The Executive Chairman of Ford, Bill Ford Jr, recently set out a vision of a world mostly free of traffic congestion as intelligent systems monitored traffic flow and directed vehicles accordingly. A world where each parking space is monitored and cars directed to specific empty spaces as they enter a city. For all those sat on gridlocked highways, or driving endlessly round a multi-story car-park this may seem fantasy. But, in principle, it is a dream we could bring about. Monitoring traffic flow is relatively easy, as is deducing where congestion is occurring and working out where to re-route cars. But there is a big piece missing from the puzzle at the moment —a way to get information from sensors to control centre and from control centre back to cars, traffic lights, roadside signage and more.
Transport is not the only area crying out for an M2M connectivity solution. Connectivity for the smart meter is far from solved. Smart cities require myriad sensors to be connected. Healthcare could be revolutionised with sensors in pill dispensers, diabetes monitors, scales, heart rate monitors and so much more. Even devices in the home might benefit from a simple and cheap connection to an external network.
There tend to be two differing reactions when the M2M connectivity problem is discussed. One is to note that there have been predictions of machine connectivity for decades and that they have never really been fulfilled. This leads to scepticism that they ever will. The other is to assume that cellular systems can provide all that is needed.
Responses to both of these require an assessment of what machines actually need. While there are many possible applications, the common characteristics tend to be: very low cost both for the chipset and the annual fee for sending data; ubiquitous coverage, even better than cellular; in some cases battery life of 10 years or so.
This is obviously challenging, but there are some characteristics of M2M traffic that can be exploited in system design including: Most messages are very short; Delays of a few seconds are rarely problematic; data rates can be low; sleep times can be long in some cases and; seamless handover is not required.
Looking at this list it becomes obvious that there is no wireless communication standard currently available that comes close to meeting these requirements. Cellular provides coverage to almost the level needed but cannot achieve the cost points and battery life required. Indeed, if it could it would have done long ago. It is moving ever further away from a design that would exploit the characteristics that would provide such a system as it becomes ever faster and more complex.
Bluetooth and Zigbee do not provide the range and coverage needed. Systems like Paknet do not have the capacity or low device costs. So we do need a new standard. One that exploits the simplifying characteristics of machine communications in delivering the cost, coverage and battery life needed. But can it be done, or are the sceptics right to note that all previous attempts have ended in failure?
Designing such a system requires a little lateral thinking but few breakthroughs. To get ubiquitous coverage from sensors with small batteries lasting 10 years, without deploying a dense network that would be overly costly requires mechanisms for extending the link budget. This can be done with spreading, just as deployed in GPS satellites.
Spreading trades off range against data rate. That is fine as long as there is sufficient initial bandwidth such that the resulting data rates when spreading of up to 1,000-fold is used are high enough to allow for up to 100,000 devices per cell to send their messages. Indeed, data rates are irrelevant in such systems; it matters little to a smart meter whether it sends a reading at 1kbit/s or 1Mbits/s. Instead, “polling times” – the frequency with which each device in a cell can be read, are much more important.
Low frequency spectrum below 1GHz helps improve propagation. Finding, say, 40MHz of bandwidth below 1GHz which is globally harmonised to generate economies of scale and is sufficiently low-cost has been a near-impossible challenge until recently. It has been the advent of the white space spectrum that has changed all of this. White space has the potential to be harmonised around the world, has an average of around 100MHz of bandwidth available in the ideal frequency band and is low-cost or free. The answer to the sceptics is that it is the advent of white space spectrum that makes M2M communications possible.
Possible but not simple. The technology needs to be able to handle short messages, often only 50 bytes long, with minimal overhead (cellular falls down very badly here). It needs to be able to cope with 100,000 simultaneously attached terminals per cell. A range of broadcast calls are needed. Operating in white space demands time division duplex (TDD) as paired spectrum may not be available. It requires frequency hopping to mitigate any interference and antenna nulling to avoid strong interferers such as TV transmitters.
The need to use spreading to reach low-power or distant terminals requires that any header information in frames is also spread, extending its duration. This results in long frame times of the order of two seconds – happily not an issue for most machines. No current standard or technology looks remotely like this. Hence the rationale for Weightless – a standard optimised for M2M communications and able to operate in white space spectrum.
Weightless is being developed as an open standard in much the same manner as Bluetooth was. There is a Weightless standards body called the “Weightless SIG” which has issued version 0.6 of the specification (around 80 per cent complete) and expects to publish the first complete and stable specification—version 1.0—early in 2013. Like Bluetooth it has a royalty-free licensing regime for the terminals in order to foster a vibrant eco-system and keep the terminal cost low. It is likely that the specification will then be ratified by a body such as ETSI to turn it into a formally recognised standard.
My book describes the Weightless technology in a much more readily understandable manner than the specification, along with important additional information describing white space regulation, the business case for network deployment and likely key applications that will use Weightless.
These applications include smart metering, healthcare, automotive telematics and consumer electronics. For smart metering Weightless must be able to deliver chipset costs of a few dollars, meet maximum annual fee charges of less than $10, provide coverage deep within homes, be able to read meters hourly or daily and be able to handle events such as large numbers of simultaneous alarms from meters in an area where the electricity supply has failed.
For automotive, costs also need to be sufficiently low that vehicle manufacturers can build it into vehicle electronics without concern for the through life cost. Typically this will need to be part of the purchase fee of the vehicle, as few consumers will want a monthly or annual bill for their vehicle connectivity. Weightless will need to provide mobility support, efficiently handling signalling loads for cars that may pass through thousands of cells per message transmitted. It needs to allow for downloads of new vehicle software, broadcast to all cars in the region overnight and to be able to transmit reports on vehicle condition and more.
For healthcare long battery life, assured and highly reliable communications and emergency modes with rapid response are needed. For consumer electronics low cost is also key as is simplicity, good coverage, worldwide availability and flexibility to handle a very wide range of applications. These are exactly the sort of markets Weightless is designed to serve. All the design, simulation and measurement work to date suggests it can indeed deliver on the requirements for all these.
So M2M has been tried many times before but Weightless has much going for it. In particular, spectrum that is ideal allowing a design without compromise and the use of novel system design optimised for machine communications.
Professor William Webb is CEO of the Weightless Standard body and CTO of Neul, one of the companies behind the Weightless standard.