Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
CALIBRATION, CLASSIFICATION AND LOCALIZATION USING CHANNEL TEMPLATES
Document Type and Number:
WIPO Patent Application WO/2020/037171
Kind Code:
A1
Abstract:
A method of parameter estimation in a multi-channel signal environment system wherein a plurality of receiving antennas receives signals or waves from targets due to one or more transmitters transmitting a predetermined signal that is reflected back from the targets or directly from one or more transmitters and then processed over multiple frequencies or channels by digital receiver. The method includes (a) comparing received voltages to an analytic or a table driven calibrated channel model without only relying on information from lossy intermediate steps such as time of arrival ("TOA") or angle of arrival ("AOA") measurements; and (b) mitigating channel model calibration errors, including multiplicative channel noise, phase noise, clutter or multipath modeling errors, by using a noise model to estimate away error nuisance parameters, either during a prior calibration process or during a real time calibration process concurrent with localization and parameter estimation during normal system operation.

Inventors:
BROMBERG MATTHEW C (US)
Application Number:
PCT/US2019/046741
Publication Date:
February 20, 2020
Filing Date:
August 16, 2019
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
MOVANO INC (US)
International Classes:
H04B17/391; H04B17/21; H04B17/309; H04J11/00
Domestic Patent References:
WO2004075577A12004-09-02
Foreign References:
US6571104B12003-05-27
US20100008406A12010-01-14
JPH11510981A1999-09-21
US20090042526A12009-02-12
Attorney, Agent or Firm:
SEQUEIRA, Antonia L. et al. (US)
Download PDF:
Claims:
1 claim:

1. A method of parameter estimation in a multi -channel signal environment system wherein a plurality of receiving antennas receives signals or waves from one or more targets due to one or more transmitters that transmit a predetermined signal that is reflected back from the targets or receives signals that are directly transmitted from one or more external transmitters to the receiving antennas and then processed over multiple frequencies or channels by a digital receiver connected to one or more processors, comprises the steps of:

comparing received voltages to an analytic or a table driven calibrated channel model without only relying on information from lossy intermediate steps such as time of arrival

(“TO A”) or angle of arrival (“AO A”) measurements; and

mitigating channel model calibration errors, including multiplicative channel noise, phase noise, clutter or multipath modeling errors, by using a noise model to estimate away error nuisance parameters, either during a prior calibration process or during a real time calibration process concurrent with localization and parameter estimation during normal system operation.

2. The method of parameter estimation in claim 1, wherein a Bayesian detection is used for the estimation of channel parameters .such as location parameters, shape parameters and reflector electromagnetic parameters.

3. The method of parameter estimation in claim 1, wherein the reflector signal from the target parameters is saved to be reused for dynamic tracking of the target.

4. The method of parameter estimation in claim 1, wherein a Bayesian particle filter is used to update mixture models for unknown parameters that might affect the reflected signal from the target.

5. The method of parameter estimation in claim 1 , wherein the predetermined transmitted signal is a frequency stepped radar.

6. The method of parameter estimation in claim L wherein the target is filtered by a delay range filter to minimize the clutter or focus in on known target parameter ranges.

7. lire method of parameter estimation in claim L wherein a channel is presumed to have multiplicative Gaussian noise or additive noise not otherwise modeled by the calibration process.

8. The method of parameter estimation in claim l, wherein multiple targets and the clutter are separated using multi-user detection techniques such as successive interference cancellation or parallel interference cancellation.

9. The method of parameter estimation in claim 1, wherein one or more targets and the clutter are separated using multi-user detection techniques such as successive interference cancellation or parallel interference cancellation in order to compute constant modulus phase normalization gains.

10. Hie method of parameter estimation in claim 1, wherein the target state is tracked by saving one or more aggregate channel estimates and their gains or whereby the channel itself is a statistical mixture or a slate machine requiring multiple discrete channel vectors.

1 . . The method of parameter estimation in claim I, wherein processing is triggered by a statistic that detects changes in the environment.

12. t he method of parameter estimation in claim 1, wherein received voltages ami calibration tables are whitened by the Hermitian inverse Cholesky factor of the measured or estimated interference covariance.

13. The method of parameter estimation in claim 1 , wherein the Bayesian parameter update is constrained to place a lower bound on the variance in an effort to allow tracking changes in a parameter vector.

14. The method of parameter estimation in claim 1, wherein the parameter model includes higher order dynamic parameters such as velocity or acceleration.

15. The method of parameter estimation in claim 1, wherein different shapes or reflector types are calibrated separately so that shape/type inference is possible in the Bayesian estimator.

16. The method of parameter estimation in claim 2, where the progression of shapes or locations are used to infer gesture types.

17. The method of parameter estimation in claim 16, wherein use is made of machine learning for the gesture determination.

18. The method of parameter estimation in claim 1 , wherein this invention makes use of a higher order singular value decomposition to compress a neural network into a single layer of opti m ized stun of products form.

19. The method of parameter estimation in claim 16. wherein the method decomposes the gesture recognition problem, into a set of discrete states whose transitions can be estimated using

Bayesian tet¼iique.s,

20. The method of parameter estimation in claim 1 , wherein the system interpolates a wave function using basis functions that are reflected through objects of predetermined shape and electrical properties.

21. The method of parameter estimation in claim 1 , wherein a reproducing kernel Is used to generate a model of the reflected signal or a waveform at the receive antennas.

22. The method of parameter estimation in claim 15, wherein the reproducing kernel is used to dynamically learn the shape of the reflector and/or its electrical parameters.

23. The method of parameter estimation in claim 21, wherein a series decomposition of the reproducing kernel is used to obtain basis functions of the reflected wave.

24. The method of parameter estimation in claim 18, wherein a higher order singular value decomposition is used to decompose a tabulated version of the reproducing kernel using far fewer singular values.

25. The method of parameter estimation in claim 1, wherein electromagnetic imaging techniques are used to find basis functions for the reflected signal or wave.

26, The method of parameter estimation in claim 1, wherein conformal mappings are used to find basis functions for the reflected wave.

27. The method of parameter estimation in claim 19, wherein a Cauchy-Kovalevskaya extension is used to predict the basis functions for the reflected wave.

28. The method of parameter estimation in claim 1, wherein the phase centers of an array of antennas are factored out for either removal or to reduce the number of bits required for accurate phase representation of the array.

29. The method of parameter estimation in claim L wherein a coupling matrix is applied to the output of the array to model mutual antenna coupling.

30. The method of parameter estimation in claim L wherein a Bicompiex Clifford algebra, that tracks both the complex pseudoscalar and the unit phasor a used as separate commuting complex units.

31. The method of parameter estimation in claim I, wherein a Vekua transform to transform a set of basis functions for the simpler Laplace equation into the basts function is used for the more general Helmholtz equation.

32. D method of parameter estimation in a multi-channel signal environment wherein a plurality of receiving antennas and/or a plurality of transmitters transmit a signal or wave that is known or estimated that is processed over multiple frequencies or channels by a digital receiver connected to one or more processors for processing the signal, comprising the steps of: comparing received antennas array voltages to an analytic or table driven channel model from a calibrated template without only relying on information from lossy intermediate steps such as time of delay or angle of arrival measurements to improve accuracy of reflected or emitter transmitted signals; and mitigating channel calibration errors from transmitted target signals by taking into account additive channel noise, multiplicative channel error, phase noise, multipath modeling errors by using the noise model to estimate away the error nuisance parameters either during a calibration process or during a calibration process in real time localization and parameter estimation during normal system operation.

Description:
Calibration, Classification and Localization Using

Channel Templates

Cross Reference to Related Inventions

[001] The present invention in a continuation in part of VS Application 16/541 ,730, and this application and the parent application claim priority to VS Provisional Application Serial Number 62/764,814 filed on August 16, 2(518, both of which are incorporated in their entirety by reference.

