Can you imagine a world where millions of interconnected wireless sensors provide early warning of natural disasters?
Well, imagination might soon not be necessary.
Our perception depends on a complex network formed by the brain neurons and their tentacle-like bodies, which extend up to a metre in length. A wireless sensor network (WSN), on the other hand, can reach up to hundreds of kilometres.
Inspired by battlefield surveillance, WSNs consist of autonomous sensors spread across large areas. These sensors monitor conditions such as temperature, sound, vibration, pressure, motion or pollutants.
A typical WSN may comprise a few hundred to several thousand sensors that register changes to physical stimuli. These sensor readings are processed by a tiny on-board computer, known as a node, which wirelessly communicate the results to a central computer.
There are significant ramifications in spatially-extending the sensing capabilities of WSNs – not least that the increased range will facilitate greater prediction of potentially-disastrous events.
Observing ecological events unfold becomes possible through long-term environmental monitoring of ecosystems with WSN.
By monitoring ecosystems and the environment in the long-term, WSNs can become an early-warning technology. Being forewarned of a cataclysmic event such as an earthquake by just a few minutes can make a huge difference. Used retrospectively, WSNs can help to track events from their point of origin.
So what else is possible as WSNs improve? Well, plenty.
Part of the fabric
Buildings with embedded sensors have long been envisaged, with rooms that cater to specific needs and produce remarkable energy savings.
Similarly, intelligent infrastructure such as in roads can increase driver safety, conserve fuel, and reduce traffic congestion.
Solutions to impending healthcare problems may also be found through the use of WSNs including being able to monitor: the elderly at home; air quality; and quarantined areas.
At present there are two main issues with WSNs.
1) The energy needs of individual sensors have to be met, which means these sensor nodes must conserve energy if very large numbers are to be deployed.
2) New types of computer algorithms are needed to combine these tiny nodes into a powerful network computer to process data and disseminate results in real-time. Such computer programs must be able to keep up with growth in network size.
Optimising the future
Industry estimates suggest there will be billions of internet-connected devices within the next decade.
To meet the challenges of this new world, researchers at Monash University – myself included – have been working on a new types of computer algorithm that make efficient use of large networks and are able to almost instantly recognise crucial patterns.
One such technique is the Hierarchical Graph Neuron (HGN) which imparts brain-like memory to the WSN, enabling the network to rapidly recognise patterns within sensory data.
Programs such as the HGN operate within the body of the network and thus continue to become more powerful as the network grows in size.
Other researcher at Luleå tekniska universitet, Sweden, has lead to the development of Mulle, an energy efficient WSN node.
The Luleå group also specialises in bio-inspired computing and investigates biomimetics – technology inspired by nature.
No flies on me
When it comes to energy efficiency, the natural word can teach us a lot. A fruit fly’s brain consumes only a few microwatts of power and yet is still able to integrate sensory information, actions of flight and control over relatively complex behaviour in order to survive.
Conventional computers are no way near as efficient. An average computer sensor node consumes about one milliwatt of power, which is around a thousand times more than that used by a fruit fly (and five times more powerful than the average laser pointer).
So what would happen if we were to combine an infinitely-scalable sensory network with energy efficiencies of a bio-inspired scheme?
We don’t yet know. But it’s apparent that such a technology will profoundly alter human experience.
Just as the internet has made it easy to communicate efficiently across vast distances, sensor networks will fundamentally change the way we interact with our surroundings.
Sensing natural events in the future, detecting an imminent structural failure, or routine actions such as avoiding traffic congestion or a hazardous driving condition will become the norm.
Preliminary results are very encouraging.