Extra sensory perception

A range of high-performance sensors under development by Southampton University could detect potential car accidents more accurately than any currently available.

These inertial sensors would measure factors such as the car's yaw rate and communicate with its central computer, which could then perform electronic stability control functions such as reducing cornering speed to prevent the car being driven in a way that would risk it being rolled.

The use of Micro Electro-Mechanical Systems (MEMS) accelerometers to sense rapid deceleration for airbag deployment is over a decade old. However, MEMS inertial sensors are now being used in many other automobile applications such as ABS braking to detect if the chassis is still moving.

The team at Southampton's School of Electronics & Computer Science, led by Dr Michael Kraft, is working with Belgium's Melexis, which produces integrated semiconductor device systems for the automotive market.

Their work aims to develop innovative control and interface systems to improve the performance of existing micromachined sensors.

In the three-year, £200,000 research project, 40 per cent of funding will come from Melexis, which will supply micromachined accelerometers — devices for measuring acceleration — so that the team can assess and improve performance using their interface and control circuits. The remainder of funding will come from the EPSRC.



Meeting specifications

Intelligent automotive systems have created demand for high-performance inertial sensors such as those that sense if a car has crashed and so deploy its airbags. According to Kraft, existing ones may not meet increasing specifications.

The research will take an existing Melexis accelerometer and use advanced electronics and control engineering to make it better, more versatile and easier to integrate at system level.

To achieve this, the team will focus on the electronic interface and control systems aspects of the sensors.

Although their first application will be aimed at improving car safety systems, the sensors could also be used to improve GPS back-up systems, virtual reality systems, inertial navigation and guidance, as well as seismology.

For instance, the sensors could be used to measure whether a shipment of fragile goods had been exposed to shocks, or integrated within digital cameras to alert users if their hand is shaking while they take a picture. Other applications include measuring the vibration of machinery, for example factory equipment, in order to minimise it.

The device will feature a digital output signal, to enable it to interface with a microprocessor. This will allow data from several sensors to be combined, while the computer will perform any complex mathematics using algorithms.

The researchers have already received interest from seismology companies working for the oil and gas industry. During the exploration stage, explosives are used to create mini-earthquakes. The technology would allow engineers to measure this groundshake with extreme accuracy. Examination of these waveforms would then reveal the presence of oil and gas.

The improved systems could also be used for navigation of aircraft, rockets or trains. In the air, they could be used in conjunction with a GPS system to measure the position of the plane, provided that its starting position is known.

Kraft added that the research could also lead to the development of an improved gyroscope.

'Micromachined gyroscopes lag behind accelerometers by about a decade in their performance,' he said. 'They may be able to sense the motion of, for instance, someone shaking a camcorder, but they are very crude.'

MEMS gyroscopes that have so far been built are very big and bulky, as well as expensive. Although used on aircraft, cost prohibits their use in motor vehicles. The gyroscopes could also be integrated within advanced in-car safety systems, improving them further.



Improved performance

Kraft's work will focus on finding ways to compensate for the inherent sources of non-linearity between the input and output signal of sensors, correcting this using electronics.

'We hope to use existing fabricated micromachined sensing elements, improving the performance of the sensor beyond what you would get if you used a micromachined sensor and amplifier,' said Kraft. 'The sensitivity of the sensor is not the only thing of interest. We will also be looking at the bandwidth — the maximum frequency under which we can function.'