Background of the Invention

[002] This invention introduces additional details regarding calibration and geolocation techniques fust introduced in the prior referenced applications. This application focuses on simplified channel models and techniques that can exploit those models for geolocation and object classification.

Summary of the Invention

[003] In one emMoiment, there is provided a method of parameter estimation in a multi- channel signal environment where a plurality oi receive antennas and/or a plurality of transmitters transmit a signal that is known or estimated, that is processed over multiple frequencies or channels. by a digital receiver and one or more processors whose processing includes the following steps. First, an analytic or table driven channel model is used for comparing received voltages vs a calibrated template, without only relying on information lossy intermediate steps such as delay or angle measurements. A statistical likelihood function is used to model the receiver noise. channel parameters, or prior channel uncertainty. Then a Bayesian detection or other Statistical Signal Processing Techniques is used for the estimation of channel parameters such as location parameters, shape parameters, and reflector electromagnetic. parameters. A saving target, reflector . /emitter parameters is reused for dynamic tracking. Finally, Bayesian particle filtering or Maximum Likelihood Methods is used to update mixture models for the unknown parameters.

[004] Numerous other advantages and features of the invention will become readily apparent from the following detailed description of the invention and the embodiments thereof, from the claims, and from the accompanying drawings.

Brief Description of the Figures

[005j The patent or application file contains at least one drawing executed in color. Copies of t his patent or patent, application publication with color drawmg(s) will be provided by the Office upon request, and payment of the necessary fee. A fuller understanding of the foregoing may be had by r eference to the accompanying drawings, wherein:

[000] FIG. .1. is a radar transceiver;

[007] FIG. 2 is a flow diagram of background cancellation:

[008] FIG- 3 is a flow diagram of a calibration algorithm;

[009} FIG. 4 is a flow diagram of a blind calibration algorithm:

[010] FIG f > is a flow diagram of a target calibration: jOllf FIG. 6 is & flow digram of a classification alogotihm;

i012j FIG. 7 is a hardware overview;

[013] FIG. 8 is an image of an antenna board:

j014{ FIG. 9 is an architecture diagram of an SIMD processor in accordance with an embodiment of the invention; and

jOlOj FIG. 10 is an architecture diagram of a parallel reduction operation processor in accordance with an embodiment of the invention.

Description of the Invention

{010] While the invention is susceptible to embodiments in many different forms, there are shown in the drawings and will be described in detail herein the preferred embodiments of the present invention. It, should be understood, however, that the present, disclosure is to be considered fin exemplification of the principles of the invention and is not intended to limit the spirit or scope of the invention and/or claims of the embodiments illustrated

[017] Channel Model

j018| Consider a radar configuration of the form shown in Figure 1. We already shown on to modify the calibration signal model in the previous reference application At the receive antennas we assume a signal of the form, where the voltage seen by the m'th sensor. frequency /channel,

and the calibration collect., whore is a channel model

basis function for the m ! th sensor and k\h frequency /channel . s is an additive receiver

thermal noise term, and where is the“position” and shape parameter vector, for the </t.h calibration event. An unknown bulk complex scalar term b <! is presumed to model any rapidly changing, highly sensitive phase dependency for each collect·. It is often easier to maintain precise relative phases between antennae and frequencies, than bulk phases over time or position. The noise is assumed to be Gaussian, and can be assumed to be white, if pre-whitening is performed as needed.

[019] The channel voltages #* .. (p v ) will typically be observed after canceling a model of the stationary background data from the current received data. The stationary background is collected over quiescent periods of the transceiver device, prior to target entering the field of view. The canceller will be preferably of the least squares type, or a simple subtraction, and can generally be written as. where is the raw receiver data, q is the collect index, is the average background

voltage for antenna m and frequency index k, and where are some cancellation constants.

