system by the use of frequency coding

Background and general description

The invention is a primarily a method to substantially reduce the required hardware in wideband Radar Warning (RWR) Systems, Electronic Support Measures (ESM) Systems and/or Electronic Intelligence (ELINT) Systems utilizing digital receivers.

Signal surveillance directed at RADAR is used to characterize different RADAR transmissions to warn a platform for threats (radar warning function), or to create a tactical picture concerning surrounding emitters (ESM function). Signal surveillance directed at RADAR is also used in strict intelligence purpose to create database description of various radar transmissions (ELI T).

It is often of great importance that the signal surveillance system is "wide open", i.e. surveil the complete frequency range and space coverage instantaneously. The main reason is not to miss signals that can pose a threat to a platform. There are different types of traditional receiver architectures that support wide instantaneous frequency coverage and the most common is the so called IFM-receiver. The IFM receiver is largely based on analogue techniques and has a number of drawbacks. Two of these are:

1. The IFM receiver in its basic design cannot resolve time overlapping signals on different frequencies. The occurrence of simultaneous signals gives measurement errors and frequency reading of only one signal (that might neither be the correct frequency for any of the time overlapping signals).

2. The IFM receiver has limitations in sensitivity that are given by the noise floor that can be determined by the receiver RF and video bandwidth according to: where BB is equalent noise bandwidth, BHF represents RF bandwidth and Bv is video bandwidth after the signal detector. The thermal noise floor at room temperature are usually specified to be -114 dBm/MHz and the formula above gives an noise level of approximately -85 dBm at 16 GHz RF bandwidth and 20 MHz video bandwidth. With the receiver's noise figure and the signal to noise margin needed for reliable signal detection in noise the overall sensitivity often achieved is around -65 dBm.

In the recent years the fast development of high performance analogue to digital (A/D) converters has made the market within RWR/ESM/ELINT community to replace IFM-based systems to systems based on digital receiver technology. The reason is to overcome the mentioned drawbacks with the IFM receiver and at the same time other advantages can be achieved, like miniaturization. A digital receiver system is designed and constructed by using an analogue RF input stage that normally filters and frequency convert the incoming RF signals to adapt them to digital sampling. A digital receiver system can be made superior to an IFM receiver regarding measurement precision for RADAR parameters, sensitivity, and handling of time overlapping signals. The latter is important since commercial communication, typically mobile base stations, increases in density and can now be found in traditional radar bands, and interfere with RADAR signal surveillance.

The present problem with digital receivers is to cost effectively create instantaneous frequency coverage over the full spectrum of interest, typically 2-18 GHz. In addition the user often wants to instantaneously measure signal direction of arrival which requires several parallel channels covering the same frequency range. Since the instantaneous frequency coverage of a digital receiver today is limited to a couple of GHz a very large number of receivers is required for a system instantaneously covering 2-18 GHz with instantaneous and accurate direction finding. As an example a 6 channel direction finding system with 4 GHz wide digital receiver building blocks requires 24 receivers (4 frequency channels times 6 complete channels). This result in a very cost driving and complex design.

There are patented methods to reduce the number of digital receivers in a wide band system. One of these methods [1] requires at least two receivers and provides one of the receivers with a time delay prior to sampling and use so called undersampling. With the help of the time delay the phase difference measured between receivers can be used to resolve frequency ambiguities even if the sampling frequency does not support unambiguous frequency measurement for each receiver. However the method requires at least two receivers and one drawback is that "blind" frequency spots are created where unambiguous frequency cannot be determined.

Another method [2] uses several undersampling receivers, each with a specific sampling frequency. By comparing all resulting frequencies in signal processing the true frequency can be determined. The drawback is that several receivers are needed.

Specific description

The present invention reduces the number of digital receivers substantially without lowering the system performance. This is done by analogue frequency channeling and frequency coding of the signals in each channel followed by sampling in one digital receiver. As an example one digital receiver covering just above 4 GHz can be used for instantaneous frequency coverage 2- 18 GHz. In the example with a 6 channel system the number of required digital receivers can be reduced by a factor of 4. The frequency coding in each RF channel is the key of the invention. This makes it possible to convert all incoming signals to a common narrow intermediate frequency range that are sampled without losing the ability to determine which frequency channel a signal originates from. Signals in each frequency channel are automatically frequency coded to consist of a pair of frequencies with a unique frequency difference. In the following signal processing the instantaneous frequency difference between frequency pairs is extracted and the origin frequency channel determined. Since the local oscillator frequencies are known for all frequency channels the signal processing can extract the true absolute frequency. The invention is typically a part of a complete signal surveillance system but the other parts of the system is technology known by prior art. A complete description is however given in this document to put the invention in its typical context. A walkthrough the signal path with reference to figure 1 is given below. Figure 1 exemplifies two frequency channels in a 4 channel architecture.

