In the present lecture, relevant examples of adaptive radar signal processing techniques are surveyed. An adaptive system performs the processing on the incoming signals by using an architecture having time-varying parameters. The corresponding filtering mask is taylored to the actual interference (i.e. clutter) which is real-time estimated by the same input signal. Additionally, the mask is able to track variations of interference power spectrum which may occur during the time. This approach overcomes the inherent limitations of conventional systems based on filters (e.g. MTI, FFT) having predetermined coefficients. These non adaptive techniques suffer poor interference cancellation when the expected environmental conditions significantly differ from the actual ones or when the experienced interferences vary unpredictably during the time. A brief recall of the theory of optimum signal processing is given. To this purpose, the most relevant concepts of the Wiener, Levinson, Kalman and Brennan Reed filtering are pointed out. All these theories provide useful insights and analytical means for the adaptive techniques considered. Five adaptive processors are examined, all sharing a common theoretical background provided by the optimum filtering theory, namely: Parametric Estimator (PE), Gram-Schmidt (GS) orthonormalization algorithm, Direct Matrix Inversion (DMI) technique, Maximum-Entropy-Method (MEM), Kalman Filter (KF). The working principle and the implementation algorithm of each above-mentioned processor will be de scribed and the performance achieved will be compared with respect to the optimum procedure. Also, the problems concerning with their hardware implementation will be mentioned, even though further researches are needed in this area. The revised adaptive techniques seem to be a valuable upgrading of the conventional fixed parameters processor.

The paper deals with the application of Gram-Schmidt decorrelation algorithm to the clutter cancellation and useful signal enhancement in radar signal processing. The proposed system architecture is an adaptive implementation of the well known optimum processor. The performance of the selected approach is evaluated and compared with that of the optimum one. As a general result, the adaptation time is a few tens of range cells and the losses are negligible. Finally, the hardware complexity is briefly considered.

A sidelobe canceller can be efficiently employed to reduce the effect of jammers received through the sidelobe of a radar system. A well established technique refers to the SideLobe-Canceller (SLC) approach, in which external aerials (called "auxiliary antennas") placed around the radar antenna (called "main antenna") are subject to control. The signals received through the auxiliary antennas are multiplied by proper weights and then summed obtaining an esti mate of the jammer received through the radar sidelobes. The cancellation is performed by subtracting the jammer estimate from the radar output. The weights are usually obtained by evaluating the correlation coefficients between each auxiliary signal and the residue of cancellation. The processing is performed by an adaptive loop, Howells-Applebaum technique, consisting of a multiplier and a low pass filter. The number of auxiliary antennas determines the degree of freedom on the sidelobe structure of the overall system and, thereby, the number of jammers which can be cancelled. Fuller details of this technique can be found in Monzingo and Miller, Hudson. Two figures of merit define the SLC system performance: the power cancellation ratio and the time required for adaption of all the loops. Unfortunately these figures are in con trast to each-other in the Howells-Applebaum implementation. In fact, the greater the loop bandwidth, the faster its response to a non-stationary jamming situation; however, a wider bandwidth reduces the filtering effect on the input jamming process. A detailed analysis of these conflicting effects on the jammer cancellation, for a SLC having two auxiliary aerials, have been described by Farina and Studer. One way to speed convergence is based on the Gram-Schmidt or thogonalization procedure which maintains the same steady state cancellation of the standard loop and it is easily implemented. This paper gives the performance evaluation of this canceller. Mathematical expression of the steady-state cancellation as a function of the ratio between the adaptive circuit bandwidth to the radar receiver bandwidth is given. A comparison, in terms of performance and implementation complexity is made with the canceller based on the Howells-Applebaum technique.

In this paper two adaptive algorithms to implement the optima radar signal processor are presented and their performance are evaluated. The first method considered derives from the application of the Gram-Schmidt orthonormalization algorithm to the input. The second one refers to an algorithm for the direct inversion of the clutter covariance matrix.

Two high resolution techniques for spectral estimation of random processes are described. The estimate is obtained from samples of the autocorrelation function or directly from samples of the process. These techniques are based on the Maximum Entropy Method (MEM). After a review of the main points of this method, an application to radar system is shown in detail. First, the estimation of clutter spectrum is considered; then, this estimate is exploited to shape a filter for clutter cancellation and target echo enhancement. The processing algorithm is an adaptive one, and its performances are evaluated, by means of computer simulation, in term of Improvement Factor and speed of adaptation.