The operation is illustrated in Figure

[020} Ni M:e that the basis functions are typically chosen from geometric wave- functions, such as spheroidal wave functions, cylinder wave-functions and elliptical wave functions. These functions were previously described in the referenced applications. Some of the first few radial spheroidal wave functions are given by,

The actual basis functions used will be of the form where is the

wavenumber. is the distance from the transmitter to the target, and r-j is the distance from the target to a receiver. The terms containing associated withh the spherical Hankel

functions of the 2nd kind represent outgoing waves and are to be preferred. Similarly we can write the first few radial cylindrical wave functions as linear combinations of the Hankel functions and

{021} We can write Equation (.! ) in matrix fonn as, where v" is a vector of length M K, whose rn element is is an

matrix with the structure,

where,

mid where.

end

[022] The log-likelihood function for the Q calibration collects can be written as,

[023] Calibration

[024] The calibration process for our transceiver hind ware consists, of setting objects to be calibrated at known positions in a calibration grid. Also measurerl are the transmit, and

receiver antenna positions all for the purpose of evaluating the basis functions

at position p . The electromagnetic channel is sampled by the M receive antennas after transmitting a stepped frequency radar signal, for the radar application of this invention.

The algorithm is illustrated in Figure T

.··

[025] Mathematically the calibration problem involves solving for target coefficients, collected in a channel gain vector or in Equation (b) either using Maximum Likelihood (ML) estimation or using Bayesian estimation with a prior distribution presumed for or. However the likelihood function can also be used to refine the otherwise known positions of the transmit†. cr(s) and receivers. The likelihood function also serves as a starting point for the classificat ion and localisation problem itself during normal operation of the invention.

[025] Suppose o: is complex normally distributed with mean or turd covariance · The joint log likelihood can therefore be written as,

From this, given knowledge of B’> the Bayesian estimator for a is given by,

if however we need to remove the nuisance parameter we can find it’s maximum likelihood

(ML) estimator and write it as. ί<)2T| Using the ML estimator for in Bqnatkm(.i) wc obtain the likelihood function,

We can use this for Bayesian estimation presuming the empirical Bayes prior, wherein the prior for fi'* is the delta function centered on the Maximum Likelihood solution in Equation

(5)· Unfortunately Equation (v) can not be optimized in closed form over <¾ except in very important special cases. 'Fherefore for the general case we optimize using the following algorithm:

(028 j General Calibration Algorithm

(029 j 1. Place target of desired type in known location arid evaluate the basis functions ·

1030 ! 2. Capture the complex channel values due to reflected waveform from the targets at position

(031 i 3. Initialize or. to otherwise.

[032] 4. Compute the optimal using Equation

[033] 5. Compute the optimal a using Equation the number of iterations is less than ?¼ go to General Calibration Algorithm

Step otherwise stop.

[037] When 8 is Independent of q.

[038] An important special case that admits a closed form solution occurs when we presume that the unknown receiver gain is actually independent of the q collect number. This situat ion models the ease where the data collection over the q index has been performed within a short time period, or is automated and where tire target environment is mostly stationary, or where gain and phase variations are adequately modeled by the (p,,) basis functions. It is often the appropriate model to use after calibration, but during the parameter estimation phase of the device e.g. during geolocation or classification.

[039] This model yields a solution for 8 as.

If we ignore any priors for a,. and ignoring superfluous constants, and substituting in Equation (?) wo can write the likelihood function as,

This can be solved by finding the largest eigenvalue and associated vector of of the generalized eigen value problem.

[040] Collect Independent Algorithm

[041] 1. Place target of desired type in known location p !f and evaluate the basis functions

[042j 2. Capture the complex channel values due to reflected waveform from the targets at position

(t>43) 3. (Compute the optimal « by finding the eigenvector associated with the largest eigenvalue Equation

[044] Independent Basis Inner Product

(045) Titis special case occurs when I tuid is not dependent on q and thus independent of the parameter vector The primary application for this is when we have a

single phase only basis function, ie The ideal phase function for preferred use is given

by,

where is the transmit. frequency for index is the position of the transmit antenna

and x is the position of the receive antenna for index m.

|04<)j In this ease b* is solved by Equation and the likelihood function reduces to the

Rayleigh quotient,

This allows us to write our calibration algorithm as.

