There has been some progress since I published Growbotics AI Vision in February 2019, and I’d like to summarise this progress and share with more people.
In the Growbotics AI Vision, I mentioned seven principles that are important to this project; however, instead of reporting progress on each point separately, I’d like to use a slightly different format and group updates in a way that it would be easier to read. Also, instead of publishing one long post, I will release ten smaller updates and will summarise what I learned, what I worked on, what failed and what’s coming next.
So this blog post perhaps could be seen as a Table of contents with just a hint on the latest status, and I’ll update it with links in the specific sections once more content becomes available.
The Cube was the starting point of this project. The goal was and still is to explore the vertical farming, hydroponics and aeroponics. The idea is relatively simple in practice — develop the stackable cubes (miniature greenhouses) that could fully function of their own, but also work in various formations. However, in reality, it became apparent that to prototype fast, release numerous versions, it’s not quite possible with just a drill and a hammer. Learning from all the pain and trying to build something that works led to more investment of time into Digital Manufacturing. There’s a dedicated section just one that.
Despite the difficulties, the first version of the Cube is near the completion. I now crystalised a lot of thoughts on which materials to use in constructing the Cube, how to produce nutrients, etc. So if you’re interested to hear more, I’ll publish a more detailed update just on this section.
I had a couple of courses on Artificial Intelligence quite some time ago at university. It was interesting, but other competing subjects and technologies took priority. More recently (since 2015), I decided to invest time and effort in gaining more knowledge related to Machine Learning, and it’s latest developments, especially around Deep Learning. To do so, I took multiple online courses, read various books, followed some people, attended related conferences and so on.
A more detail post on this topic will be on sharing what helped to gain more knowledge. How the progress in Computer Vision, Deep Reinforcement Learning, etc. could be applied in practice working on the goals of this project.
The electric gardening cart, autonomous mapping drone and the robotic arm aren’t quite tirelessly working day and night (yet). However, there’s been quite some progress made in this area that is worth summarising and sharing.
The goal is to build an electric gardening cart in as simple as possible way. Then control it programmatically and use it as a base for various activities that would require navigating outdoors. A basic version is already on its wheels, but some wiring still needs to happen. And more time will have to pass before it can navigate autonomously using just cameras. As it won’t have to navigate through the streets of London, this task seems rather doable.
I haven’t assembled the drone yet, but I do have most of the parts, and it’s just a matter of finding some time to complete it. I mainly intend to use it for the autonomous drone missions for mapping purposes. The heavy-lift drone is currently out of scope, but I do follow several people who are actively developing them.
The third component under this section is a robotic arm. I already purchased all the parts, and I have received them. The task of 3D printing other parts is also complete, so it’s again just the assembly that is waiting in the queue.
More on these three pieces in the detailed post a bit later.
At the moment, we are getting ready for the second winter! From knowing absolutely nothing about the bees to the current point, is a massive jump. This activity took an enormous amount of time and effort. I’m glad it becomes a bit easier with time and practice. Also, all the knowledge gained trying to understand how bees live their lives and build a symbiotic relationship, avoiding the use of chemicals is highly rewarding.
Besides what I already mentioned, the feedback from some lucky friends who already tasted the honey was very pleasing too.
There are some exciting opportunities in applying Machine Learning in this section, so more details to follow.
An attempt to grow mushrooms indoors on the used coffee grounds failed, but the desire to learn how to grow mushrooms didn’t die at all. There will be more trials and errors on this front. I have an idea on what could have possibly gone wrong so will share it in the more detailed post, and hopefully, there will be someone out there to educate me a bit on this.
Talking about education, I did attend a dedicated workshop on growing mushrooms indoors, so I’ll share the experience (which was very positive) later in more detail.
Worms, worms, worms. My worms are still alive, and they are lovely. Aren’t composting worms the true heroes? They can take our food waste and turn them into nutrients that people can use to grow more food and repeat such a cycle. I think this is an excellent example of a Circular Economy.
I received the first mix of worms a couple of years ago, and by now, I acquired experience of collaborating with them to process food waste and produce nutrients for the Cube. The worms aren’t demanding creatures, temperature and humidity control is essential; certain food types such as meat, dairy or citric fruits are not good options. Besides that, it’s pretty much straight forward.
The Great Struggle of building the Cube in a way led to investing more time and effort into Digital Manufacturing. It’s also quite rewarding when you design an object in your laptop, press a “Go” button and watch the machine manufacturing the item for you. I already had a tiny 3D printer when I started this project, but later I made a lot of progress adapting larger 3D printers, modifying them, assembling CNC machines and so on.
Another important aspect was to learn the required software that is at the core of Digital Manufacturing.
Solar, Wind & Batteries
I attended a couple of workshops related to this section: one where we built a solar panel individually (led by the instructor) and another where we made a wooden wind turbine together. Before these workshops, I already worked on a simple prototype — installed solar cells on the custom-built wooden blinds and 3D printed a vertical axis turbine. I also learned how to access all metrics from the Solar controller and visualised it in the Home Assistant. In addition to all this, I built a custom 48V battery pack. I combined 16850 battery cells in a 3D printed frame, connected them using spot welding and attached Battery Management System (BMS). This battery is currently used to power the electric cart. There are more plans under this section, and I’ll be happy to share more details.
An idea of owning a small size digital factory (call it a Kilofactory) is quite appealing to me for this project, but also the future ideas as well. Perhaps the easiest way to start would be to buy a 3D printer and start printing. However, the appetite grows with time, and I can see other people too who started with 3D printing and later expanded their horizons, bought and assembled CNC machines, lasers, plasma tables, built injection moulding machines, etc. Some of such devices already are fully compatible with Digital Manufacturing, and others perhaps could be turned into computer-controlled machines. A more detailed post in this section will be in a way similar to Digital Manufacturing post, but focused a bit more on which devices and why I decided to build or buy and in what order and what else is on the wishlist.
I have to admit that the Business Model isn’t the current focus. Also, this project isn’t about creating something, making it profitable, then selling it and moving on. This project is a more fundamental look at how we grow food, how can we use modern technologies to avoid using dangerous chemicals. It’s also to look at how we could collaborate globally, produce and consume more locally so that people don’t have to fly lettuce or an apple that is grown on one side of the globe to the other side of the world (which would also reduce CO2 emissions). It’s is also to explore new ways of working. Given the progress we have made on Internet speeds and accessibility (which should further improve once Starlink becomes operational) combined with modern collaboration tools, there’s no longer a need to share the same physical office from 9 am till 5 pm.
Before the pandemic, I did regularly attend the Ethereum London meetups in Imperial College London for a couple of years. There’s still quite some inspiration from those times, so I do have some concrete ideas on how to fund this project in the future so it can be sustainable on its own.
Also, if you are interested and haven’t seen yet, please have a look at my previous posts related to this project:
With that, I’d like to thank you for your time, and if you’re interested in the details of one section more than the other, please consider leaving a comment below. Also, for more frequent short updates, I invite you to have a look at Instagram.