1. The input signals are received from antenna Ant.

2. The signals are band pass filtered in a wide band filter BF1.

3. The signal levels are adjusted if needed by a digital controlled attenuator Al.

4. The signals are amplified in preamplifier Gl.

5. The signals are split into frequency channels in splitter SI.

6. The signals in each frequency channel are split up in two paths in splitter S2, S3...

7. The signal in each path is mixed with local oscillator 01, 02, 03, 04... Each frequency channel has a unique frequency difference, fl, f2..., between the two oscillators. Often the frequency conversion has to be performed in several steps (with several mixers and oscillators) with adopted filtration in a wide band system in order to avoid undesirable mixing products (harmonics and spurioses) in the pass band. Methods for wide band frequency conversion is known by prior art. The end result however is two signals with a frequency difference fl, f2... in each frequency channel.

8. The signals from all frequency channels are converted into the same intermediate

frequency range to support wide band signal detection with only one band width limited digital receiver.

9. Level adjustment with amplifiers and possibly attenuators and extra bandpass filtering for RF channels are not included in figure 1 but is implemented according to prior art for the specific implementation and selection of components. The two signal paths in each frequency channel are combined in combiner Cl, C2... The signals from all frequency channels are combined in combiner C_l. The resulting combined spectrum is sampled and quantized in an A/D convertor in the digital receiver. Preferably complex sampling is used which provides a quadrature pair, I and Q, that describes both amplitude and phase for each signal sample. When a signal has been detected, frequency estimated and RF channel determined in a fast frequency estimation function (such as Fast Fourier Transform), digital data is routed to a narrowband digital channel that is tuned to one of the frequencies in the determined frequency pair. The narrowband channel needs to be preceded by a time delay to not miss signal content (the frequency estimation takes time). The narrowband digital channel is preferably designed as a digital superheterodyne receiver. A great number of such narrowband channels can be used simultaneously and tuned arbitrarily which supports signal measurements on several simultaneous signals. Measurement of signal Time of Arrival (ToA) and pulse width is preferably made in the time domain in a conventional fashion. Measurement of accurate frequency, amplitude and modulation on the signal is preferably made in the frequency domain. The solution with a digital superheterodyne receiver and methods for signal measurement and parameter estimation is known by prior art and alternative measurement methods is compatible with the invention. The signal descriptor data is signal processed to isolate e.g. pulse trains from each emitter and extract emitter parameters and perform emitter identification with the use of a database storing known emitter parameters. 15. The digital signal processing in a complete system is typically performed according to the flow in figure 2 with clarifications below.

A. The frequency content in the sampled spectra is estimated through a Fast Fourier

Transform or similar algorithm. Several parallel frequency estimation processes can be used to create a kind of matched filtering. Longer data sequences into the frequency estimation gives more narrow digital frequency channels and thus lower noise in each channel. Longer sequences provide therefore higher sensitivity for detection of longer pulse widths. Digital thresholding is used to separate signals from noise with a certain false alarm rate.

B. Frequency separation between signal peaks is extracted and matched against expected separations for each frequency channel.

C. When frequency separation has been extracted the absolute frequency can be

calculated based on the local oscillator frequencies in the determined frequency channel.

D. One of the frequencies in a frequency pair is isolated in a digital narrowband receiver preceded by a time delay to not miss signal content.

E. In the digital narrowband receiver the signal parameters are extracted, such as Time of Arrival, pulse width, accurate frequency, parameters for modulation on pulse and amplitude. Methods for signal parameter measurement and estimation is known by prior art.

F. When the signal parameters is estimated a so called signal descriptor word (SDW) is created. The SDW contains all characteristics of the signal. Method for creating SDW:s is known by prior art.

G. When SDW:s has been created SDW:s are sorted in to groups to isolate signals

originating from the same emitter. For pulses, inter pulse parameters such as pulse repetition interval (PRI), are extracted. Processes for signal sorting is known by prior art.

H. When the signal sorting and inter pulse extraction has been carried out all available data for an emitter is matched against a data base containing known emitter data in an attempt to identify the emitter and present the emitter name to a user. The process for emitter identification is known by prior art.

Figure description

Figure 1 describes the core of the invention.

Figure 2 describes a complete system signal processing flow.