Lab to field — a journey in robotics
Build up
In 2020, when the word pandemic was still confined to thriller movie plots instead of 24 hour news cycles, I was thinking hard about the next work move. I’d left my graduate role with a large corporate a few years prior along with it’s stable, sensible approach to professional development. I was impatient, ready to meet people, travel and see engineering in action and learn about the business of automating processes. The team at Beijer, a Scandinavian automation provider gave me the opportunity to develop an emerging, ambitious market around industrial IoT. I joined the UK team who gave me freedom, trust, support and lots of opportunities to learn and grow. By early 2020 I’d learnt about business development, getting results, but also feeling further from hands on engineering solutions. It was too soon to give up on this part of my professional development, I knew I had to keep my technical skills sharp.
Northcoders, a software development programme I’d joined just as the world descended into lockdowns, provided the opportunity to immerse myself in learning about open-source software development, remote teams, early startups and personal projects. Taking on a new type of risk and accepting relative uncertainty in my work felt less outlandish when the world outside was going through such unprecedented turmoil. Tuning into my intuition, I knew I was on the right track to join a startup and see a project grow.
As well as the relative challenges of lockdown (spare a thought for those key workers putting in a real shift), it was a time for introspection and contemplation. After years of travelling every week, I suddenly had time to pause, reflect and pivot towards work that would keep me engaged, motivated and satisfied. In the language of tired business speak, it was time to come up with a mission statement. A naive, inaccurate attempt at best, it would have went something like this at the time:
I want to use my engineering-problem-solving mindset to offer meaningful solutions to some of our most pressing, systemic problems — around climate, food and our environment.
With these scribbles in my notepads and travel restrictions easing, it was time to take a train journey to Europe to see family and friends and work on my next move.
Remote start
I met the E-Nano founders during the summer of 2020, remotely. We had good talks and I joined the team with a product manager and developer. We started building a cloud-based robotics platform with relevant UI and infrastructure. Productivity was high, people were getting ready for a long winter of isolation. There was a certain ‘modern, remote startup’ smell in the air that kept people engaged. Despite the distance in timezones and geographies, our work had started to take shape, but so was the end of our runway. During that time there was real focus on building the future of the company with many personal risks taken. It paid off and further funding allowed us to plan the next year in London. By early 2021 I’d started to get familiar with the robotics architecture of the project. It was time to get into the lab and start building physical interfaces — robotics APIs — between our platform and 3rd party hardware. Bench-top proof of concepts were being shaped into working prototypes capable of automating meaningful data collection (and other essential) tasks.
Into the lab
In the lab, we started integrating 3rd party sensors with our robotics operating system. This meant standardising on an edge architecture in terms of electronics modules, embedded code and interfaces. Mechanical integrations were a constant feature of our calendar and so was the noisy metalwork echoing around the bowels of Somerset House. Our systems started to move, we developed controls and interfaces. A series of field visits to our turf research partner in Yorkshire meant real-world testing with added environmental challenges (read, rain). It was good quality, dynamic, hands-on engineering. Late in 2021 we had a number of early demos with our first prototype, Sprout. This proved invaluable in terms of learning reliability lessons. After a re-think, the winter was spent on the re-design, sourcing and building of our second, more robust and refined system. Many of the critical and more practical parts of the solution were overhauled. For our fast approaching field trial, we would be entering with a much more capable robotic solution.
Fresh air
Our robot was working in lab conditions, the flow of information from sensors to our data platform proven — ready to capture some early feedback from real-world users. A Premier League football club provided an area at their training facility and we collected data from pitches every other day for 3 weeks. This raw data and feedback pushed us forward in terms of developing our reporting UI, core robotics functionality (sample resolution, grid layouts) but most importantly, it allowed for a window into the world of elite grounds management. This type of learning is impossible to simulate in an academic or even commercial research environment — developing the personal relationships means trust is built and information starts flowing from the user to the engineer. This ability to capture feedback, iterate on it and offer an ever-improving product and complementary services has allowed us to build confidence in our collective abilities to deliver value.
Where next?
I’m happy and proud of the role I played in the team getting the project this far. No mean feat — developing hardware-based robotics solutions is complex, risky and costly business. Software startups come and go, but to build something that moves around and delivers tangible value in the real world is a different story. The first step of this journey — building a working prototype, testing it the lab, iterating on it, then seeing it take on real-world environments is quite a moment. But of course the work doesn’t stop here…
For E-Nano, the immediate focus is to gain early clients in the sports industry, continuing the product development iterations and start generating revenue. I’m positive about their success, a hard-working, technical and committed team, pragmatic problem-solvers who’ve learnt quickly about their market. The type of engineering problems to solve on the agenda are related to scaling up — building, deploying and supporting multiple units with early clients. It continues to be a fantastic opportunity to keep learning and growing.
We are starting to build our network externally — with the market, sure — but also with academic institutions, other startups, corporates and investors. We are interested in where robotics is heading, we want to understand the innovation landscape in the UK, we want to engage the next generation of engineers to helps us solve meaningful problems.
This final point is the most personal. Elite sports is the perfect training environments for developing robotics — good infrastructure, helpful people, immaculate playing surfaces (kind of the point!). There are many efficiencies and quality improvements autonomous data collection tools can deliver. But in parallel to this, I feel my ideas to build impactful, scalable solutions to some of our most pressing issues — climate, food, environment — are starting to take shape. These themes intersect in the way we produce food, something I’ve been proactively learning about for the last 2 years. But how to make these two worlds of engineering/technology and agriculture overlap?
To bring robust, reliable, worthwhile robotics into agricultural environments, a few things must happen. Perhaps least worrying for an engineer is ruggedising their solution — better suspension, navigation, finding our way back home — all technical issues that can be solved. What’s more critical (and tricky to achieve) is getting feedback from the end-users, the clients, in this case the farmers and growers close to the land. It’s clear that no farmers can justify funding robotics R&D that may or may not provide direct benefit to his or her operations in the not too distant future. Yet the value better tools can help deliver could very much benefit them. This is an area where there is concentrated effort in building solutions — public and private investors, researchers, non-profits and companies are all trying to bridge this gap. Personally, I see the quickest route being simplifying the value proposition to a point where the core of the solution can start to deliver even a very small, measurable impact. Once this feedback loop is implemented, iterations can follow and scaling can happen.
The task taking shape ahead that I want to engage in is to work with my network to propose a series of experiments. I’d like to put appropriate technology solutions to use on working farms, making life easier for people working hard on regenerating our natural environments. To push this forward I am starting more conversations with farmers, entrepreneurs, makers and frankly anyone who can align with the above. If your mission is to push for more ecologically sound practices in agriculture I want to connect, meet and exchange thoughts with you. Do get in touch so we can all feel like this!
Originally published at https://github.com.
