Image Processing Reference
In-Depth Information
•
Financial costs for establishing a localization infrastructure must be minimal. This
implies cost-efficient beacon hardware as well as minimal number of used beacons.
•
The limited resources require minimal effort in computation, communication, and
memory consumption.
•
Localization must be processed fully distributed to achieve robustness and reliability.
•
Algorithms must be adaptive to changing environmental conditions, e.g., fluctuations in
the quality of the signal strength indicator.
•
Mobility of nodes must be considered.
However, not all requirements can be simultaneously fulfilled. This is subject to the underly-
ing application. For example in a highly mobile network, it is very important to estimate positions
very quickly, otherwise they are not up to date. Summarized, both precise localization and minimal
resource requirements are fundamentals in wireless sensor networks.
6.3.3 Algorithm Classification
Localization algorithms can be categorized as follows:
•
Approximative versus exact precision
•
Central versus distributed calculation
•
Range-based versus distance-free (or angle)
•
Relative versus absolute localization regarding point of reference
•
Indoor versus outdoor usage
•
Beacon-free versus beacon-based
The limits between these characteristics are floating but allow a general classification. Next, the
most relevant algorithms are described in detail. Figure . gives a short comprehensive overview.
6.3.4 Classical Methods
From mathematical point of view, a position can be calculated by triangulation, namely trilater-
ation. If angles are provided, the sensor node's position
S
(
x
;
y
)
is estimated with a triangulation
asshowninFigure.a.Indetail,
B
(
x
;
y
)
and
B
(
x
;
y
)
are basis for a line of position
B
B
.
Localization algorithms
Classical
methods
Proximity
based
Optimization
methods
Iterative
methods
Pattern
matching
Sweeps
APS
Fingerprinting
Pattern
Recognition
Trilateration
Triangulation
Nearest beacon
Centroid
localization
Fine-grained
localization
Distributed least
squares
Convex position
estimation
ABC
Weighted centroid
localization
Range free
Bounding box
FIGURE .
Classification of localization algorithms.