The performance of an optimum radar signal processor and more conventional techniques (such as MTI. adaptive MTI, and coherent integration) are compared. A mathematical method is suggested and applied to several cases of practical interest, A number of operative conditions are discovered in which the conventional processing techniques give very poor performance and the optimum radar processor becomes necessary.

This paper provides general models of radar echoes from a target. The rationale of the approach is to consider the echoes as the output of a linear dynamic system driven by white Gaussian noise (WGN). Two models can be conceived to generate N target returns: samples generated as a batch, or sequentially generated one by one. The models allow the accommodation of any correlation between pulses and nonstationary behavior of the target. The problem of deriving the optimum receiver structure is next considered. The theory of "estimator-correlator" receiver is applied to the case of a Gaussian-distributed time-correlated target embedded in clutter and thermal noise. Two equivalent detection schemes are obtained (i.e., the batch detector and the recursive detector) which are related to the above mentioned procedures of generating radar echoes. A combined analytic-numeric method has been conceived to obtain a set of original detection curves related to operational cases of interest. Finally, an adaptive implementation of the proposed processor is suggested, especially with reference to the problem of on-line estimation of the clutter covariance matrix and of the CFAR threshold. In both cases detection loss due to adaptation has been evaluated by means of a Monte Carlo simulation approach. In summary, the original contributions of the paper lie in the mathematical formulation of a powerful model for radar echoes and in the derivation of a large set of detection curves. An additional contribution is related to the use of the important concept of estimator-correlator for radar applications, a concept not usually employed by radar engineers.

The paper deals with the problem of radar detection of a target echo embedded in log-normal clutter and white Gaussian noise. Relevant features of this article, with respect to previous papers on the same subject, refer to the coherent model assumed for the clutter and the processing chain. In more detail, the in-phase and quadrature components of clutter have been modelled to give a log-normal amplitude distribution and a near uniform distribution of the phase. Any shape of the correlation among consecutive clutter samples is also allowed in the model. At the same time, the processing chain is also coherent, i.e. it operates on the two components of the signals. Two architectures have been considered for the processor. The first, used in current practice, is formed of a linear transversal filter (for the clutter attenuation and the target echo enhancement) cascaded with a quadratic envelope detector and a comparison with a suitable threshold. The second processor considered differs from the previous one in the filter for clutter cancellation. A nonlinear homomorphic filter has been conceived to obtain a better suppression of clutter. The detection performance of the two processing chains have been evaluated, by means of computer simulation, in a number of operational cases of interest. The paper gives a first contribution to the problem of finding better models of disturbance and of deriving more efficient processing chains.

The first part of the paper provides flexible and reliable stochastic models for the radar signals scattered by target and clutter sources. The models allow to consider any shape of autocorrelation function between consecutive pulse echoes and any probability density function for their in-phase and quadrature components. The second part of the paper revises the theory of detecting targets, with any type of probability density and autocorrelation function, embedded in a disturbance having any type of probability density and autocorrelation function. In the third part of the paper, the theory is applied to the cases in which target and/or disturbance may have a log normal probability density for the amplitudes. Several processing schemes are suggested and corresponding detection performances evaluated. Finally, adaptive implementation schematics are suggested for some of the processors presented.

The present paper pursues, extends and concludes the theory and the presentation of results initiated with the papers /1/, /2/ and /3/ prepared by the same authors. The purpose of the present paper is manifold. First, a brief revision is provided of the mathematical background to derive radar detection algorithms for any type of probability density and autocorrelation functions of target and clutter. A relevant requirement concerning this theory is the derivation of adequate mathematical models for the target and clutter processes. This is done with particular reference to the Lognormal and Weibull cases. A remarkable result refers to the derivation of models for the coherent echoes train case. The leading concept of “Whitening and Gaussianing” filter is then introduced as a fundamental block to derive radar detection schematics. The aforementioned theory is applied to the derivation of completely new detection schemes and to the evaluation of the corresponding detection performance when the amplitude probability density of the clutter is Lognormal or Weibull. Another novelty of this paper refers to the presentation of detection schemes having adaptive features. More in detail, methods are suggested -for the on-line estimation of the “Whitening-Gaussianing” filter weights. Results are presented concerning the detection loss versus the number of range cells along which the average of weights estimate is performed. Detection loss are evaluated for different number of processed pulses and for different parameters of clutter and target signals. Another adaptive feature explored refers to the on line evaluation of a CFAR detection threshold. Even in this case, an evaluation of the corresponding detection loss is enclosed.

The paper deals with the problem of radar detection of a target echo embedded in Weibull clutter and white Gaussian noise (WGN). Relevant features of the paper, with respect to previous papers on the same subject, refer to the coherent nature of the Weibull process (that modelling the clutter) and of the processing chain. In more detail, the in-phase and quadrature components of the clutter echoes have been modelled to give a Weibull probability density function (PDF) of the amplitude and a uniform PDF of the phase. Any shape of the correlation function among consecutive clutter samples is also allowed in the model. The so called 'coherent Weibull clutter' (CWC) introduced in the paper represents a suitable generalisation of the conventional 'coherent Gaussian clutter' (CGC). The processing chain is also coherent, i.e. it operates on the in-phase and quadrature components of the signals. To derive a suitable processing scheme, we resort to the general theory of radar detection which applies to any type of PDF and autocorrelation function (ACF) of the target and clutter. Briefly, it is found that the processing scheme is based on two nonlinear estimators of the clutter samples in the two alternative hypotheses (i.e. H₀ and H₁), the fully fledged architecture being discussed in the paper. The detection processor turns out to be a suitable generalisation of that concerning the CGC case. A new family of processing schemes is derived in accordance with the statistical model assumed for the useful target. The following cases are covered: target known a priori, Swerling 0, 1 and 2 models, and partially fluctuating target. Attention is also paid to the interesting case of a target modelled as a coherent Weibull process. The detection of such a target against white Gaussian noise is worked out. The detection performance of the afore-mentioned processing schemes has been evaluated in a number of operational cases of interest. Another novelty of the paper refers to the conception of detection schemes having adaptive features. In more detail, methods are suggested for the on-line estimation of the parameters of the two nonlinear estimators of the clutter. Results are presented concerning the detection loss against the number of range cells along which the average of parameters estimate is performed. Another adaptive feature explored refers to the on-line setting of a constant false-alarm rate (CFAR) detection threshold. Also in this case, an evaluation of the corresponding detection loss is made. To summarise, the paper represents a suitable generalisation of the detection theory of a coherent Gaussian target embedded in a CGC. This new theory proves its usefulness in advanced radar systems that should have high performance, notwithstanding challenging environments including spiky clutter and/or spiky target (e.g. stealth).

The optimum receiver, allowing for the presence of Gaussian noise and Rayleigh fluctuations of the target echoes, is synthesized for a multi static radar system. The performance of such a receiver is analyzed, showing that in certain respects the target-detection characteristics of the multistatic system compare advantageously with those of the monostatic system. The inherent complexity of the multistatic structure may be alleviated by resorting to a simplified scheme, here suggested, which envisages a number of peripheral decisions in as many conventional receivers, followed by a central decision based on an OR-criterion. The latter de sign admittedly yields a sub-optimum receiver, but an assessment of its performance shows the existence of a wide range of operating conditions of practical engineering interest where performance degradation with respect to the optimum receiver is negligible.

The paper deals with the design and performance evaluation of optimum and suboptimum multi static radar receivers. Several detection schemes are considered according to the possible models for the wanted signal: completely known; random phase; Rayleigh amplitude and random phase; and one-dominant-plus Rayleigh amplitude and random phase. The inherent complexity of the multistatic structure may be alleviated by resorting to a simplified scheme suggested here which envisages a number of peripheral decisions in as many conventional receivers, followed by a central decision based on an OR criterion. The latter design admittedly yields a suboptimum receiver, but an assessment of its performance shows the existence of a wide range of operating conditions of practical engineering interest where performance degradation with respect to the optimum receiver is negligible.

