Smart structures to get lift-off

A US research centre aims to produce a network of sensors that connect the physical world much as the internet links computers.

If the technology is successful water systems will automatically detect sabotage and isolate the danger. Buildings will change their structural properties to prevent collapse during earthquakes by avoiding resonance with seismic energy.

The Centre for Embedded Network Sensing (CENS), at California’s UCLA, aims to develop tiny sensors and actuators that will be densely distributed within a natural or manmade environment.

The sensors will be able to organise themselves into a network, repair themselves and manage their own power consumption. They will monitor and collect information on subjects from soil contaminants to physiological information for doctors about their patients.

‘Not only will we be able to collect information not available before, but this will allow us to design systems to automatically take action once a pollutant, structural failure or other hazard is detected,’ said CENS director Prof Deborah Estrin.

The UCLA project, which is to receive up to $40m (£27m) from the National Science Foundation over the next 10 years, aims to incorporate wireless channel communication.

So far the promise of smart structures has failed to materialise as they have depended on a central computer to receive, process and send information and commands to each sensor.

Once they grew beyond 100 nodes, the networks became too complex for a centralised computer system to handle. In addition, the weight, power con-sumption and cost became prohibitive, meaning that it was almost impossible to increase network size.

But recent advances in micro-electromechanical systems (MEMS) and distributed computing are starting to overcome these limitations.

A first test for the new technology will be to tackle the high noise and vibration levels inside rockets during launch. Vibration significantly increases the cost of manufacturing satellites and other space equipment, as these structures must be strengthened to withstand it.

But a system that reduced vibration by even a small amount would allow designers substantially lower payload development costs.

Researchers at Tennessee’s Vanderbilt University, led by assistant professor of mechanical engineering Kenneth Frampton, have designed a smart vibration- reduction system for a 4.5m rocket payload fairing using embedded systems.

In their embedded system each node contains a PC-strength microprocessor with a relatively simple program and a modest amount of memory, allowing it to control the sensors and actuators wired to its node.

The microprocessor is also able to communicate with its neighbours to work together, and as its workload does not increase as the system grows, larger networks can be built.

Frampton’s team now plans to place a 100-node system into a rocket fairing to test its performance.

On the web