WO2005071433A1

WO2005071433A1 - Method and radar system for clutter filtering of broadband radar signals

Info

Publication number: WO2005071433A1
Application number: PCT/SE2004/000077
Authority: WO
Inventors: Kent Olof Falk
Original assignee: Telefonaktiebolaget Lm Ericsson (Publ)
Priority date: 2004-01-23
Filing date: 2004-01-23
Publication date: 2005-08-04
Also published as: EP1771747A1

Abstract

A method is disclosed for creating a tool for clutter filtering of broadband radar signals by obtaining a measure of the impedance of an echo target. A distance resolution is selected and then a power spectrum is calculated for the necessary transmitting signal. Further a target area profile ρ(b) is estimated by utilising a correlation between the radiated signal and the received signal in form of a convolution. The target impedance is then calculated and thereafter, by means of the achieved impedance characterising the echo of the target, the target can be filtered out.

Description

Method and radar system for clutter filtering of broadband radar signals TECHNICAL FIELD

The present invention relates to a method for differentiating echoes from valid radar targets of interest from targets of no interest as well as disturbances considered as clutter.

BACKGROUND

A desire to be able to distinguish targets from clutter has always has been an objective with radar sensors. Traditionally this has been based on the fact that the velocity of a moving target will be different from background disturbances such as present clutter. Clutter echoes may, for instance, result from small gradients in a surrounding medium like air or a surrounding water surface. Therefore different Doppler filters have been developed to solve this task.

Thus, all present solutions generally seem to be based on the Doppler signal information extracted from received echo signals. However typical solutions of today utilise fully range coded long pulses, for instance using a binary phase code or spread spectrum signals then operating at large receiver bandwidths. Some of these signals will resemble signal noise or clutter signals or a jamming signal source.

Such broadband signal radar signals generally do not exhibit any pronounced carrier, which in turn leads to that a regular velocity filtering can not take place using ordinary Doppler methods. However, with new broadband radar systems there will be further possibilities to measure also other echo parameters and to construct filters which, for instance, may directly filter out reflections from objects having metallic surfaces and separate those for example from those targets with generally non-metallic surfaces. SUMMARY OF THE INVENTION

A method for creating a tool for clutter filtering of broadband radar signals is disclosed by obtaining a measure of the impedance of an echo target. A distance resolution is then selected and a power spectrum is calculated for the necessary transmitting signal. Further, a target area profile p(τ) is estimated by utilising a correlation between the radiated signal and the received signal in form of a convolution. The target impedance thereafter is calculated as a function of time using the relation

where Zo is the impedance of free space. Thereafter, by means of the achieved impedance characterising the echo of the target, the target can be filtered out.

The present method is set forth by the independent claim 1 and further aspects are defined by the dependent claims 2 to 4. A radar system utilising the present method is set forth by the dependent claim 5.

DETAILED DESCRIPTION As indicated, the new broadband radar systems open up possibilities for measuring also other parameters, for example an indication of impedance of the targets and then constructing filters which filter out, for instance, reflections from metallic objects {Z=0).

To simplify calculations, just suppose a bandwidth limited broadband noise signal having an average value equal to zero in accordance with: s{t) = S_j{t) -cos(2-π • f_Q -t)-s_Q(t) -ήn{2 -π ■ f₀ -t) (1)

where sι(t) and SQ(t) are two Gaussian noise processes and fo is the centre frequency. Gaussian noise does have a rather large amplitude variation, which involves that an amplitude-limited transmitter will be ineffectively used. Phase modulated noise of the type s(t) = cos[ωot + θ(t)] having some suitable statistical distribution may therefore practice be more preferable. Equation (1) may then alternatively be written: s(t) = A(t) -cos[2-π -f₀ -t + φ(tj] (2)

The envelope function A(t) is Rayleigh distributed and given by:

A(t) = (t) + f_Q (t) (3)

The phase function φ(t) is equally distributed over the interval \-π, τi , and given by:

We also may represent the signal in a complex analytical form according to:

*( =rW -^ ^{β< +}<( -^^{" β (5)}

Suppose that the radar station is stationary and transmits an ergodic wave form s(t) with an average value of zero. Ergodic theory can best be described as the statistical and qualitative behaviour of measurable group and semigroup action on a measurable space possessing a non-negative measure. (Ergodic theory has its origins in the work of Boltzman in statistical mechanics problems where time- and space-distribution averages are equal.) This adoption simplifies the calculations, but implies still no limitation. A moving radar station having the velocity v implies that time is scaled according to t(τ) = to + 2-v-τ/co. The received signal from the reflection profile p(t) for the distance 0 to Rmax can be written according to:

r(t) = ^° p(ζ)-s(t-ζ)-dζ + n(t) (6)

where Co is the velocity of light and n(t) represents additive noise and possible disturbances. The reflection profile p(τ) lies within the unity circle in the complex plane and represents the sum of all subset reflections within the resolution cell. For a resolution cell for instance constituting a mix of air and metal the reflection profile p(τ) will be a sum of contributions having a reflection coefficient of 0 and -1, respectively. Thus, the resulting reflection coefficient will be positioned along the line between 0 and -1.