[011] Independent Basis Inner Product

[045] i. Place target of desired type in known location and evaluate the basis functions

[046] 2. Capture the complex channel values due to reflected waveform from the targets at position

[047] 3. Compute the optimal d by finding the eigenvector associated with the largest eigenvalue associated with the Rayleigh quotient

[048] Blind Calibration Algorithm

(049) For target classification, it may be accessary or desirable to perform the classification independently from any knowledge of the target position. We can support this by blindly estimating unknown basis function parameters or a sufficient statistic that depends on them. For the independent Basks inner product case of Independent Basis Inner Product Section we only require knowledge of the phase ramps for each antenna, or equivalently the round trip delay,

|05( ) | The blind algorithm is ii) nitrated in Figure 'Hie Basks Phase parameters replace the known parameters and are estimated blindly assuming known target coefficients ct * and using the Likelihood function fox each q for the p, } or r* optimization. The delays can be estimated by matched filter estimation using various hypothesized delays. This is facilitated by fast Fourier Transforms (FFT) for linearly sp iced frequencies / ¾ .

1051 j For the independent basis inner product algorithm we am write the likelihood from Equation (i .O as.

If is a unit pliasor per Equation (9), then, ignoring superfiuous constants,

we caxi write this as.

We can thus choose optimal delays by hypothesis testing by taking the FFT over the k index of for the case of linear frequency spacing /*.

[052j Blind Calibration Algorithm

[ 053] 1. Initialize ex to otherwise. j054[ 2. Capture the complex channel values due to reflected waveform from the

targets target coiled number </.

[055] 3. Evaluate the basis functions optimizing blindly the unknown depen

dency on p,. { or equivalently r* using Equation (3) or Equation (12) for each q.

j0¾G| 4. Compute the optimal d‘‘ using Equation (o).

[057] 5. Compute the optimal « using Equation ( 1)

[058{ 6. If the number of iterations is less than go to Blind Calibration Algorithm Step () otherwise stop.

|05<)| Classification and Localization

[060] In order to support target classification we need to calibrate each target category and obtain the channel coefficients <y. specific to category c, as shown in Figure · > . In general one desires to separately calibrate each type of target we desire to classify

(061 j Let ct c be the channel coefficients for category e. From Figure the category index might correspond to a human target, pet target or weapon target, though of course other target types can be envisioned. We can estimate all free parameters by using Bayesian inference and the likelihood functions provided in the Calibration Section, herein defined.

[062] Suppose we collect Q data vectors from our transceivers with v 4 the associated channel seen for collection q. Let the aggregate collection of the v 4 be written its V º

Defining,

from 2 we can also define the likelihood function.

where d is the optimal scalar dependent on the desired model chosen from the Calibration Section. if we choosy the maximum likelihood solution for /P, we can write the likelihood function as.

10(13] From Bayes theorem we can write,

where p,. is the optimal position parameter set for the category e. However it is possible to simply choose the phases blindly without proper array knowledge if you are primarily interested in classification instead of localization. The classification algorithm is shown graphically in Figure i>

[064] For the ease where we don’t have position or array information, we can classify blindly by learning the sufficient statistics fox the unknown position parameter by optimizing Equation! M) over p. This can be done using FFT operation» as suggested in the prior Section 'the blind classification algorithm is illustrated in Blind-Classification Algorithm.

