Lucy Zodion

A referral from Amazon

It’s not every day we get a call from AWS, the cloud computing subsidiary of Amazon.com, Inc, asking us to help one of their clients with a major IoT project. We met with Lucy Zodion to learn about their vision for a connected city, before advising them on how to hook up their devices to the cloud, working together to sketch ideas, draw up specifications and create entirely new protocols.

Let There Be Light

Our initial project was to build a proof-of-concept, which is where we usually like to start with a large technical project. Let’s make the light switch on and off, and make it look relatively pretty in the process.

Lowering the Costs through Innovation

Following on from our initial project we set out to improve how streetlights operate while offering huge cost savings at scale. We introduced brightness control allowing operators such as city councils to save energy by lowering brightness or switching off bulbs when they weren’t needed. We reduced the bulb failures maintenance overhead by removing the need for costly ‘scouts’ to check if the bulb is working, letting the system report faults automatically.

Over-the-air Citywide Update

Updating the firmware on hundreds of thousands of streetlights is no easy task. In order to achieve this, we developed a bespoke over-the-air update protocol. Each light has a radio transmitter and receiver (using LoRaWAN). We took ideas from fast UDP-based transfer protocols and reworked them to within the limits of the technology. We can only send 50 byte (yes byte!) packets at a time. Combining this with an extremely efficient patching algorithm that can be used in low-memory environments, this brought the whole thing together.

WebGL Maps

Plotting lots of data points is tricky. Many tried to do this with Google Maps by collating the points together, then drawing a number with the number of points, requiring you to zoom in for more detail. This is clunky. Instead, we take a different approach. Using modern technologies we can actually plot all the points in real-time on the GPU of the computer. Using similar tech to that employed by Uber’s.

Automated Testing

A big part of this project is the automated test suite. When dealing with low-level binary protocols, you want to make sure that each bit matches the specification exactly. Sometimes this meant testing before firmware had finished being developed. To make this work, we developed simulated versions of the nodes themselves, which send and receive packets in the same fashion. A larger simulation was developed to simulate the actual airwaves, and clashing radio transmissions on the same frequency at the same time, much like in the way physics simulation would work in a video game. This can give us confidence in the rollout of the lights.

The Future

While we’ve been rolling out street lighting applications, we’ve noticed many parallels between this and other applications. Street lights are perfect for IoT applications, they are a source of power on every street, and with the connectivity we’re building in, can be put to work at a plethora of different tasks.

We think starting with this will give us a robust platform for future IoT applications, such as mapping out temperature for gritting. Being in the unique position of having a power source, and connectivity on every street, we can then look at EV charging, waste management, Air Quality, Gritting, Flood and traffic flow.

Adding additional sensors to the lamp can be done with relative ease, and it’s outputs monitored using the same tooling. Air quality is becoming a hot topic at the moment, if we can collect and map data it can help provide actionable insights for councils and other organisations.

Getting the data is the first step in solving many problems we have in our cities, and we’re all thrilled to see where we can take it to next!