When the Internet of Things and Smart Machines Collide


Slowly at first and then suddenly — that’s the signature of exponential growth. Out at sea, a tsunami is an almost imperceptible swell. But when it rises up and hits you, unlike a normal wave it just keeps on coming.


We are all constantly bombarded by hype on technology trends but there are two genuine technology tsunamis heading our way right now, namely the Internet of Things and Smart Machines. Where these two forces collide with one another and with us, they will create an explosion of new opportunities in areas as diverse as entertainment, healthcare, disaster management and smart cities. If you thought the mobile revolution has had a huge impact on individuals and businesses, you ain’t seen nothing yet!


At a recent Ocado conference, we gave all the delegates a small electronic gizmo and asked them to put it behind one of their ears. We then demonstrated that we could locate them and remotely sense information such as their heart rate, temperature, stress and arousal levels. The demonstration was in fact a spoof, not because anything we did was impossible just that doing it for real would have been challenging in terms of the size of the device and its cost; but all that is changing.


There are already relatively low cost consumer EEG brain sensing devices available that can feedback your stress level and its only a matter of time before it becomes practical to do the sort of sentiment analysis we pretended to demonstrate. Indoor location is ironically harder to do than outdoor location, especially in environments such as our highly automated warehouses (we call them Customer Fulfilment Centres or CFCs), but new technologies are becoming available and their cost is falling.


The last ingredient was two way communication to a dense population of thousands of sensors. Imagine a restaurant where the lighting, music, decor, temperature, humidity and even smells are controlled by feedback from wearables attached to customers. Or a nightclub where the music and lighting are controlled by the movement and bio-feedback of the dancers. However what if you wanted to do the same for a large rock concert, or the real time control a herd of robots on an automated construction site or choreographing a swarm of nanobots cleaning up an oil spill.


Those are genuine Internet of Things challenges where conventional communication solutions won’t cope with the density of devices. It’s actually a problem Ocado has just had to solve for our next generation of CFCs and the solution we have created has far wider applications.


So what about the second tsunami, smart machines. For a system to be considered a genuine smart machine, I believe it needs to do something that previously you would have thought only a human could do and it must display a high level of autonomy. Products such as Google Now or Microsoft’s Cortana are smart(ish) machines that give us a sense of what truly smart machines will be able to do for us in areas such as proactive search, cognitive analytics, digital assistants and smart agents.


We already employ a lot of machine learning, optimization and data science techniques to make our end to end solution very smart but like the holes in the proverbial Swiss cheese, we can see so many opportunities where advances in areas such as deep learning will enable us to make it much much smarter. These are the sort of technologies we are building into our new Ocado Smart Platform, to put grocery retailers around the world online using our disruptive business model.


But the really exciting opportunities come where these two tsunamis collide. Then we will have smart machines talking to all these internet enabled devices and in so doing, becoming much more aware of the world around them. And smart machines will be talking to other smart machines, creating a lattice of smart machines and devices all talking to one another.


Let’s imagine you are driving to work and your smart car develops a fault. The smart engine management system communicates with the car manufacturer and additional diagnostics are downloaded and run to identify the cause of the fault. A repair is required and your car communicates with your garage to find a suitable service slot tomorrow. The necessary parts are ordered from the manufacturer. Your car is aware that its MOT and annual service are due soon, so it arranges for these in to happen at the same time.


Your car contacts your digital assistant at work. You have an early meeting this morning that needs to be cancelled as you will now be in late. The digital assistants of the other meeting attendees are also contacted to handle the consequences of the meeting cancellation. You had a one day business trip arranged for tomorrow and your digital assistant knows you were planning to drive to and from the airport. Since your car will be unavailable, it asks if you would like to go by train or taxi. Having opted for the train, your smart luggage is re-programmed accordingly.


You were due to collect the kids on your way back from the airport which now you won’t be able to do and so your partner’s digital assistant is contacted to arrange for them to do the school pickup instead. You have an Ocado order booked for tomorrow afternoon but since now neither of you will be around, your digital assistant asks if you would like it moved until later or converted to a click and collect delivery.


A temporary digital key is downloaded to your car to allow the garage to collect it from the roadside. Your home management system is alerted to allow the garage to open your gate and drop off your car in the afternoon. Your digital assistant knows you like to watch Suits tomorrow evening but since you will be back late, it records the episode for you. As nobody will be around to give the dog her second walk, your digital assistant asks if you would like the dog walker contacted. Finally it’s your partner’s birthday tomorrow, so your digital assistant reminds you to buy a present and asks if you would like dinner booked for the two of you; and so on. The initial car fault is like dropping a stone into a pond, with the implications rippling out and being handled by a lattice of smart machines and devices.


So who will build out this smart internet of things? The rise of mobile app stores and cloud services has democratized the art of the software start-up, making it genuinely possible to setup a multi-million (or even billion) dollar worldwide business from your bedroom. In a similar way, the explosion of low cost hardware platforms for embedding computing have democratized building the internet of things, enabling SMEs, start-ups and even makers to be players in this revolution. This will be an innovation battle fought with asymmetric warfare.


So what are the implications of all this technology coming our way? Well one of our holy grails is using technologies such as deep learning to process all the data we collect to predict what our customers want before even they have clue. And then we need to integrate with our customers’ increasingly smart worlds. Here smart packaging is really exciting. Imagine a portal in your home that warns you when products enter that might trigger your child’s peanut allergy. And forget fizzy drink bottles with your name on and think about smart labels that interact with you.


And smart packaging will interact with smart utensils. And smart packaging and smart utensils will interact with smart appliances and robotic cooking assistants. Devices that require supplies, such as your coffee machine or your dishwasher, will be going online and taking responsibility for replenishing themselves. And all of this is part of the fabric of the truly smart home. And smart homes, smart offices and smart communities will be part of smart cities and smart countries.


Many of us already inhabit a hyper-connected always-on world with mobiles next to our beds but wearables are going to take this to a whole new level and with it, data overwhelm. To the rescue will come another breed of smart assistants, the Smart Agents. These will help you manage your data privacy and when you finally get tired of splurging all your personal data over social media, these agents will negotiate what data you are prepared to share with whom. And since these data have value, they will also administer the auction to get the best deal for selling them.


So what are the next steps? Well if you are up for building this stuff, then there are some capabilities to go shopping for such as data science, natural language processing, machine learning and deep neural nets. This is a revolution where fortune is going to favor the early adopters and those willing to get their feet wet experimenting before these technologies mature.


Returning to the theme of tsunamis, the water is drawing back now, the tsunamis are on the horizon and the time to act, is now.


Paul Clarke is Director of Technology at the UK online grocery giant Ocado.



No comments:

Post a Comment