Page 8

EETE OCT 2014

Vast numbers, vague predictions at Trillion Sensor Summit By Christoph Hammerschmidt Ever heard of a Brontobyte? No? This is the amount of data generated by the trillions of sensors the Internet of Things will comprise, at least according to the group of experts trying to fathom out the shape and nature of this scenario where everything is connected to everything else. One assumption of this group seems logical: If sensors are transforming information about the analog world to its digital image, and more and more things are connected, then the necessity for a growing number of sensors is obvious. Likewise it is clear that these sensors will emit impressive amounts of data - with their exact amount up in the air so far, hence the somewhat hazy expression brontobytes which can stand for just about anything between ten to the power of 15 and ten to the power of 27. At their recent meeting in Munich, the Trillion Sensor Summit experts exchanged their projections and predictions, and though for the unsuspected observer much of these predictions might appear fantastic, these projections have some real-world foundations. The reason why the IoT now is at the verge of the rapid expansion we are currently observing is that electronic components now have shrunk wide enough in terms of size, price and power consumption to create autarkic, affordable systems. Plus, the data communications technologies, wire-bound or wireless, needed for such sensor networks are available, explained Christoph Kutter, Local Chair of the TSensors Summit in Munich. For the unbiased observer, the sheer numbers of connected sensor nodes predicted by IoT evangelists like TSensor Summit Chair Janusz Brycek seem rather optimistic - he projected that within one generation, some 45 trillion sensors will be online. According Fig. 1: Sensor production will continue to rise at staggering speed, predicts TSensor Summit Chairman Janusz Bryzek. to the presentations at the summit in Munich, hardly any aspect of human live will be spared of being sensed, measured and networked in the future - all for the best of mankind, of course. High volume sensor applications will swamp the markets for mobile and wearables, health and wellness, the home (which by then will be a smart one), cities (also smart), cars and so on. Connected sensors are the answer to many global challenges like hunger, environmental pollution, rising costs for healthcare, claimed Ira Feldman, owner of California-based consultancy Feldman Engineering. Some presenters indeed provided hints how the abovementioned global challenges could be met at the level of connected sensors. For instance, in the area of health care, as Luc Van den hove suggested, CEO of Belgium’s highly regarded research institute imec. “Today’s drugs are generic - they are not optimised for the individual case”, Van den hove said. He sketched an image of smart sensors for blood analysis, heartbeat and blood pressure monitoring, glucose determination and many other purposes. Smartphones will play the pivotal role for these sensors - a smartphone app will read out the sensor signals, process the data and make a diagnosis. “Currently there are sensors under development that can conduct a full blood analysis within 10 to 15 minutes”, Van der hove said. These sensors would be designed and produced for one-time use; their low price would make them affordable for many parts of the world’s population who today have no access to medical services at all. Other sensors will be enablers to build smart cities, with waste management being a big driver for this part of the market, explained Jérémie Bauchaud, analyst for MEMS and sensors at market research company IHS. In the industrial arena, the demand for sensors will be driven by applications for pipeline and tank monitoring. The automotive market offers huge opportunities, too: the trend towards autonomous driving multiplies the number of sensors used in cars. Advanced Driver Assistant Systems and eCall, the European automatic emergency call system, will be the first sensor systems to connect directly to some form of cloud services. Currently, the industry is exploring ways to create additional value out of the data generated by these sensors. If myriads of sensors are collecting data, in many cases the data may involve user’s privacy or the public or personal security. Already today, cyber security is an issue; for the realisation of the abovementioned scenarios, security will be paramount. “All these connected sensors can be hacked”, warned security expert Raj Samani, CTO of McAfee. With medical devices being potential targets for hackers, the lack of security would be the “ultimate disabler” for the trend towards putting sensors online, he said. Researchers and security companies are already developing methods to lock out uninvited guests - and here the TSensor summit finally came up with more substantial ideas. Georg Sigl, Director of the Fraunhofer Institute for Applied and Integrated Security, suggested utilising an unavoidable and basically unwanted property of existing semiconductors for security purposes: tiny variations of each individual chip, caused by minute variations in the manufacturing process, result in a unique pattern for each chip, comparable to a fingerprint. Known as Physical Unclonable Functions (PUFs), these variations cannot be copied, Sigl emphasized. They however can be used to generate encryption keys - keys that are hopefully secure enough to protect the myriads of data generated by the billions of sensors in the Internet of Things. 8 Electronic Engineering Times Europe October 2014 www.electronics-eetimes.com


EETE OCT 2014
To see the actual publication please follow the link above