Smart Dust

As described in [19], Smart Dust nodes as envisioned by the Berkeley Smart Dust project will consist of a small battery, a solar cell, a power capacitor, sensors, a processing unit, an optical receiver, and a corner-cube retroreflector (CCR) within a space of about one cubic millimeter. Later versions might also contain an active transmitter based on a semiconductor laser diode. However, the high power consumption of the laser diode significantly limits the usefulness of such a component. Therefore, in the near future, communication will be possible only between sensor nodes and a so-called base station transceiver (BST).

The BST mainly consists of a steerable laser and a compact imaging receiver. For downlink communication, the BST points the modulated laser beam at the optical receiver of a node. For uplink communication, the BST points an unmodulated laser beam at the node, which modulates the laser beam and reflects it back to the BST using its CCR. Using its imaging receiver, the BST can receive and decode transmissions from dust nodes.

Obviously, this communication scheme requires an uninterrupted line-of-sight path. For many of the environmental monitoring applications envisioned for Smart Dust, however, this is not a major problem. Additionally, communication is only possible if the node's optical receiver and CCR point to the BST, so that only a fraction of deployed nodes will be able to communicate. This should not be a problem, however, if the dust node density is high enough.

Having these Smart Dust characteristics in mind, what are the differences to state-of-the-art RF-based WSN with respect to location sensing? The main differences clearly stem from the tremendous size reduction from several cubic centimeters to a few cubic millimeters. The small size also imposes tight limits on the available energy, which in turn restricts communication, memory, and processing capabilities of dust nodes. Another difference is caused by the passive optical communication scheme of dust nodes, making near future Smart Dust systems essentially single-hop networks without direct node-to-node communication. We can summarize the differences between current WSN and future Smart Dust systems as follows with respect to location systems:

• Small size: current RF antennas for radio waves and transducers for ultrasound are too large for dust nodes.
• Mobility: future Smart Dust nodes are likely to be small enough to be moved by winds or even to remain suspended in air, buoyed by air currents. This is in contrast to current sensor nodes, which are typically immobile due to their size and weight.
• Large scale: the expected small size and low cost of future Smart Dust nodes will allow very large scale deployments in terms of the number of nodes.
• Limited energy: the power consumption of current RF transceivers, for example, is too high for dust nodes.
• Limited computing and memory resources: many wideband ultrasound location systems, for example, sample at more than 40kHz and do signal processing on the sampled data, resulting in large memory and processing overheads [13], which is not possible on a Smart Dust node.
• Single-hop network topology: current WSN location systems often assume multi-hop networks, where a node can cooperate with its neighbors in order to compute its location [3,25], which is likely not true in near-future Smart Dust systems.