MICA MOTE 2 DotSince I’m searching documentation about Wireless Sensor Networks, Smart Dust and MOTEs for a college paper, I am going to post there the results of my search. Let’s introduce the topic: Wireless Sensor Networks (WSN) are special networks where nodes are small, low-powered embedded systems with limited processing capabilities and wifi communication capabilities. Tipically the devices are placed in position at the setup time (for example, for structural monitoring in bridges or skyscrapers), or they can be placed on moving “things” (for example: patients in a medical appliance, wild animals in a environmental monitoring appliance and so on). Sometimes the sensors are literally thrown over a wide area (Smart Dust), for example in military surveillance applications.

There are many issues when using this tecnology:

  • How to deal with energy consumption? The devices usually rely on batteries, so you must carefully design your software and on board-os (like TinyOS) to “power off” the sensor that are not needed, and to limit the duty-cycle of the device to extend its lifetime
  • How to deal with network reconfiguration? We are working with an “ad-hoc” “self organizing” networks here. In tipical appliances, the network is very dynamic and the comm links are not reliable. Moving nodes can get “out of contact” and be partitioned from the network; even with fixed sensors, not all of them are always active (power saving mode) so we must ensure that, at any time, there is always a pattern of nodes capable to build a coherent network
  • How to collect data? How to process it? Usually, from the system’s perspective, it’s not important to get a specific data from node X, but rather to collect all the sensor’s data in a basestation and then calculate a coherent aggregate information useful for our application. For example, if we use a WSN to track the movement of an object in the network’s range, we must ensure that the flow of information from the sensors to the basestation is mantained “ordered” wrt time (synchronization issues). Another interesting point is to determine the tradeoffs between node/basestation data processing. For example if we process the data in the node we may be able to optimize the information over the network, but we spend more processing power (and energy) on the nodes

Talking about Wireless Sensor Networks, we can also introduce a widespread known object, the MICA MOTE: this device was initally developed at UC Berkeley, now there are many commercial implementations of it. The MICA MOTE (and MICA2 MOTE) is a small embedded system with modular design. We have a CPU (8 Mhz), RAM (4KB) and Flash Memory (5Mbit). There is a wifi communication module (ranges of 150-300m, 38.4 Kbit/s) and we can mount a wide variety of sensors on the board. The MICA2 DOT (the one in the photo) has basically the same features, but it’s very small! The prices of these devices aren’t high: you can buy’em for a few tens of bucks!

Well, now that I’ve introduced the topic, I will present you some of the links I am collecting about this. To have all neatly ordered, I’ve also opened a list on listible about WSN. That will be updated every time I find something useful, and I hope you can contribute to it too! Just point out at these sites for now: