How Companies Can Rapidly Accelerate Their Automation and Autonomous System Development

Cambridge Consultants is a Business Reporter client.

Automation and autonomous system development do not have to be prohibitively expensive or time-consuming. That’s the clear message from Niall Mottram of Cambridge Consultants, as businesses continue to react to the unprecedented disruptions to supply chains and consumer buying habits caused by pandemic restrictions. 

In an interview with Business Reporter, Niall explains how this disruption is creating opportunity, which is causing companies to look at automation and autonomy with entirely fresh eyes. However, their view is often obscured by a common barrier: how to deliver ambitious automation and autonomy projects on budget and on time. 

It is clear that building differentiated, industry-leading autonomous systems does not work using off-the-shelf system integration solutions. Every business is different, with unique needs and use cases. But to date, building bespoke or tailored solutions has proved costly and time-consuming. It’s no surprise that many companies have hesitated to seize the opportunity. 

What’s the reason for the sky-high costs? It all comes down to the fact that such systems need to be intelligent. Intelligence requires training and training requires data—and the more diverse your use cases, the more data you need. Crucial data acquisition is hugely expensive. For instance, an autonomous car must be driven thousands of miles, or a robot needs to be tested and simulated in model warehouses, and the cost can go from millions of pounds well into the billions. Very few companies can afford that. 

But, as Niall reveals, it doesn’t have to be this way. Cambridge Consultants believes in a future unconstrained by current thinking. By challenging conventional approaches and drawing on its deep expertise, it has developed a unique modular tool for autonomous system development. Data acquisition is performed in a digital, simulated environment instead of a real-world setting. This reduces the cost of data acquisition by up to 100 times—and makes building autonomous systems economically viable.

This means there’s no need to send an autonomous vehicle out into the real world and hope a cat jumps in front of it. Such very specific conditions, along with countless variations of them, are created digitally. Similarly, it’s not necessary to throw pallets at robots in a warehouse, as thousands of variations of that scenario can be performed digitally to generate the data that teaches the machines. 

This tool dramatically reduces development timelines and offers the luxury of testing for hundreds of “what-if” scenarios that are either very hard or very dangerous to replicate in a real-world testing environment. The real difference comes from the modular way that the simulation system has been built; it’s composed of different building blocks, each of which is interchangeable. As with toy building blocks, a huge number of different outcomes can be created from a small number of elements. This allows tailored developments for specific client requirements.

Frustrated by the pace of your automation development? Here’s how to accelerate it.

Image: iStock id1193101728