Giving a flexible edge to the IoT

As the Internet of Things (IoT) continues to revolutionise our daily lives, the demand for smaller, smarter, and more diverse flexible technology has never been greater. Increasingly complex demands have driven the development of smart sensors to monitor everything from velocity and proximity to pressure, humidity, and more. Future devices will need to interact with the ambient environment by performing intelligent activities such as fingerprint, vein, and odour recognition, with sensors so small and flexible that they can be integrated into almost anything. 

Such technology could enable clothing that automatically deodorizes you, food packaging that alerts you to contents past their best, or bandages that can detect early-stage infections, but these activities require embedded and customized processing hardware built on a flexible substrate such as plastics. A new project could make these devices a reality sooner than you think.

PlasticArmPit

The ‘PlasticArmPit’ project is a 30-month project that started in October 2017. It has been part-funded by Innovate UK, and is led by British tech company Arm, based in Cambridge. It brings together a number of key partners with complementary expertise; Cambridge-based PragmatIC, a world leader in ultra-low cost flexible electronics, The University of Manchester, which has an extensive background in printed Organic Field-effect Transistors (OFET) sensor technology, and global consumer goods company Unilever, which brings the opportunity for a real-life application from the consumer industry. 

The consortium aims to create a wearable device consisting of an ‘e-nose’ sensor, sensor interface and custom machine learning (ML) processing engine, manufactured on a flexible substrate like plastic, as a proof-of-concept, before extending this potentially revolutionary technology to a wide range of alternative applications. The initial use case is for a sensor which can detect body odour and inform the wearer when they need to apply more deodorant. An enhanced system could even apply a top-up directly, removing the need for human monitoring and intervention.

Looking beyond silicon

Unsurprisingly, the device created needs to be low cost, to have small form factor, and be specifically designed bespoke to the application, so that it can be seamlessly embedded into any garment, container, or wrapping. This means looking beyond traditional chip manufacturing methods and materials to achieve the project goals – using plastic in place of silicon, and specialist flexible electronics fabrication processes that allow for low-cost hardware customization, which is very expensive to achieve in silicon fabrication process.

Over the last 50 years, Moore’s Law has delivered significant benefits which silicon designers have been able to exploit with chips made up of many billions of transistors. However, as flexible electronics technology is much newer, the largest published designs to date are just thousands of transistors. For silicon designers, this is like time travelling back to the 1970s! As a result, flexible integrated circuit designs need to be very frugal in terms of transistor usage, without spending more in design costs and efforts. For this reason, unlike most silicon chips, the device won’t be reprogrammable. A common base design will be configured at chip level for each specific purpose. 

How can you make ML work on plastic? 

The consortium believes that customized processing engines such as neural networks (NNs) are the key to accelerating development of low-cost and customized flexible, integrated smart systems. Customized for a specific application and capable of operating in extremely parallel fashion to achieve high performance, and consume low power, this will be the first time that a flexible smart device has been created to take advantage of machine learning algorithms in hardware.

Emre Ozer, the project coordinator from Arm Research, describes the project:

"PlasticArmPit aims to develop the underpinning technologies such as OFET-based e-nose sensors and a custom machine learning processing engine on flexible substrates, which can be embedded into the packaging of items. Without Innovate UK’s support, these four unique partners would not have been brought together to collaborate on such a disruptive industrial research project."
Emre Ozer, Arm Research

Not just a vanity project

While not many would object to embedded technology that keeps you smelling good, the development of such innovative technology could pave the way for other valuable applications. 

Corporations and consumers around the world are becoming increasingly more conscious of their environmental impact. Food waste is of particular concern, with over 1.3 billion tons of food wasted in 2016 (Source: The Grocer), including 45% of all fruit and vegetables and 33% of all seafood produced globally. In the UK and elsewhere, moves are being made to remove confusing best-before dates from packaging, which are thought to contribute to around one third of annual UK food waste. 

Intelligent food packaging using the e-nose coupled with ML processing engine technology could result in significant additional reductions in food wastage. Particular application might be found with items such as fresh meat or fish, which become potentially dangerous rather than just unpleasant to consume when past their best. A sensor that reliably detects the presence or concentration of a particular odour or chemical indicating toxicity in the protective packaging atmosphere, displayed in an easy-to-understand format, would give the consumer confidence to consume products which may have been otherwise wasted when relying on traditional ‘use-by’ dates. Wide adoption of flexible IC technology in packaging could even pave the way for smart recycling centres, enabling automatic sorting of various plastics and reducing contamination.

Applying the technology in healthcare could enhance smart bandages, which can not only promote faster healing but could also provide crucial early indicators of infection. This would be particularly useful where resources are scarce and need to be distributed according to need. In an additional safety application, they could also be usefully applied in mechanisms to detect air quality or pollution, or develop more specialised air purification systems. 

What’s next?

The ‘PlasticArmPit’ project is the first to pioneer ML hardware as the processing engine of choice of the flexible electronics world. It will serve as an important proof-of-concept for reliable and efficient building of smart, integrated, fully flexible systems. The work completed by the consortium could pave the way for significant advances in industries spanning food, healthcare, and agriculture.

PlasticArmPit Partners

“Like silicon electronics, I believe that flexible and plastic electronics can change the world. The cost and form factor advantages allow it to address different problems to silicon; in particular we see opportunities to improve health, food safety, and sustainability. The PlasticArmPit project is generating the fundamental building blocks required to support our vision for these affordable applications, and putting the design flows in place to make this technology available and accessible to everyone.”
James Myers, Director of Devices and Circuits Research, Arm

“The PlasticArmPit project is part of our vision to add increasingly complex sensing and processing into everyday objects. Our unique, ultra-low cost, flexible ICs can deliver electronic intelligence at a dramatically lower cost point than silicon, while also enabling physical device architectures that are more naturally matched to the neural processing structures of the human brain.”
Scott White, CEO, PragmatIC 

“At The University of Manchester we have been developing a low cost sensor platform through support from the EPSRC Centre for Innovative Manufacturing in Large Area Electronics.  These sensors can detect volatile compounds in a range of environments at very low concentrations.  We are delighted to be working with such world leading companies as Arm, PragmatIC and Unilever in the PlasticArmPit project to realise an integrated system that can mimic human perception.”
Prof. Michael Turner and Prof. Krishna Persaud, University of Manchester 

“Participation in this Innovate UK funded project has provided Unilever with a unique opportunity to collaborate with leaders in the fields of OFET-based sensors, custom machine learning and flexible electronics.  The aim of the project is to develop a bespoke and low cost flexible electronic solution to tackle a complex, real industrial challenge. The proposed device will be designed with sustainability high on the agenda, and the customised highly innovative solution will provide a step change in measurement capability for our applications.”
Susan Bates, Research Manager, Unilever Research and Development