Image Processing Reference
In-Depth Information
FIGURE .
Wireless sensor node or Mote made by Crossbow Technology, Inc. in San Jose, California.
speed, storage capacity, and communication bandwidth. Additional design challenges include limited
power that a Mote can harvest or store, ability to withstand harsh environmental conditions, ability
to cope with node failures, mobility of nodes, dynamic network topology, and unattended operation.
These constraints require new hardware designs and novel network architectures. Several standards
are currently under development for wireless sensor networks. One example is ZigBee, which is a
mesh-networking standard intended for uses such as industrial control, embedded sensing, med-
ical data collection, building automation.
∗
Another recent standard is WirelessHART, which is an
extension of the HART Protocol for industrial automation.
†
From a signal-processing point of view, the main challenge is the distributed fusion of sensor data
across the network. This is because individual sensor nodes are often not able to provide useful or
comprehensive information about the quantity under observation. Furthermore, due to the variable
environmental conditions in which sensor devices may be deployed, one can expect a fraction of the
sensor nodes to be malfunctioning. herefore, the underlying distributed algorithms must be robust
withrespecttodevicefailures.
In this chapter, we introduce a powerful class of information fusion algorithms, which are
based on formulating the sensor fusion problem as a “convex feasibility problem.” There are
great advantages to formulating a sensor network fusion problem as a convex feasibility prob-
lem. The most basic advantage is that a solution can be found in a distributed fashion by using
a series of independent projections onto independent convex sets. Another key advantage is that
the individual projections can be computed, very reliably and efficiently, using methods for con-
vex optimization []. These solution methods are reliable enough to be embedded in real-time
system.
∗
See the ZigBee Alliance Web site at www.zigbee.org
†
See the HART Communication Foundation Web site at http://www.hartcomm.org/index.html