[Oίίό j Hardware; Implementat ion

[066] One implementation of this invention is shown in the functional hardware block diagram in Figure 7. M antenna receivers are coupled with a transmit antenna. These are feci by an RF ASIC, which is modnlated by a phased locked loop and an oscillator. Both the transmit and receive oscillators are looked.

j(HV7; The return signal is downcon verted in the RF ASIC and then digitized by a bank of analog to digital converters, followed by a decimation filter where the data is buffered. The algorithms are implemented by DSP processors on board an ASIC device. A demonstration board containing 1 transmit antenna and 8 receive antennas is shown in Figure » 8.

|0C8| Digital Signal Processing (DSP) Architecture

j009j in order to implement our geoioeaiion algorithms, we need a processor architec lure that is able to compote multiple location hypotheses simultaneously This could be implemented in a multi-processor layout, but it would he more efficient to use either a single instruction multiple data (S1MD) or single instruction multiple thread (SIMT) architecture, since the same code is executed for each hypothesis. The problem is known as“embarrassingly parallel", since each likelihood computation can be performed without sharing much data between hypothesized positions.

[070] In one embodiment, we implement a SIMD architecture, with some changes made from traditional approaches to make it closer to a SIMT (Single instruction Multiple Thread) paradigm, used in heterogenous computational models such as Oil DA or OpeuCL. Consider the architecture shown in Figure 9. While for the most part every arithmetic-logic unit (ALU) can execute the same instruction on each clock cycle, we have to support more independence for each ALU than is traditionally allowed in most SLMD architectures.

[071] Independently Addressable Local Memory Cache

[072] For a variety of reasons we need to be able to us e either the A or B registers to address a local cache of memory unique to each processor. We need this for example to support the parallel computation of transcendental functions, either using independent look up tables and mult ipliers or look up tables and shifts for COR.D1C style algorithms.

[073] We also need local per ALU memory to support something like a parallel FliT, which could be facilitated by distributing the FFT input, with the appropriate decimation, over each independent ALU.

{0741 Word Shuffle

[075] We also implement some communication from one register to a neighboring ALUs dedicated register. For parallel reduction over an associative operator like addition or maximization, it suffices to simply be able to communicate with your neighboring ALU.

[076] Consider for example the structure shown in Figure 10. A reduction step allows oper ations using inputs from neighboring registers. In general we wish to add the operation where o is an associative operator, e.g addition multiplication, maximum or minimum. The output of the operator applied to register is mapped to the register A, t reducing the data registers in use by a factor of 2.

jOTTj To support the FFT operation we may need some more complicated operations depending on the type of FFT algorithm. For the decimation in time FFT we have the basic relationship,

where </ : Q : A/ is the Matlab notation for the .sequence that starts at. ¾, increases bv Q and ends at M or before. For the ease of Q— 2. we have the standard power of 2 recursion.

1078] 'Go support a decimation in tune FPT algorit hm we need to be able to distribute the input data to each local cache in a decimation order. Furthermore we need parallel reduction steps to combine the result or the FFT over each ALU section.

}C)79] For full generality some standard shuffle operations such as register swapping or circular shifts should also be supported.

[080] External IO

1081] We also implement the ability to pass data to all the memory cache’s and/or registers from an external source Each register A » , must be accessible from an external source, both for reading and writing, though this operation does not require any sort of multi-port access.

1082] Branching

[083] While we don’t need to support hilly independent branching, it will be necessary to support branching by globally looping over all branches until they are exhausted over all the A LU / Register files. We also need an instruction that obtains an ALU number, so that we can index and select based on which ALU/Register pair is in consideration. Additionally having instructions that, support a conditional move (similar to the C ? operator), would be helpful. [084] In accordance with various methods and embodiments outlined herein, there is provided in one embodiment a method of parameter estimation in a multi-channel signal environment system wherein a plurality of receiving antennas receives signals or waves from one or more targets due to one or more transmitters that transmit a predetermined signal that is reflected back from the targets or receives signals that are directly transmitted from one or more external transmitters to the receiving antennas and then processed over multiple frequencies or channels by a digital receiver connected to one or more processors. The method comprises steps including

(a) comparing received voltages to an analytic or a table driven calibrated channel model without only relying on information from lossy intermediate steps such as time of arrival fTOA") or angle of arrival (‘V\OA * ’) measurements; and (b) mitigating channel model calibration errors, including multiplicative channel noise, phase noise, clutter or multipath modeling errors, by using a noise model to estimate away error nuisance parameters, either during a prior calibration process or during a real time calibration process concurrent with localization and parameter estimation during normal system operation.

