Faster, Smarter, Better: MySeat’s Automated Sensor Programming

Boosting Business Efficiency: MySeat's Innovative IoT Sensor Setup

Electronic | IOT

MySeat develops and sells space utilization analysis services for businesses. The goal? To measure if workspaces are properly sized, validate the usage and size of meeting rooms, and know how many employees could be hired before having to change premises

Project

MySeat sells / rents and needs to deploy hundreds of IoT sensors to equip workspaces. A tedious step is the programming and deployment of these sensors. I have participated in deployments of 800 sensors to be completed in a single night, which is no small feat. Furthermore, each sensor must be programmed specifically for each client.

To simplify this procedure, I was able to create an automated programming bench connected to a label printer to print QR codes containing a unique identifier for each sensor. The QR code is then attached to the sensor housing and allows each sensor to be identified via a web application accessible via phone.

The Programming Bench

Raspberry Pi

A Raspberry Pi was used for programming the sensors. A serial programmer is connected, and Python scripts automatically detect the presence of a sensor to program it. A database of programmed sensors records each new programming with the sensor ID and an error code.

Brother QL-570 Label Printer and Camera

At the end of the programming, a QR code is generated by the Python scripts and printed using a label printer connected via USB to the Raspberry Pi.

A USB camera also allows scanning the QR code on the Raspberry Pi using image recognition and the OpenCV library.

3D Impression

In the end, since MySeat only needed two units of the programming bench, we chose to design a casing on Solidworks and print it in 3D.

This casing, approximately 20x12x8cm in size, integrates:

Raspberry Pi
USB Camera
Electronic connection board to the sensors
Control buttons (program / erase)

The Result

At the end of the project, we measured the gains from this programming bench:

Programming time reduced from 60 seconds/sensor to 20 seconds (3 times faster, resulting in a gain of 9 hours for 800 sensors).
“Easy to scale“: Just print a new casing and duplicate the Raspberry Pi’s SD card to have a new bench ready to use.
Simplified setup: Before the bench, programming required a PC with many libraries and IDEs to configure.