APPARATUS AND METHOD FOR PROVIDING ENERGY - BANDWIDTH TRADEOFF AND WAVEFORM DESIGN IN INTERFERENCE AND NOISE

Description

FIELD OF INVENTION

The invention relates to techniques related to a tradeoff between transmit signal Energy and its Bandwidth using new transmitter-receiver waveform design methods that are applicable for radar, sonar and wireless communications.

BACKGROUND OF INVENTION

In the general problem, a desired target is buried in both interference and noise. A transmit signal excites both the desired target and the interference simultaneously. The interference and/or interferences can be foliage returns in the form of clutter for radar, scattered returns of the transmit signal from a sea-bottom and different ocean-layers in the case of sonar, or multipath returns in a communication scene. In all of these cases, like the target return, the interference returns are also transmit signal dependent, and hence it puts conflicting demands on the receiver. In general, the receiver input is comprised of target returns, interferences and the ever present noise. The goal of the receiver is to enhance the target returns and simultaneously suppress both the interference and noise signals. In a detection environment, a decision regarding the presence or absence of a target is made at some specified instant ωt=t_ousing output data from a receiver, and hence to maximize detection, the Signal power to average Interference plus Noise Ratio (SINR) at the receiver output can be used as an optimization goal. This scheme is illustrated in FIG. 1.

The transmitter output bandwidth can be controlled using a known transmitter output filter having a transfer function P₁(ω) (see FIG. 2A). A similar filter with transform characteristics P₂(ω) can be used at a receiver input 22a shown in FIG. 1, to control the processing bandwidth as well.

The transmit waveform set f(t) at an output 10a of FIG. 1, can have spatial and temporal components to it each designated for a specific goal. The simplest situation is that shown in FIG. 2A where a finite duration waveform f(t) of energy E is to be designed. Thus

$\begin{matrix} \int_{0}^{T_{o}} {\langle f (t) \rangle}^{2} \partial t = E . & (1) \end{matrix}$

Usually, transmitter output filter 12 characteristics P₁(ω), such as shown in FIG. 2B, are known and for design purposes, it is best to incorporate the transmitter output filter 12 and the receiver input filter (which may be part of receiver 22) along with the target and clutter spectral characteristics.

Let q(t)Q(ω) represent the target impulse response and its transform. In general q(t) can be any arbitrary waveform. Thus the modified target that accounts for the target output filter has transform P₁(ω)Q(ω) etc. In a linear domain setup, the transmit signal f(t) interacts with the target q(t), or target 14 shown in FIG. 1, to generate the output below (referred to in S. U. Pillai, H. S. Oh, D. C. Youla, and J. R. Guerci, “Optimum Transmit-Receiver Design in the Presence of Signal-Dependent Interference and Channel Noise”, IEEE Transactions on Information Theory, Vol. 46, No. 2, pp. 577-584, March 2000 and J. R. Guerci and S. U. Pillai, “Theory and Application of Optimum Transmit-Receive Radar,” IEEE International Radar Conference, Alexandria Va., May 2000, pp. 705-710):

$\begin{matrix} s (t) = f (t) * q (t) = \int_{0}^{T_{o}} f (τ) q (t - τ) \partial τ & (2) \end{matrix}$

that represents the desired signal.

The interference returns are usually due to the random scattered returns of the transmit signal from the environment, and hence can be modeled as a stochastic signal w_c(t) that is excited by the transmit signal f(t). If the environment returns are stationary, then the interference can be represented by its power spectrum G_c(ω). This gives the average interference power to be G_c(ω)|F(ω)². Finally let n(t) represent the receiver 22 input noise with power spectral density G_n(ω). Thus the receiver input signal at input 22a equals

r(t)=s(t)+w_c(t)*f(t)+n(t), (3)

and the input interference plus noise power spectrum equals

G
₁(ω)=G_c(ω)|F(ω)|²+G_n(ω). (4)

The received signal is presented to the receiver 22 at input 22a with impulse response h(t). The simplest receiver is of the noncausal type.

With no restrictions on the receiver 22 of FIG. 1, its output signal at output 22b in FIG. 1, and interference noise components are given by

$\begin{matrix} y_{s} (t) = s (t) * h (t) = \frac{1}{2 π} \int_{- \infty}^{+ \infty} S (ω) H (ω) e^{jω t} \partial ω and & (5) \\ y_{n} (t) = {w_{c} (t) * f (t) + n (t)} * h (t) . & (6) \end{matrix}$

The output y_n(t) represents a second order stationary stochastic process with power spectrum below (referred to in the previous publications and in Athanasios Papoulis, S. Unnikrishna Pillai, Probability, Random Variables and Stochastic Processes, McGraw-Hill Higher Education, New York 2002):

G
_o(ω)=(G_c(ω)|F(ω)|²+G_n(ω)|H(ω)² (7)

and hence the total output interference plus noise power is given by

$\begin{matrix} \begin{matrix} σ_{I + N}^{2} = \frac{1}{2 π} \int_{- \infty}^{+ \infty} G_{0} (ω) \partial ω \\ = \frac{1}{2 π} \int_{- \infty}^{+ \infty} (G_{C} (ω) {\langle F (ω) \rangle}^{2} + G_{n} (ω)) {\langle H (ω) \rangle}^{2} \partial ω . \end{matrix} & (8) \end{matrix}$

Referring back to FIG. 1, the signal component y_s(t) in equation (5) at the receiver output 22b needs to be maximized at the decision instant to in presence of the above interference and noise. Hence the instantaneous output signal power at t=t_ois given by the formula below shown in S. U. Pillai, H. S. Oh, D. C. Youla, and J. R. Guerci, “Optimum Transmit-Receiver Design in the Presence of Signal-Dependent Interference and Channel Noise”, IEEE Transactions on Information Theory, Vol. 46, No. 2, pp. 577-584, March 2000, which is incorporated by reference herein:

