Why software will win the race for autonomy


Software is a huge disrupter. Examples are easy to come by – for entertainment, it’s Netflix, for consumer goods, it’s Amazon. AirBnB own no hotels and Uber own no taxis. But let’s be clear, what makes software brilliant is that unlike hardware, it has no mass – it is atom-free. In whatever industry you look at, software is both the disrupting factor and the innovation enabler. So why would it be any different for vehicles?

Oxbotica was founded on that principle. Folk have been making vehicles for hundreds of years and the fundamentals of vehicles are firmly established. We know how to make them cheaply and quickly, and they can last (almost) forever. Not much changes week by week in this sense. What is changing rapidly, however, is computing. And if software is what makes computers useful – effectively the cortex of the machine – then no prizes for spotting that software is central to the ways in which we run, use and leverage our vehicles. All of them.

Software scales in a way that physical objects don’t. In the context of software for autonomous vehicles, effort and bottom-line cost is pretty much independent of the number of vehicles you deploy to. The returns are also proportional to the fleet size, so you’re buying one piece of software to be used on 10,000 vehicles, rather than 10,000 self-driving vehicles.

Achieving interoperability in hardware is tricky, expensive and slow. Physical atoms get in the way, whereas digital bits don’t. Smart design and modern software techniques allow the creation of future-proof and open APIs between software components that enable collaboration and create an ecosystem of partners and contributors big enough to address massive opportunities. From day one, we’ve had this in mind in everything we do. We want to enable any third party to fully realise the potential of autonomy – from OEMs through to owners and operators – in ways that best suit them.

Independence and flexibility are essential for accelerating the adoption of any new technology. Autonomy is a technology for everyone, so we need to lower the barriers to its usage. Removing reliance on any and all infrastructure is one way of doing this.

Freedom to operate is vital and it takes on new meaning in the sphere of vehicle autonomy. We have built Oxbotica’s autonomous vehicle software to enable any vehicle to operate in any environment with zero dependence on external systems. That means no GPS is required. No third-party maps are required either. We built our stack from the ground up, component by component and targeted any vehicle – be that a truck, shuttle, tug, wagon or saloon working indoors or outdoors, it doesn’t matter. We didn’t choose a particular sensor modality – vision, lidar and radar all have their roles and can be used solely or jointly depending on application type. We didn’t limit ourselves to a particular way of technical thinking. We take pride in our algorithmic diversity, going both deep and wide, drawing on the best techniques not only from the past year, but also the last 20 years. That’s not always easy, but doing so creates a heady mix of ideas which mutually amplify themselves.

Over the past 5 years, we’ve deployed our software worldwide in a dizzying array of environments indoors and outdoors (whatever the weather!), from city centres, ports and airports to mines and forests. In doing so, we bring autonomy to different organisations and markets. Another example closer to home is our recent partnership with Addison Lee, London’s premier prestige taxi company. We will be providing autonomy to them for trials in London over the coming months. This speaks to our model of being an autonomy software provider into existing verticals (e.g. MaaS) rather than being the vertical.

The way we move people and goods is changing, and software is at the heart of this transformation.