Redefining the notion of sensors

26 March 2014

Today’s vehicles use sensors to control emissions, correct driving behaviours, receive satellite radio signals and provide entertainment to passengers.

National Instruments (NI) draws on its experience of working with over 35,000 customers, spanning automotive to telecommunications; aerospace to medical, to predict which emerging test and measurement technologies will be at the forefront in 2014 and beyond. By Kevin Ross, National Instruments, UK & Ireland.

Annually, the company publishes its Automated Test Outlook, which highlights the top trends that it believes will significantly influence automated test in the next few years. This month, we will focus on redefining the notion of sensors, as technologies incorporate more sensors, and more unusual sensors, making them increasingly challenging to test.

Sensor manufacture

Janusz Bryzek, a Fairchild Semiconductor executive, says that sensors could reach a manufacture rate of one trillion per year, from today’s 10 million, within the next 10 years. With that sort of growth, MEMS (microelectromechanical systems) could become 30% of the total global semiconductor market, worth in excess of $100billion per year compared to a value of about $11billion in 2012.

This eruption of sensors poses clear challenges on test departments and their engineers. Many industries have been forced to change their fundamental understanding of what a sensor is. The notion that sensors only measure temperature, strain, force and other basic data points is obsolete.

Greater expectations

Greater expectations create greater challenges. Some of the clearest examples are in high-volume, consumer-facing markets like automotive and telecommunications. In this space, consumers, suppliers and even legislative bodies have high product expectations. Consider the automobile: sensors were used to monitor key data points like engine temperature and oil pressure, but because of the rise in consumer and legislative demands, manufacturers have been forced to significantly increase their cars’ electronic components and capabilities. Today, vehicles are required to control emissions, correct dangerous human driving behaviours, receive satellite radio signals and provide a level of entertainment and convenience to passengers. Engineers must expand the idea of a “sensor“ to technologies like O2 sensors for catalytic converter output, cameras for monitoring the driver’s eyes, an antenna for picking up digital radio and navigation signals, and a display for video and information communication, just to list a few. 

This only scratches the surface of these new types of possible sensor, a view supported by Tom Pierce, vice president and general manager for Test and Measurement at Honeywell Sensing and Control.

“In the future, people are going to put sensors in places we never have thought about,” stated Pierce. “The need for sensors is exploding and there are many more potential sensor applications than we could have ever predicted.”

Exceptional applications

The robotics industry is a lucrative one, where a multitude of sensors is required to accurately measure the surroundings and sense its environment. Pierce’s words are echoed in an extraordinary and innovative application of sensors in the UK, a robot controlled by the humble fruit fly.

An inter-disciplinary collaboration of engineers from Imperial College London, Optotune AG, ETH Zürich, ViSSee Sagl and Tufts University, developed a flexible robotic device to measure and simulate flight patterns in winged insects. 

All animals, including the common fruit fly, employ a profusion of rapid and precise biological “sensors, controllers and actuators”; desirable features for any control system. A coal miner carrying a canary to warn of gas is a simple example of a biological sensor. In the robotics case, these biological components are replaced by electronic and mechanical sub-systems.

Using the NI PXI vector signal transceiver and the NI WLAN Measurement Suite, Qualcomm Arteris improved test speeds by more than 200 times.

“Feedback from three linear cameras and eight proximity sensors determine the visual stimulus shown to the fly,” explained Optotune’s Chauncey Graetzel. “Flight parameters such as the wing beat frequency and amplitude control the robot’s movements.”

Whilst biological sensors may be some way from commercialisation, the mobile phone industry quickly exhibits a more timely effect of this explosion of sensors. By 2015, the average consumer mobile phone is projected to have nearly 20 MEMS sensors, compared to two sensors in 2000. For example, the Samsung Galaxy S4 combines a number of sensors: an accelerometer and gyro; geomagnetic sensor; temperature and humidity sensor; Hall Effect sensor; and RGB light sensor.

Implications for test

Capital costs of automated testers can account for more than 60% of overall test costs, so minimising hardware changes can significantly reduce overall expenditure. For example, a dedicated test solution that addresses mobile devices, which have a typical shelf life of 18 months, has obsolete sensors and technology for every new design. Architecting a test system that can adapt to changes occurring once or twice a year requires an agile or proactive test approach. Unlike an ad hoc approach, with dedicated box instruments that specialise in one specific measurement, a proactive test approach features modular hardware and anticipates technology changes. A modular approach minimises the sustaining costs of a tester with incremental changes instead of whole product replacement.

NI advocates this ethos of a modular approach: the NI PXI platform lends itself to such applications, offering a modular embedded controller and a variety of modular instruments, from single-point DC to RF and microwave measurements. This approach contributed to a success for device characterisation at wireless technology company, Qualcomm Atheros.

Engineers faced serious challenges when adapting high-throughput wireless technologies to meet the demands of new connected applications. As wireless standards become more complex, the number of operational modes for the MIMO transceiver devices increases exponentially. In order to utilise the 802.11ac WiFi standard, the devices require new modulation schemes, more channels, more bandwidth settings and additional spatial streams. Additionally, characterising WLAN transceivers is especially challenging when faced with thousands of independent operational gain settings.

Modular approach

The team swapped traditional bench top instruments, which took only 40 meaningful data points per gain setting, for a modular PXI system which increased throughput 10 times. To increase performance to the level required for 802.11ac, Qualcom upgraded its instrumentation. By using the existing modular PXI platform, and swapping out RF instruments for the latest Vector Signal Transceiver module, it was able to achieve 300,000 meaningful data points per gain setting.

 “Using the software-designed NI PXI vector signal transceiver and the NI WLAN Measurement Suite, we improved test speeds by more than 200 times, compared to traditional rack-and-stack instruments, while significantly improving test coverage,” said Doug Johnson, Qualcomm Atheros.

Complexity challenges

If the vision of a trillion sensors per year by 2024 is true, product complexity will be growing and advancing at a significantly more rapid pace than the current trajectory. This trend will continue to impact test organisations as more frequent product redesigns dramatically affect the total cost of test. Companies that use a test strategy featuring a modular approach that can accommodate the changing sensor market will reduce total cost of ownership and improve redesign time to meet more stringent time-to-market demands.


Contact Details and Archive...

Print this page | E-mail this page