Page 202 - Radar Technology Encyclopedia
P. 202
filter, microwave filter, pi-[section] cavity 192
belong to the class of reciprocal multiports. Ferrite filters, Multistage filters are used, for example, for matched fil-
which can be retuned within a frequency band, possess nonre- tering of pulse-compression waveforms through cascaded
ciprocal properties. connection of delay lines. IAM
Design of microwave filters is quite complex and is done
A nonlinear filter is one whose input and output signals are
in several stages, beginning with a simplified circuit, the pro-
related through nonlinear differential equations. This is typi-
totype circuit, and ending with parameter adjustments from
cal when signals or interference have non-Gaussian distribu-
the results of experiments.
tion or when interference is not additive. In practice,
Microwave filters are used in microwave circuits and
quasilinear filtering methods are often used. SAL
also to form inter-stage connections in receiving and trans-
Ref.: Korostelev (1987), p. 260; Fal’kovich (1981), p. 203.
mitting devices. They are also called RF filters. IAM
nonrecursive filter (see recursive filter).
Ref.: Saad (1971), p. 153; ITT (1975), Ch. 7–9; Sazonov (1988), p. 122; Gas-
sanov (1988), p. 51.
An optical filter is a filter based on the phenomenon of varia-
A mismatched filter is one that fails to meet the specification tion of light amplitude and phase when it passes through an
of a matched filter and hence provides an output signal-to- optically inhomogeneous medium. It has broad bandwidth
noise ratio less than the ratio of input signal energy to noise and makes it easy to implement multichannel and multifunc-
spectral density. It thus has a mismatch loss, relative to that tional signal processing. Practically any frequency and phase
optimum performance. It is often used in radars to simplify response can be synthesized in the optical filter. The basic
the filter design, to reduce time sidelobes in a pulse-compres- components of the optical filter are the optical system, units
sion system, or to compensate for expected presence of non- storing input signal and filter frequency response that are
white (colored) noise. IAM light modulators, and the units of photoelectron registration at
Ref.: Barton (1964), p. 20, (1969), p. 84; Sloka (1970), pp. 19, 64. the output. The advantages of optical filters are very suitable
for analog filtering of pulse-compression waveforms. IAM
A moving-target-indicator (MTI) filter is a filter intended
to reject clutter components at and near zero frequency, or Ref.: Sloka (1970), p. 221; Zmuda (1994), p. 408.
near the center frequency of the clutter spectrum, while pass- An optimum filter provides the maximum attainable perfor-
ing as much as possible of the target signal spectrum. The fre- mance for a specified optimization criterion. Two typical
quency responses of the single- and double-delay line examples are filters that provide maximum output signal-to-
cancelers, often used in MTI systems, are given by noise ratio or minimum rms error in signal fidelity. In radar
applications, the term optimum filter is usually applied to a
æ f ö
H f () 2 sin= p --- linear filter providing the maximum possible signal-to-inter-
1 è f ø
r ference ratio. In this case the filter does not retain the shape of
the input signal but deliberately distorts it so that spectral
2
æ f ö
H f () 4 sin= è p--- r components of the signal are superimposed to maximize the
f ø
2
signal-to-interference ratio. The useful by-effect of optimum
filtering in this case is a pulse compression phenomenon,
where H(f) is the frequency response and f is the pulse repeti-
r widely exploited in modern radar.
tion frequency. More complicated response is available when
In general, when the interference consists of white noise
feedback is used around the cancelers, or when range-gated
of spectral density N , and clutter has an energy spectrum
0
channels are filtered in low-pass digital or analog filters (see 2
|S(f)| , the square of the signal spectrum, the optimum filter
also CANCELER). DKB
transfer function is
Ref.: Barton (1988), pp. 236–238.
S * f ()
Multidimensional filtering is the filtering of a multidimen- Hf () --------------------------------
=
sional signal (a signal that depends on a set of parameters), N + kS f () 2
0
typically during integration, for the purpose of more precise
where k is a constant related to the clutter-to-noise ratio.
estimation of signal parameters. In the simplest case, a scalar
When the interference is random white noise, the optimum
parameter linearly dependent on the multidimensional signal
filter is called a matched filter, and its transfer function from
is evaluated; for example, when evaluating generally grouped
the equation above is simply the complex conjugate of the
coordinates of complex targets and when integrating uniform
signal spectrum. DKB, SAL
measurements in radio navigation. Similar problems are
Ref.: Nathanson (1969), pp. 301–306; Leonov (1988), p. 61.
solved using linear filters. In the general case, nonlinear fil-
ters are used to evaluate multidimensional parameters. IAM A pi-[section] cavity filter is a microwave bandpass filter
Ref.: Fal'kovich (1981), p. 237. with pi-shaped bent cavity. It uses striplines and has smaller
dimensions compared with other types of bandpass filters and
A multistage filter consists of a cascade of similar stages.
also uses a simple production technology. It is used in the
The frequency response of a multistage filter is the product of
low-frequency microwave range as well as for relative bands
frequency characteristics of the stages.
of more than 20% through the use of its advantages in pro-