The purpose of this paper is to determine receiver structures suitable to detect a target, by properly processing a number of distinct radar observations obtained by the diverse sensors. In view of this, optimum receivers are synthesized according to a Neyman-Pearson strategy, and their performances are evaluated in order to assess the effects of the number of diverse sensors, of the number of distinct observations at each sensor and of the values of the signal-to-noise ratios in the diverse paths. It is assumed that the wanted signal is embedded in white Gaussian noise, exhibiting no correlation among the diverse paths. At first attention is devoted to the case of coherent pulse trains, for which previously published results can be directly applied. Then the case of non-coherent pulse trains is considered, for which the extension is not trivial. It is anyway of great practical interest that, under the assumption of low signal-to-noise ratios in the diverse paths, the receiver structure is robust with respect to the fluctuation models usually assumed for the target, (namely the four Swerling trains and the non fluctuating incoherent trains).

This paper deals with the problem of optimum detection of radar signals by means of a multistatic configuration of sensors. Starting from the case of multiple observations, a systematic analysis of the model leads to a unified optimum receiver for signals fluctuating according to Swerling I, II, III & IV models. In these instances the receiver performances are evaluated by adopting an original method. Finally, possible suboptimum solutions are found, through which considerable streamlining of processing structures can be achieved.

The detection problem with multistatic radar systems is considered, resorting to the theory of detection of coherent target signals having a Gaussian probability density embedded in coherent Gaussian-distributed disturbances (i.e. clutter and/or directional jamming). A novelty of the paper is related to the capability of dealing with any type of time autocorrelation function of both target and disturbance. The spatial correlation among the signals scattered towards the receivers can also be taken into account. The latter topic is relevant when the receivers are deployed in a small area with respect to the spatial extension of the autocorrelation function of the scattered signals. Mathematical procedures have been conceived to derive the detection processors and for evaluating the corresponding performance, and have been applied to a number of original cases of interest. Finally, problems concerning the on-line implementation of the proposed processors family are discussed.

The purpose of this paper is manifold. The first issue is a comprehensive survey of signal processing techniques presently implemented in modern surface based radar and those which will be likely applied in the future years. Another major part of the paper refers to the use of these techniques in different, rather promising, system concepts such as multistatic and dispersed radars, Airborne Early Warning (AEW) and Spaced Based Radars (SBR). The paper starts with the requirements for enhancing surveillance performance, then the techniques currently implemented in the today radar are considered. Attention is devoted to signal coding, anticlutter and ECCM techniques for three-dimensional radars. Limitations of these systems are also mentioned. New techniques which are expected to be available in the future systems are then described.The following topics are considered: Low Probability of Intercept (LPI), anti ARM, anti-stealth, digital beam forming, adaptivity, high-resolution, multidimensional processing and target classification. The fleldability of each technique is also estimated according to the expected benefits to be achieved in terms of surveillance performance, the degree of confidence and the technical risk. Attention is then turned to system concepts alternative to today's surface based radars. Advantages accruing to the multistatic radar concept and corresponding drawbacks are examined in some detail. A generalization Of the multistatic radar brings to the other relevant system concept of distributed array. The class of airborne and space based radar is finally considered. They are particularly suited for surveillance because of the extremely large field of view. Pulse Doppler signal processing for clutter cancellation and unambiguous target detection is a relevant technique considered in AEW radar. SBR are described in terms of required technologies and criteria for selecting radar parameters (e.g. frequency and power-aperture product), number of satellites, satellite altitude and type of orbit. The last part of the paper describes the relevance of new technologies and processing architectures. The description of the abovementioned topics is corroborated by an extensive list of References covering the last years.

The purpose of this paper is to discuss the problem of achieving constant false alarm rate (CFAR) operation in radar systems and the techniques used for this purpose. First, the characteristics of non-coherent clutter in both space and time domains are assessed. Attention is then focused on the two main approaches to CFAR detection, spatial adaptive thresholding and temporal adaptive thresholding. For each approach, a number of currently used techniques are reviewed. Performance evaluation and comparison between the two classes of de vices also are considered. Other CFAR techniques based on coherent processing are outlined and discussed.

The problem of the synthesis of pulse compression radar signals is examined. A family of waveforms suitable for digital compression is identified in digital (or discrete) chirp. Such a code, obtained through sampling and quantisation of known chirp waveforms, has attractive features such as relative insensitivity to frequency shift (Doppler effect) and low sidelobes. Within the digital pulse compression technique, with compression ratio higher than a few tens, digital chirp codes perform belter than other codes suggested.