Chatty factories

12 December 2017

Interview with

Peter Cowley - Angel Investor

We live in an increasingly interconnected world. There are now smart fridges that order your shopping when certain foods run low, and even intelligent pills that will tell the doctor when you swallowed them. Now the UK’s Engineering and Physical Research Council have funded a £1.5m project to get products, production lines and manufacturers talking to each other. Katie Haylor spoke to Angel investor Peter Cowley, who has been taking a look at what they’re planning...

Peter - A chatty factory is, I think, a term that’s being produced from this grant which is, as you say, a grant being done by academics throughout the UK. Chatty factory in that the product will chat back to the factory, so let’s just explain that: the ability for a product once it’s left the factory to chat back needs some sensors and it needs some connectivity. So the sensors could be strain gauges, they could be temperature, they could be humidity etc., it could be location even, and that data is then passed back from the device, whatever it is, a consumer device or an industrial device, back to the factory. The factory will then use that information to make modifications to the process to improve the product as it moves out of the factory. Now how long that feedback loop is I’m not sure, but it could effectively almost happen simultaneously. So if something happens out in the field, a minor change is made to the processes.

Katie - This is through the internet of things, right?

Peter - Correct. We’ve had our desktop computers connected for what 30 years, 40 years. We’ve had our mobile phones connected for 10 or 15 years. This is where it’s not a mobile phone, it’s a thing, and these things can be very small, they can be almost invisible.

Katie - Can you give us an example of what type of product might be hooked up to this feedback cycle?

Peter - Yeah. The example given in the press release to do with this grant is a bike helmet. If you imagine a pushbike helmet that has sensors on it, those sensors could then detect something has happened. Possibly it was dropped, possibly the temperature was too low or too high and pass that back to the factory so the factory could then change something. It’s unlikely they’ll actually change the design, but they might change the process so it might change the material, it might change the curing temperature, it might change whatever.

Katie - Is this changing something then going to applicable to literally that product, say my bicycle helmet, or is it for future designs of that same helmet that are going to be making their way into the stores?

Peter - Good question. No, it won’t change the helmet you’ve got on your head, but what it will do is change the next ones that come out of the factory. Maybe even the same day or over the next few weeks or months. Can I just point out though that the bike helmet is not a very expensive product - 10s of pounds, and the actual cost of manufacture is a quarter of that so adding sensors to a bike helmet I’m not sure makes that much sense.

Katie - That seems a relatively simple thing to do. This grant is quite large,1.5 million pounds, why do they need that much money?

Peter - It’s being spread between five research departments in five different universities so they obviously have to be working together. It’ll be used primarily not for the sensing, I suspect, but for the processing of the data. The reason that it’s so relevant, and connecting products back to the factory - I had a car in the early 80s that would actually tell through the garage what was wrong with the car. Not connected in real time, but it would tell that. What’s enabled this is interconnected things which are coming on rapidly, and also big data crunching - machine learning, deep learning.

Katie - What’s the motivation for this? Is this we want to improve our products and, ultimately, make more money or is there another motivation?

Peter - There will be a number of motivations but this process has, in fact, being on for decades to some extent. More reliable products so, therefore, the warranty claims will be less. The consumer will feel happier about it because things won’t be breaking. Also the factories would be able to produce less of something that’s not selling very well, possibly because of usage.

The bike helmets a bad example there but I’ve found an example where, for instance, adjusting clutch pedals. Some regions of the world don’t ever adjust their pedals on a car. If that could be fed back to the factory, they wouldn’t  offer that product in the future.

Katie - Oh, I see. Because you know how much it’s being used, and if it’s not being used very much you can just take it out?

Peter - Correct.

Katie - What about this data? Companies have your usage data on products, is there a question around the security of that data?

Peter - There is with any interconnected devices. It’s not just the companies having that and you sort of trust that the company you buy from will look after it, it’s the connectivity back to that factory. It’s between the device at home or whatever and going back to the factory, if that’s intercepted. There's actually part of the grant specifies that. They definitely will be addressing that security of data. The whole world is moving in that direction. We’re all worried about our data leaking.


Add a comment