According to the theory of ergodic processes r(t) also becomes stationary and ergodic. Therefore time correlation may be utilised to approximate the cross correlation between a time delayed copy of the conjugate of the transmit signal s*(t-ζ) and the received signal r(t) according to:

g_T (t,τ) = n(t)-s^* (t-τ) - dt (7)

where T represents the correlation of the integration interval. For a large T the second term in Eq. (7) will go towards zero since n(t) is not correlated to the transmit signal s(t). This gives: g{τ) = mg_T (t) = ^' p(ζ) -p(τ -ζ) -dζ = p(τ)®p(τ) (8) o <8> symbolises a convolution. Besides it is valid that:

where p(τ) is the auto correlation function of the radiated noise signal s(t). It clearly comes out from Eq. (8) that the received signal is the convolution between the reflection profile and the auto correlation function of the radiated signal. Thus the distance resolution is given by the auto correlation function p(τ) of the transmit signal, which depends on the bandwidth of the transmit signal and the form of its power spectrum P(ω). It is easy to show that there is a simple relation between the auto correlation function p(τ) of the signal and its power spectrum P(ω), since these two expressions form a pair of Fourier transforms:

(10)

Consequently it is possible to select a distance resolution p(τ) and calculate the power spectrum P(ω) for the necessary transmit signal. A convolution in the time domain then corresponds to a multiplication in the frequency domain. Therefore the target area profile p(τ) can be estimated according to:

and the impedance as a function of time (distance) can be calculated according to:

Z{τ) = Z₀ (12) l + p(τ) where Zo = (μo/εo)'^A is the impedance of free space (120-π or « 377 ohms). The time τ can be recalculated into distance R according to:

R = ^l (13)

We know for instance, as already mentioned, that metallic objects are characterised by p = - 1 and Z = 0. Thus, a target having another typical reflection coefficient and impedance can be filtered out in the same way as in the example with metallic objects. However it should be noted that the reflection coefficient of each individual resolution cell represents the sum of all partial reflections within the cell. In practice a threshold value representing a certain distance from the origin of coordinates need to be exceeded in order to obtain a detection of the impedance of an echo. The angle within the complex plane for the detection concerned decides the phase of the sum of partial reflection within the resolution cell. For instance a dominating metallic contribution will result in an angle close to -π.

Generally it is valid that the impedance for an object is given by its conductivity σ, dielectric constant s and the permeability μ according to:

Z(ω,σ,ε_r,μ_r) - J J - a ^{w '} - -rr ^' >"o0 (14) σ + j - ω - ε_r - ε₀

The detection of metallic objects is extremely simple since the frequency dependency of Eq. (14) vanish as Urn z(ω,σ,ε_T,μ_T) = 0 . σ→∞

In other words a filtering of the impedance in accordance with the present invention will be able to extract echoes from a metallic target in a situation with a stationary target against even a stationary background clutter, whereby a standard Doppler detection then would not be able to distinguish the target from the background clutter. For the further separation of an achieved target impedance from another detected impedance any regular filtering method may be used, which is well known for a person skilled in the art.

It will also be understood by those skilled in the art that various modifications and changes could be made to the present inventive method without departure from the spirit and scope thereof, which is defined by the appended claims.

Claims

1. A method for creating a tool for clutter filtering of broadband radar signals by obtaining a measure of an impedance of a target, characterised by the steps of: selecting a distance resolution (p(τ)) and calculating a power spectrum [P(ω)) for a necessary transmit signal; estimating a target area profile (p(τ)) by using a correlation (g(τ)) between a radiated signal and received signal; calculating a target impedance Z(τ) as a function of time using a relation

where Zo is the impedance of free space; and filtering out by means of a filter the target using the target impedance achieved.

2. The method according to claim 1, characterised by the further step of calculating the correlation (g(τ)) between the radiated signal and received signal as a convolution of the target area profile (p(τ)) and the distance resolution (p(τ)).

3. The method according to claim 1, characterised by the further step of recalculating the time (τ) into distance {R) using the fact that R = Co ^• τ -Vz, where co is the velocity of light.

4. The method according to claim 1, characterised by the further step of visualising an angle of the target impedance being the sum of partial reflections within a resolution cell using a complex presentation plane, whereby a dominating metallic contribution of the detected target echo will result in an angle close to -π.

5. Radar system characterised in that said radar system utilises the method according to the preceding claims 1 to 4.