Semicoherent channel estimator

1. A method for processing a complex ambiguity function (CAF) surface comprising:

providing incoming data to a processing unit;

providing a model vector to the processing unit;

creating by the processing unit of a CAF surface from the incoming data and the model vector, wherein said CAF surface has at least one feature;

processing of said CAF surface by the processing unit so as to locate a peak on said CAF surface;

establishing by the processing unit of a sub-window about said peak, wherein said peak is within said sub-window; and

flattening by the processing unit of said peak.

2. The method according to claim 1 , further comprising:

processing of said CAF surface by the processing unit so as to locate a next peak;

establishing by the processing unit a next sub-window about said next peak; and

flattening by the processing unit of said next peak.

3. The method according to claim 1 , wherein said flattening is accomplished by a flattening function defined by performing an autocorrelation of the model vector and inversing the autocorrelation.

4. The method according to claim 1 , wherein said sub-window is a range about the feature such that said sub-window includes sufficient pixels to represent an entire autocorrelation function.

5. The method according to claim 1 , wherein said sub-window is represented as: h H (τ 0 , f 0 ) h(τ, f), wherein: h(τ, f) represents a neighborhood.

6. The method according to claim 1 , wherein said feature contains a plurality of two dimensional pixels having some time delay and frequency offset.

7. The method according to claim 1 , wherein said peak occupies at least one pixel and represents an area on the CAF surface of increased amplitude.

8. The method according to claim 1 , wherein said flattening includes application by the processor of a flattening function, wherein computation of said flattening function further comprises centering said peak in said sub-window, performing an autocorrelation function, and inverting said feature.

9. The method according to claim 1 , wherein said flattening function is represented by: f(τ,f)=1−∫ h H (τ 0 , f 0 ) h(τ, f).

10. The method according to claim 1 , wherein said processing continues until a condition is satisfied according to one of a group consisting of: constant false alarm processing; setting a margin level; and setting a customizable level.

11. A system for processing a complex ambiguity function (CAF) surface comprising a processing unit, and software operable on the processing unit, the software being configured to cause the processing unit to:

generate a pixel array CAF surface from incoming data, wherein said CAF surface has at least one feature and at least one peak;

find at least one pixel representing said at least one peak on said CAF surface, wherein said peak has an amplitude greater than a background level;

frame a sub-window around said peak, wherein said sub-window includes said peak; and

apply a flattening function to said feature in said sub-window.

12. The system according to claim 11 , wherein said flattening function cancels at least one said peak.

13. The system according to claim 11 , wherein said flattening function is represented as: flattening function=(1−autocorrelation).

14. The system according to claim 11 , wherein said sub-window is sized so that pixel values along an edge of the autocorrelation function are about approximately zero.

15. The system according to claim 11 , wherein the software is configured to cause the processing unit to continue finding, framing, and applying until a condition is satisfied according to one of a group consisting of: constant false alarm processing; setting a margin level; and setting a customizable level.

16. A method for high resolution processing of a CAF surface, the method comprising:

Computing by a processing unit a CAF surface of an array of pixels at a desired over-sampling rate;

Locating by the processing unit of a peak on said CAF surface;

Forming by the processing unit of a sub-window about said peak;

multiplicatively removing by the processing unit of said peak using fractional delays; and

repeating by the processing unit of said steps of locating a next peak and multiplicatively removing each said next peak.

17. The method according to claim 16 , wherein said peak is calculated by at least one of a matched filter section and a maximum likelihood estimation (MLE) section.

18. The method according to claim 16 , wherein said CAF surface is computed at the desired over-sampling rate by CAF (w, T)=r H *S(w, T).

19. The method according to claim 16 , wherein a size of said sub-window is selected from a group consisting of: fixing a reference signal to a reference point; empirical testing so said sub-window extends past a coherence time and a Doppler extent; using a priori knowledge so said sub-window extends past a coherence time and a Doppler extent; warping said window to ensure it encompasses a sufficient region; and using a priori information about a canonical shape for some range of values.

20. The method according to claim 16 , wherein said sub-window is a range about the peak such that said sub-window includes sufficient pixels to represent an entire autocorrelation function.