[085] The method may also use a Bayesian detection for the estimation of channel parameters such as location parameters, shape parameters and reflector electromagnetic parameters. A Bayesian particle Alter may also be used to update mixture models for unknown parameters that might affect the reflected signal from the target the Bayesian parameter update may also be constrained to place a lower bound on the variance in an effort to allow tracking changes in a parameter vector, it is further contemplated that wherein different shapes or reflector types are calibrated separately so that shape/type inference is possible in the Bayesian estimator.

[086] The method may also save the reflector signal from the target parameters to be reused for dynamic tracking of the target. In other aspect of the method the predetermined transmitted signal is a frequency stepped radar. In yet another objec t of the invention the target is filtered by a delay range filter to minimize the clutter or focus in on known target parameter ranges, in yet other embodiments, a channel is presumed to have multiplicative Gaussian noise or additive noise not otherwise modeled by the calibration process.

[087 j In yet further embodiments, multiple targets and the clutter are separated using multi-user detection techniques such as successive interference cancellation or parallel interference cancellation. In yet other embodiments, one or more targets and the clutter are separated using multi-user detection techniques such as successive interference cancellation or parallel interference cancellation in order to compute constant modulus phase normalization gains.

[088] The method may also track the target state by saving one or more aggregate channel estimates and their gains or whereby the channel itself is a statistical mixture or a state machine requiring multiple discrete channel vectors. The method may also trigger processing a statistic that detects changes in the environment. Other aspects of the embodiments may cause received voltages and calibration tables to be whitened by the Hermitian inverse Cholesky factor of the measured or estimated interference covariance.

[089] In addition, the parameter model includes higher order dynamic parameters such as velocity or acceleration, lire method may also use the progression of shapes or locations to infer gesture types. In yet further embodiments, the invention makes use of a higher order singular value decomposition to compress a neural network into a single layer of optimized sum of products form. The method may also decompose the gesture recognition problem, into a set of discrete states whose transitions can be estimated using Bayesian techniques.

[090] 'Hie method may also interpolate a wave function using basis functions that are reflected through objects of predetermined shape and electrical properties. As outlined herein, the method may also use a reproducing kernel to generate a model of the reflected signal or a waveform at the receive antennas. The reproducing kernel may be used to dynamically learn the shape of the reflector and/or its electrical parameters. A scries decomposition of the reproducing kernel may be used to obtain basis functions of the reflected wave. Alternatively, a higher order singular value decomposition may be used to decompose a tabulated version of the reproducing kernel using far fewer singular values.

[091] In yet further embodiments, electromagnetic imaging techniques are used to find basis functions for the reflected signal or wave. The method may also use conformal mappings to find basis functions for the reflected wave. A Cauchy-Kovalevskaya extension may be used to predict the basis functions for the reflected wave. A Vekua transform may be used to transform a set of basis functions tor the simpler Laplace equation into the basis function is used for the more general Helmholtz equation.

[092] The method may factor out the phase centers of an array of antennas for either removal or to reduce the number of bits required for accurate phase representation of the array. As further contemplated a coupling matrix is applied to the output of the array to model mutual antenna coupling, A Bicomplex Clifford algebra, that tracks both the complex pseudoscalar and the unit phasor may be used as separate commuting complex units.

(093 } From the foregoing and as mentioned above, it is observed that numerous variations and modifications may be affected without departing from the spirit and scope of the novel concept of the invention. It is to lx? understood that no limitation with respect to the embodiments illustrated herein is intended or should be inferred. It is intended to cover, by the appended claims, ah such modifications within the scope of the appended claims.