Light Sensor for Coca Cola CoolerHack2016
High Level Architecture
#How does it work?
The solution consists of photocells that detect light, or, if covered (by a can or bottle), the presence of a product. The photocell sensors are installed in the product rack columns and will read a value of light (if no can or bottle is detected above the sensor, or dark if a can or bottle is present. The values (light or dark) are captured in real-time using an Arduino device and output to a serial port. A Raspberry-Pi monitors the serial output of the Arduino and determines inventory for a given product (added or removed). The Raspberry-Pi transmits processed information reports over over a private cellular network connection to a (cloud) database, so that analytics/KPIs/metrics can be gathered and alerts can be sent to stakeholders/suppliers. The solution also includes a front-end web application that monitors the MySQL database in real-time and illustrates movement of inventory (added or removed). Because existing shelving in coolers of all sizes and formats can be easily retrofitted for this system, this solution is the most affordable option, while still including real time data, for the over 14 million Coca Cola coolers deployed across the globe.
# What does it measure?
The solution measures light to determine presence of a product in real-time. Photocell readings are evaluated based on a configurable threshold that is tuned for specific lighting conditions.
Does the retailer/bottler/distributor have to do anything different to make your solution work? (if yes, please explain)
The solution will require minor modifications to the existing process. The product shelf columns will need to be retrofitted with the photocells and the coolers will need to include the hardware/application components (e.g. Arduino/Raspberry-Pi). This would be a very simple process, easily replicated by distributors or retailers. It is assumed that upon initial installation/setup an application will be used to ‘commission’ the cooler/solution which would require administrative setup (e.g. planogram). Additionally, future coolers could ship already equipped and pre-wired replacements shelving could be provided as well.
What materials does your solution require?
The materials currently being used in the solution include the following: photocells, resistors, Arduino, Raspberry-Pi, USB/Serial cable and Wi-Fi USB adapter.
Best guest on cost to implement
When considering volume, we have estimated that the total costs for manufacturing would be $20.00 or less per cooler. We believe that there will be implementation costs with regards to installation/administration, operations and maintenance of the overall solution. The installation/administration costs are estimated to be 1-2 additional service hours per cooler, while the operations and maintenance costs would vary depending on the scale of the solution to be deployed.
Challenges you/your team ran into
During the initial design discussions, one challenge included finding a technology that would solve all aspects of the problem at hand – we examined the possible use of cameras, RFID tags, crowdsourcing, etc. but it quickly became evident that photocell light sensors were the most cost effective solution. In having decided upon photocells, we found that the lighting conditions (e.g. low-light, reflection, etc.) can skew the photocell readings, which required us to build a solution that could be configured for varying thresholds to support varying conditions. Lastly, we found that the solution was best implemented when the sensors were under a particular product rather than on the side.
What you/your team learned
This experience has enabled our team to appreciate the use of low-cost materials/equipment to build a solution that can quickly scale. We have also found that the use of light to detect presence may be applicable in other market areas for a wide-variety of uses, areas that will begin to research further.
Our next steps include refining the solution / approach we have designed. We firmly believe that there are areas for improvement in the overall solution that we could consider given the appropriate time and budget. We intend on moving forward with this project to identify viable partners/sponsors/market areas where a technology using real-time presence is directly applicable. We would love the opportunity to work further with Coke and help them tackle one of their biggest operational issues.
Anything else you want to add
The design we have defined was meant to support a large-scale implementation effort. We focused on a solution that will go beyond the initial prototype/demonstration. Our solution includes design elements behind the scenes (e.g. product/cooler scalability, database management, configuration) to support a wide variety of deployments in locations with varying conditions.