Smart Robotics — what’s all this then?
Technology never sits still. That’s both a good and a bad thing for us here at Cambridge Consultants. On the one hand it keeps us in business, there’s always something new we can help a client with. On the other hand, it means a constant effort to stay at the forefront of technology, right on the edge of things that are ready to be implemented into products this year or next. With that background, we’ve decided to invest some time this year into the development of smart, mobile robotics. By “smart”, I mean cable of interacting with people, and able to respond to a changing environment. By “mobile” I want to focus on wheeled, ground based robots, but I expect much of the technology will be applicable to drones and other vehicles. And by “robotics” — I don’t want to get hung up on this definition, but anything that falls into the classic paradigm of “sense, think, act”, and crucially, is innovative and exciting, counts as a robot for me.
Cambridge Consultants have a battle-hardened product development process, one we’ve used in a myriad of forms for many different clients. It’s achieved great success when we’ve used it before, so we’re using it on our internal project too. Once we’d had the idea of Smart, Mobile Robotics, the first step in this process was to look into possible markets, to see if a need existed for the concept. After all, even though most of the people on this project are engineers, and would rather be digging into lines of code than worrying about grubby commercial realities, nobody likes working on a project that goes nowhere. Fortunately, we were able to identify a couple of target clients and a few broad market areas where the sort of technology we were interested in could make a difference. The next step in the process was to identify the knowledge or technology gaps — if we were going to make a smart, mobile robot, what tech would be need? Does it exist or do we need to create it? Next, we spent some time thinking about what an example use case for the robot would be, and then that led to defining the split between an overall, generic architecture for the general mobile robotics problem, and a specific architecture for a potential product based on our use case.
Robotics crosses several market sectors
Mobile robotics is a growing sector, across a variety of markets. Just a bit of research on crunchbase shows the amount of money going in this direction, and more detailed research into market reports confirmed it. Sectors such as agritech, construction and security are ripe for new robotics products that improve the experience of workers or customers. There are key drivers around labour shortages that mean the commercial imperative is there — and that’s vital for us. Employing CC requires a significant outlay of cash, so it only makes sense where the potential upside is large. However, when the upside is there, we can be incredibly valuable — many of our clients have seen massive growth on the back of work we’ve done for them. Looking at what technology is out there — there’s plenty of investment already happening in drones and autonomous cars. We want to play in a slightly different niche, where ground-based robots have to closely interact with people to be successful.
Aim of investment
One of the challenges with an internal investment — especially one such as this, where the aim is to grow our skills and experience, and push the art of the possible — is knowing when to stop. We don’t want to make a product (we don’t sell products, we sell design services), but we do want to make a tangible, physical demonstrator. We need to find a way to embody the key innovations, without having to do the high-effort, time consuming engineering necessary for turning a prototype into a product. To borrow an analogy from one of my engineers — if you were making an anti-gravity car, there’s tons of engineering effort needed to make the car part — design of the exterior and interior, safety features, aerodynamics etc — but if you haven’t demonstrated the anti-gravity part, you’ve got nothing. Designing a car is definitely possible — it’s not easy, or cheap, but it’s a known process — so if you were tackling the anti-gravity car, you‘d start with the anti-gravity part!
That’s what we’re doing here, tackling the hard part up front. We think there are some key technologies that need to be invented, implemented from existing academic papers, or brought together in order to get a useful mobile robot. By going through a rigorous design process, we’re going to gain lots of experience around the technologies, as well as hopefully pushing the boundaries somewhat.
Our goal
The challenge we’ve decided to set ourselves is that of a mobile robot, about the size of a large child or small adult, that can autonomously navigate around pedestrian areas. We’re avoiding “automotive” challenges and keeping out of fields (but we’re tackling both those challenges in other projects here). We’re imagining something that could be used in environments like a mall, theme park airport or hospital, and its purpose would be to lead or follow people, carry or deliver small items, or provide information and interaction. One of the key challenges to achieving this is performing well in crowds. The places we’re targeting will be full of people, and any new robot is going to be immediately mobbed by curious people. We’ve got to do better than simply stopping whenever someone comes within 1m of the robot, but we’ve got to do it in a way that’s safe.
