The shopping list: national Instruments has set out six grand trends and issues for 2014, for itself as well as for the broader test and measurement sector.
It's making no claims to originality or completeness in the six it has picked, but what it's hoping to do is to spark discussion and debate about the broader implications of changes in technology not just in the areas that it covers as a group, but more widely. And though it's been written up as six separate 1000-word essays in a downloadable PDF, there's some overlap between them. Here's a brief run-through of the six
1. The cyber-physical design challenge
Being American by birth, NI tends to refer to the combination of embedded software inside hardware products as "cyber-physical systems" or CPSs, though in other parts of the world terms such as "Internet of Things" or "Industrie 4.0" are used. The gist of a CPS is that they are products and systems that have a degree of inbuilt "intelligence" or self-awareness and are able to some extent to communicate their status and measure and control their performance.
That CPSs are going to be everywhere is pretty much a given in most forward projections about technology But what NI's Trends people see as less known is that this kind of technology demands a rather different design methodology from the practices that have gone before to make products.
Some of this links into a later point NI wants to make about skills. But it's also that there is a significant difference between something that's been designed as a CPS and something that's just evolved, through complexity, into a system that has both hardware and software features, such as the electricity supply network. Lack of the CPS-style design methodology, NI suggests, is why the electricity grid in the north-eastern US suffered a massive and embarrassing time-day black-out in 2003: no shortage of software-based models in the grid, but they failed to model the behaviour of the control systems that shut it down without warning.
NI says that there are essentially two methodologies for designing a CPS that have been proven to work, and in some cases you can take your pick of which of the two you use, while in others it'll be pretty obvious that one is appropriate and the other isn't.
Model-based uses modelling and simulation to design, analyse, verify and validate dynamic systems, with the models drawn from the specifications or requirements and analysis of the environment. Models enable fast development of systems by doing a lot of the design work in the abstract but, says NI's UK technical marketing manager Kyle Voosen, "you have an element of risk" - when you move from the abstract to the concrete and try the system out, it just might not work.
A second design method for CPS then is pretty much a diametric opposite: the platform-based approach is essentially an incremental design process in which you bring together proven elements step by step, so that at each point you know that what you have is going to work. Maybe it's slower, but it's more certain.
There are tunes that can be played between the two approaches. Patently a car manufacturer, with an installed base of already complex systems that it's been making before, is more likely to adopt a platform-based design approach. But where you're adding embedded software systems and smart functions to a previously dumb product, however, either approach can work.
2. Big Analog Data
NI has been flagging this as a trend and an issue for some time, and in fact has trademarked the term with Capital Letters. The gist of the argument is that, amid all the proliferation of data of all kinds from many sources, the analog data from instruments measuring the physical world - the environment, nature, people, machines--is going to be the biggest of all, partly because in many cases it's continuous measurement.
This huge volume has to be managed in some way, and Ni has developed a three-stage architecture for handling it, analysing it, extracting the useful stuff and then storing it.
3. The SDRification of RF instrumentation
Software-defined radio or SDR transfers many of the physics functions of wireless systems to software, and is now a key technology in RF (radio frequency) instrumentation, broadening its use in areas such as FPGA-based instruments.
4. Common languages for system design
A recurring theme across NI's six trends is that CPS and other modern devices and products bring together a lot of technologies and talents that were formerly seen as separate. It gives the example of a "smart" appliance which includes expertise on RF standards, power management, physical design, heat dissipation, image capture and analysis and perhaps video--and maybe others.
These diverse areas will tend to have their own "domain experts", their own design, development and test methods, and their own computation systems - and experts are prone to create silos in which they hone and perfect their expertise and are reluctant to modify their behaviours or processes.
With the new kind of products and systems, NI takes the view that this kind of silo mentality is a barrier to progress and believes that the tools for system development such as computation need to be much more closely integrated, with the aim of getting the job done.
5. The touchscreen user interface
Kyle Voosen, of NI's UK office, quotes a provocative American futurologist as likening the computer scientists of today to the typists of 30, 40, 50 years ago: expert in what they do, but ultimately proficient in something that adds nothing to the real business of designing, developing, testing and making things work or making things happen. In the same way that we're all typists (more or less proficiently) now, we'll all be computer scientists in the future.
And the device that will take us there is the touchsereen of the smartphone and the
tablet, which presents programs and information to our fingertips, enables access to multiple systems and increasingly doesn't require us to be in a particular place at a particular time.
Voosen -says that already there are signs that "Generation Y" people will not accept the keyboard or other analog methods for access and control. "Particularly, you do not need to have specialised devices," he says. "If all the apps are based on the single platform, then you need only the one device. We're moving beyond the button and the knob and even the computer mouse."
6. Changing technology education
In the same way that systems and products and devices are changing, so too are the people who develop them. Traditional engineering disciplines of mechanical and electrical engineering have been blurring in any case, since functions that were previously delivered purely in one or other engineering format often now cross over into another. The universality of CPSs - smart systems - pushes this further and in two directions, NI suggests.
On one side, there's a case for saying that the skills that today's mechanical engineers, for example, should be being taught have to include circuit design, control theory and measurement and instrumentation, because these are the parts of the mechanical engineering design of, say, a vehicle transmission where there is likely to be real innovation - not in the transmission itself, which is pretty much known technology.
But the other side of this, Voosen explains, is to question the whole nature of an education system that encourages engineering specialisation without having a view as to whether those specialisms are for the long term. That's not to decry expertise, he says, which will always be wanted: but if the future of many products and systems is putting together assemblies of "black boxes", the task is to have people who have a broad engineering education that enables them to understand what the black box does and what it contributes to the whole, not necessarily in detail what is going on inside any individual black box. Someone has to know that, but not everyone.
In short, what systems need is systems engineers able to take a broader view who know what they want to achieve but don't need the fine detail. *
|Printer friendly Cite/link Email Feedback|
|Date:||Apr 1, 2014|
|Previous Article:||Joint ventures: Hip replacements are pretty commonplace operations these days. Yet there are still a lot of unknowns.|
|Next Article:||Free thinkers: the Millbrook vehicle testing centre has a new owner and a new boss and is looking forward to independence.|