AstraCity Solutions

IoT Data-Driven solution for smart city waste management and much more.

Comment

Inspiration

IoT smart city solutions often face one great challenge - connectivity. In this challenge, we decided to try and address that, by creating a multistep plan, on how to make cities connected, and truly smart. And doing all of that, in a way that local governments will like and be interested in, will save them money in the long (_and sometimes even short _) run, as well as won't cost a fortune of taxpayers' money to adopt. The first step of that plan is to be bootstrapped at this wonderful event.

What it does

Our first step is to make a city connected. For that, we chose a combination of machine learning, and custom hardware solutions, that use intelligent technology, to create a bunch of clusters, connected to each other in a network, capable of transmitting data to our (or the cities) server. Because of the route we chose, these clusters will copy and occur in places that they naturally should, as the city progressed through its development and growth. Onto those networks, we can continually connect a range of IoT solutions, to improve the quality of life of the citizens, at a rate, that the city is ready to adopt a new technology.

So how do we create these clusters and networks?

In order for these networks to work in the future, their clusters need to be rather dense, in order for other IoT solutions to be ready to be installed at whatever place necessary, and easily connect to the network. But we can't really just come and install a transmitter on the door of each household. Sure a lot of households have cars, so what about parking spaces? Well that wouldn't exactly work either, as they are a big problem in many cities, and there are not enough of them - hardly can we create a network spanning the whole city using them. And it would be really expensive. So what then? We realized, that one thing that 99.9% of households have in common, is that they produce waste and that waste needs to be thrown out somewhere - bingo. Garbage bins occur regularly enough to facilitate that network, while also not being so dense to drive costs up. And it works out great, as smart waste collection is also a big problem, especially in big cities. But it's not only that, only in Slovakia alone, we know of at least 4 municipalities that are currently looking for a smart waste solution, without being able to find a company to provide it for a reasonable price. So the first point of our plan consists of exactly that - providing a city with smart waste bins, that are all connected to each other. Those create the network, and to that network, we connect other smart cities' IoT solutions like traffic optimization, optimized public lighting, or accident prevention. With all of those solutions collection data and a potential public dataset, that small business owners can use to optimize their business hours, etc...

How we built it

We designed a smart garbage bin, that has 4 sensors, 1 ultrasonic on top of the lid, which measures the approximate volume of the bin, and 3 weight sensors, measuring the weight of the waste. Those sensors are connected to an ESP board. That board is then connected to boards of the neighboring bins, using low-range Bluetooth, creating, what we previously called a cluster. Then using long-range Bluetooth we connect to neighboring clusters, creating a network. What we think is the beauty of this solution, is only a few of these clusters in a network need to be able to communicate with our servers, saving costs on network adapters massively. This also helps us to get smart city solutions to areas, where the public network isn't available.

Then we need a server, to facilitate all of that data, process it, and save it in our databases. Because we wanted to have that done here, we choose NextJS as our backed framework, to be able to fast prototype. We have created an API with 39 endpoints, to facilitate that process.

But we need to bring value to the cities, even in the first step. That's where the utilization of the data comes in. We have created and _ sort of _ trained a machine-learning model, that predicts the fullness of the bin, based on historical data, and then decides if it needs to get collected. With that, we can direct garbage trucks, only to the places, where they need to go, just before the bin gets full, and becomes a problem for the citizens., while also saving the city money, and being more ecological.

How about a route generation for those garbage trucks, it wouldn't be good if they started going from one side of the city to the other.

We'll because, all of this was already a big challenge for a team of two, and we started to get tired, we decided to not invent the wheel, and use already existing solutions by Google, utilizing distance matrixes, and a bunch of cool stuff they do on the background.

And because we are very people focused in our solutions, and we know that some cities reward their citizens for recycling, we decided to start optimizing that as well, and we are _ sort of _ prepared to do also that and match changes coming from the waste bins, to individual households, automatically calculating the deduction from their taxes they'll get at the end of the year.

Now, it's almost 9:00 AM, we both didn't sleep, and consuming unhealthy amounts of coffee, but we realized, that except for thousands of lines of machine or serverside code and city network database prepared to be easily modeled and extended in the future, we don't have anything visual. _ Short state of panic, we are going to bootstrap something_.

Here we go, almost 2 hours later, we prototyped a very simple preview of the core functionality of the waste management administration panel, particularly (not particularly, but exactly that and nothing else) generating the route on request, and displaying it on a map, one click away from getting automatically navigated with Google Maps. And as we do not like to forget about people, we also have a very light prototype of a mobile app, that displays for users, their savings, approximation of carbon emissions saved, etc. The app also allows users to get Request a Pickup of their garbage, which solves a completely different problem, but as it's 10:55 AM, we do not have time to talk about it, and probably also do not have time to answer the rest of the questions, but we encourage everyone interested to come and talk to us, as well as judges, to come and see.

Challenges we ran into

Machine learning, data, time, full bathrooms

Accomplishments that we're proud of

How much we managed to actually do, considering the timeframe, and our reduced numbers.

What's next for AstraCity Solutions

That is definitely something we want to pursue, in the near future finishing what we started here of step 1, going to the municipalities, and persuading the politicians in charge, that this is a way of the future.

Also I forgot to mention somewhere at the top, that we do not need to buy new garbage bins, but modify the existing ones, reducing costs even further.

And monetization? Monthly subscription, duh

Built With

Share this project:

Updates