Database Reference
In-Depth Information
Table 3.1.
Energy consumption of radio in different status
Radio Status
Power
Sending
60mW
Receiving
45mW
Listening
45mW
Sleeping
90
μ
mW
2.2 Other Constraints
Sensors have limited computational and communication resources.
The technology of sensors has evolved for several generations. Mica [37]
is a popular series of second-generation commercial sensors. A Mica node
has a 4 MHz, 8 bit Atmel microprocessor and a 40 Kbits/second radio
device. More advanced sensors, such as Mica2, a third generation com-
mercial sensor, are equipped with 7 MHz processors, 128Kbytes program
flash memory and radio devices with bandwidth 38.4 Kbits/sec. Even for
Mica2, the computational and communication resources are extremely
valuable; i.e., it is undesirable for an algorithm on sensor nodes to store
large amounts of data and perform complex computations. During data
transmission, there is always an upper bound on the packet size. A large
message is divided to fit in the packet before transmission. Similar to
other wireless radios, the radio devices of Mica nodes are half-duplex;
i.e., they are not able to listen to the incoming signals during message
transmission. In order to avoid collisions, before sending a message, the
nodes listen and detect whether the transmitting channel is in use. If
so, they delay their own transmissions for a random period of time and
then try again. In general, unreliability in WSNs is attributed to the
following aspects:
Sensor nodes are unreliable. A sensor may occasionally take wrong
sample readings. Errors may occur during message transmission.
More severely, a node may stop functioning for a short period of
time and come back again.
The links between sensors are unreliable. Since the links are wire-
less, they are sensitive to the physical status of the environment.
Turbulences in the environment (e.g., some object passing through,
the change of humidity, etc.) may affect the link quality.
Thus, an important parameter in a WSN is its packet loss rate, which
indicates the average probability that a packet gets lost in the path to the
base station. Algorithms developed for wireless sensor networks should
be able to handle message losses and possibly recover the missing data.
