was not as mature as today, as it was basically a project under development. We lacked some showcase examples that would confirm how well LIPS performs in various cases. So we took the challenge and demonstrated that our wireless indoor positioning system can be used for retail customer tracking. This is how it went…

The problem

Supermarkets don't exactly know what customers are doing while they are inside the premises. How do they move around the shelves, where do they stop, what path do they follow etc. They lack the understanding of customer behaviour.

The solution

Our partners had a pretty straightforward plan - to develop a solution that would allow supermarkets to track trolleys in real-time. Such a solution may have brought many benefits to supermarkets:

  • CUSTOMER BEHAVIOR: Supermarkets could truly track consumers and analyse their habits - combined with the supermarket's loyalty card, such a solution could revolutionize the way supermarkets are approaching us with their offers. Imagine receiving only advertisements that really interest you... image

  • PRODUCT PLACEMENT OPTIMIZATION The possibility of optimizing how products are placed on shelves is something an indoor tracking system could make possible. By analysing movement of consumers, a supermarket could see which shelves need change or optimization and, therefore, optimally place products or add advertisements to boost revenue. Basically, they could generate heatmaps and see "hot" and "cold" areas... image

There are, of course, many more benefits once an indoor tracking system is set up in the enviroment. It is only up to the supermarket's marketing team and their imagination to generate new use cases that would exploit shopping trolley tracking data.

So basically, the guys who approached us with this idea had it all - they knew they wanted a customized tablet to be placed on the shopping trolley and had already picked one. They were also very good software engineers and prepared their own tablet app that would let the consumer upload a shopping list, and the tablet would navigate you through the store. But they needed hardware that would provide them with real-time positions of shopping carts.

The hardware infrastructure

First, we needed the floorplan of the supermarket where they intended to set up the pilot.

In order to provide accurate RTLS of indoor moving assets (which shopping trolleys are), we need to establish some infrastructure of detectors. Devices that are used as fixed references are called anchors. We need to place them on the ceiling of the enviroment where we want to track assets or persons. The higher we mount them, the more ground area they will be able to cover. Finally, we mount tags on the moving assets, allowing them to determine their position by means of communication with the anchors.

Using the floorplan and shelf dimensions, we were able to calculate the required mounting heights and distances between anchors. Here is a quick look at what we did:

image

We figured out that anchors needed to be mounted exactly above the shelf center at the height of 6 meters in order to fully cover the required area between two shelves. Furthermore, we suggested the anchors to be placed 3.5 meters apart, which meant they would need three anchors per shelf that was 7 meters long.

In order to confirm our calculations our engineers also prepared 3D simulations where light was used to simulate RF signal coverage.

image

With all the needed calculations we were eager to try the proposed solution in production environment. So we went to the mall.

Testing in the supermarket

The guys arranged that we could set up a pilot in a local supermarket.

image

As the anchors were installed, the time was to test how accurately could we track shopping trolleys using LIPS. Since they intended to fit each trolley with their own customized tablet, we proposed that instead of using a LIPS tag on each trolley, they could use a LIPS module and integrate it within their tablet device.

image

This LIPS module is the minimum possible device we can offer to anyone who wants to perform real-time tracking using LIPS. With its small size (35x30x6 millimeters) and well-documented pinout and APIs, it is ideal for developer use and can be fitted into devices such as custom tablets, gadgets, etc.

At the time of testing, they unfortunately did not have a tablet developed yet, so we decided to improvise. We used a LIPS module, closed it in a plastic enclosure together with some basic electronics, added an antenna, and connected it to a notebook computer that was put onto the shopping cart. The hardware configuration allowed us to track a single shopping cart in the supermarket.

image

Since there was no tablet, the guys of course did not have any software. But we found a solution to that, too. We quickly threw together some code, defined the shelves and prepared a graphical user interface that allowed us to see what we were doing - the black dot represented the real-time location of the trolley:

image

At that point, the testing was done and we proved that LIPS can be used to perform accurate real-time tracking of shopping carts in supermarkets. You can see the video on YouTube:

Therefore, when receiving questions about the possibility of using LIPS in supermarkets, we are more than confident that LIPS is indeed most suitable. We know the capabilities of our devices and our system, and we can't wait to put it to the test in different environments and show all its potential.

Would you like to find out more about LIPS?

see LIPS webpage