A state of the art Smartphone such as the iPhone 4 or Samsung Galaxy S2 is loaded with various sensors, such as motion sensors and accelerometers, a gyro, a camera (or two), a microphone, WiFi and Bluetooth radios (which can be used both for communication and for sensing RF signals), Near Field Communication (NFC), a touch screen, etc.
In addition, they are increasingly equipped with high performance processors, such as the 1.2 GHz dual core Cortex-A9 processor of the Galaxy S2. And of course, these phones run a fully-fledged operating system and modern developing languages such as Java. All this comes with a price tag of $500-$900.
Compare this with modern mote sensing nodes, having a memory of a few Megabytes, a terribly slow processor, and running a special purpose OS like tinyOS, which requires unique programming skills and development environments, while priced at $100-$150, with each sensor costing an extra few bucks. Given the economy of scale of smartphones and the tough competition in this market, their prices are likely to be halves within a year, as new models come in. Moreover, when cellular communication in not important, one can settle for the Galaxy S WiFi for around $400.
Given the cost of programmers, with the exception of extremely large deployments, it is cheaper to use smartphones than sensor nodes. In terms of lifetime, sensors can typically survive longer than smartphones. Yet with an extra battery, one can dramatically prolong the latters’ capabilities as well and still be affordable.
Even more interesting is the ability to utilize smartphones for crowdsourcing sensor networks. Consider for example earthquakes. When an earth quake occurs, naturally all phones in the area will shake.
Hence, by having all phones that shake beyond some threshold report to a cloud server both the magnitude and nature of the shaking as well as their position, either over their data channel or by SMS, the server could compute whether an earth quake is happening and what its magnitude is. Based on this, warnings can be sent out to other areas to prepare for the earthquake, e.g., by shutting elevators and power plants, securing hard disks of important computers, and the likes.
Similarly, mobile phones can be crowd-sourced to detect noise hazards, excessive RF radiation, etc. One company that already does something along these lines is Waze, which crowd-source mobile phones to detect traffic jams. I am sure others will soon follow.