Case study: How, against all odds, we created an IoT device for cities

Implementing new technologies is a complex process. In addition to creating the technology itself, there is the human factor. People are not always able to immediately grasp the essence of changes and accept them, even if it makes their life easier. This idea was especially noticeable in the project to create an ISA device. This project aimed at developing cities according to the Smart City principle and focused on simplifying the control of meter readings collected by housing and communal services.


The city wants to develop, but what about its inhabitants?

When we got a project idea, we had been working in the software development market for more than ten years and tried to do hardware development. We had a customer with a software project. During our long-term work, we learned that our client could not solve the problem with hardware development. He turned to various hardware developers. Every other convinced him that it was impossible to make an acceptable or inexpensive solution, that they were experts, and nothing would come of this venture. So we wondered since we have already worked with software for this domain.

You will learn how we implemented this project below, but the result deserves interest. We created our hardware development department, created a device for the customer and continue to work on this project.

Are you sure you want to hear all the details?

The qualitative implementation of the project requires a business analysis stage, which is particularly important for a serious task. We started with BA. We learned that in residential buildings, there is a planned replacement of old meters with smart ones. The plan was as follows: to install 500 units per year. It was possible to install 250 units, and it was necessary to install 10,000 units. At such a pace, the plan in 20 years turned, at best, into 40, or even never at all. In addition, the same scheme for collecting indicators looked “slightly” outdated.

It turned out that some “unknown” number of people were walking around the city and taking meter data. At the same time, the city’s power company is doing the same. Due to the difference in the removal dates, the indicators do not match, leading to conflicts and the inability to track the amount of electricity consumed.

During the verification process, it turned out that some metering devices were not installed at the facilities in principle, but data were regularly taken from them. Because of such situations, the relevance of all data tended to zero.

If it were necessary to make the most unreliable system for the operation of something, then this would be its example.


How to solve the data transfer issue and not pay several billion for it?

The idea of the project was simple – to simplify the data transfer procedure into three steps: take a picture, recognize and transfer the data to the server. The critical problem was cost reduction. Otherwise, there is zero practical sense in such a solution.

The situation turned out to be interesting. In the process of in-depth market research, we found the following: most meters are either induction or already smart. They have a digital interface and the possibility of remote transmission. But in both cases, there were nuances. Smart – super expensive, and induction gave a significant error.


If the others can’t, we’ll do it ourselves

The solution was to use the standard set of components. That is, hardware engineers (in Ukraine, by the way, these are electronic engineers) used a relatively common set of parts for the board.

In searching for the correct answer, we were so carried away by the idea that we already intuitively felt this could be done. Having hired our hardware-electronic engineer, we set him a task. By the way, our new teammate was very optimistic.

A hot young engineer calculated everything, and we concluded that two months would be enough. And as a result… we spent more than a year on the first version. A few more months were spent adapting the device for operation in actual conditions.


Half the battle is to invent, and half the competition is to implement

Having reached the stage of practical use, the question arose – which camera to choose. It seemed like the best option was to select the cheapest one, but then we failed. It turned out that the form factor of the cameras is very different. As a result, they went through hundreds of cameras but found the necessary one. We installed it. And then, it turned out that you need to read the readings but do not remove the device from the meter. We will not describe the entire path of searches, selection of the required dimensions of the device, calculation of the focal length, and struggle with exposure. One of the main problems was the lack of connection at the counter installation sites. Many meters are located in the basement.

After the first report, it turned out that half of the devices could not send information. Of particular pleasure were the public utility workers. The illuminated meter became a topic of conversation for them for a long time.


What did we finally get?

Throughout the project’s development, there were many reasons to refuse and stop, but we did not give up. We took on this project and confidently moved forward with our clients, supporting each other through tricky development turns. Fighting humidity and temperature regulators, setting up a data recognition system using neural networks, fighting with the local population so that new technologies are not hyped for spare parts, and much more.

Not to mention the manufacture of the devices themselves using 3D printers. Of the required five, there were only two, but it must be necessary. Then our 3D printers appeared in our arsenal.

Now, ISA devices are actively used by the state energy company “Oblenergo” to take meter readings. In the future, it is planned to develop systems to transition all city services to the concept of using smart devices.

And perhaps, the central insight of the project was thinking out of the box. In an attempt to find outsourced performers, we often encounter blinders. At the same time, our goal was to complete the project in any way and fulfil our obligations to our client. Therefore, despite the beliefs of hardware engineers, software engineers and neuroscientists, the project was successful and integrated into urban systems’ operation.