Ultra Low Power Sensor Networks for Agriculture
Sensor motes today do not seem to achieve the aims of those who coined the term ?smart dust?. Sensor networks have not yet become pervasive, perhaps because of the cost. One of the main aims of this project is to design and implement prototype motes whose cost on mass production would be just a few dollars or less. The motes we are interested in should be:
* Cheap: Both the hardware and the software must be low-cost. For software, this probably means that specific application code will need to be generated using sophisticated and intelligent model driven techniques.
* Very low power. They will possibly operate using light or vibrations as a source of energy, and perhaps just a capacitor or small battery to maintain operation when environmental conditions have not provided enough power.
* A number of integrated sensors: Such as time, light, temperature, voltage, location etc.
* A very low-power processor: Designed for low-power rather than full speed, with the ability to sleep and wake up quickly, perhaps with sensor specific instructions.
* A very low-power wireless transceiver: The transceiver may possibly use novel techniques such as ultra wideband.
* Environmentally aware adaptable programming (model driven software engineering): For example, the ability to dynamically offset communications power usage with power usage for error coding. Another example being dynamically adaptable, cross layer optimisation of the protocol stack and tuning of individual protocols behaviour and syntax (e.g. packet/frame formats) where needed.
* Location aware: The motes may not need to know their exact location but should still be able to provide location information when required.