$\begin{matrix} P_{O} = {\langle y_{s} (t_{o}) \rangle}^{2} = {\langle \frac{1}{2 π} \int_{- \infty}^{+ \infty} S (ω) H (ω) e^{jω t_{o}} \partial ω \rangle}^{2} . & (9) \end{matrix}$

This gives the receiver output SINR at t=t_obe the following as specified in Guerci et. al., “Theory and Application of Optimum Transmit-Receive Radar”, pp. 705-710; and Pillai et. al., “Optimum Transmit-Receiver Design in the Presence of Signal-Dependent Interference and Channel Noise”, incorporated herein by reference:

$\begin{matrix} SINR = \frac{P_{o}}{σ_{I + N}^{2}} = \frac{{\langle \frac{1}{2 π} \int_{- \infty}^{+ \infty} S (ω) H (ω) e^{jω t_{o}} \partial ω \rangle}^{2}}{\frac{1}{2 π} \int_{- \infty}^{+ \infty} G_{I} (ω) {\langle H (ω) \rangle}^{2} \partial ω} . & (10) \end{matrix}$

We can apply Cauchy-Schwarz inequality in equation (10) to eliminate H(ω). This gives

$\begin{matrix} \begin{matrix} SINR \leq \frac{1}{2 π} \int_{- \infty}^{+ \infty} \frac{{\langle S (ω) \rangle}^{2}}{G_{I} (ω)} \partial ω \\ = \frac{1}{2 π} \int_{- \infty}^{+ \infty} \frac{{\langle Q (ω) \rangle}^{2} {\langle F (ω) \rangle}^{2}}{G_{c} (ω) {\langle F (ω) \rangle}^{2} + G_{n} (ω)} \partial ω \\ = {SINR}_{\max} . \end{matrix} & (11) \end{matrix}$

Thus the maximum obtainable SINR is given by equation (11), and this is achieved if and only if the following equation referred to in previous prior art publications, is true:

$\begin{matrix} \begin{matrix} H_{opt} (ω) = \frac{S * (ω)}{G_{c} (ω) {\langle F (ω) \rangle}^{2} + G_{n} (ω)} e^{- jω t_{o}} \\ = \frac{Q * (ω) F * (ω)}{G_{c} (ω) {\langle F (ω) \rangle}^{2} + G_{n} (ω)} e^{- jω t_{o}} . \end{matrix} & (12) \end{matrix}$

In (12), the phase shift e^−jωt^ocan be retained to approximate causality for the receiver waveform. Interestingly even with a point target (Q(ω)≡1), flat noise (G_n(ω)=σ_n²), and flat clutter (G_c(ω)=σ_c²), the optimum receiver is not conjugate-matched to the transmit signal, since in that case from equation (12) we have the following formula given by Pillai et. al., “Optimum Transmit-Receiver Design in the Presence of Signal-Dependent Interference and Channel Noise”, incorporated herein by reference, Papoulis, “Probability, Random Variables and Stochastic Processes”, and H. L. Van Trees, Detection, Estimation, and Modulation Theory, Part I, New York: John Wiley and Sons, 1968, incorporated by reference:

$\begin{matrix} H_{opt} (ω) = \frac{F * (ω)}{σ_{c}^{2} {\langle F (ω) \rangle}^{2} + σ_{n}^{2}} e^{- jω t_{o}} \neq F * (ω) e^{- jω t_{o}} . & (13) \end{matrix}$

Prior Art Transmitter Waveform Design

When the receiver design satisfies equation (12), the output SINR is given by the right side of the equation (11), where the free parameter |F(ω)|²can be chosen to further maximize the output SINR, subject to the transmit energy constraint in (1). Thus the transmit signal design reduces to the following optimization problem:

Maximize

$\begin{matrix} {SINR}_{\max} = \frac{1}{2 π} \int_{- \infty}^{+ \infty} \frac{{\langle Q (ω) \rangle}^{2} {\langle F (ω) \rangle}^{2}}{G_{c} (ω) {\langle F (ω) \rangle}^{2} + G_{n} (ω)} \partial ω, & (14) \end{matrix}$

subject to the energy constraint

$\begin{matrix} \int_{0}^{T_{o}} {\langle f (t) \rangle}^{2} \partial t = \frac{1}{2 π} \int_{- \infty}^{+ \infty} {\langle F (ω) \rangle}^{2} \partial ω = E . & (15) \end{matrix}$

To solve this new constrained optimization problem, combine (14)-(15) to define the modified Lagrange optimization function (referred to in T. Kooij, “Optimum Signal in Noise and Reverberation”, Proceeding of the NATO Advanced Study Institute on Signal Processing with Emphasis on Underwater Acoustics, Vol. I, Enschede, The Netherlands, 1968.)

$\begin{matrix} Λ = \int_{- \infty}^{+ \infty} {\frac{{\langle Q (ω) \rangle}^{2} y^{2} (ω)}{G_{c} (ω) y^{2} (ω) + G_{n} (ω)} - \frac{1}{λ^{2}} y^{2} (ω)} \partial ω where & (16) \\ y (ω) = \langle F (ω) \rangle & (17) \end{matrix}$

is the free design parameter. From (16) (17),

$\frac{\partial Λ}{\partial y} = 0 gives$

gives

$\begin{matrix} \frac{\partial Λ (ω)}{\partial y} = 2 y (ω) {\frac{G_{n} (ω) {\langle Q (ω) \rangle}^{2}}{{G_{c} (ω) y^{2} - (ω) - G_{n} (ω)}^{2}} - \frac{1}{λ^{2}}} = 0. & (18) \end{matrix}$

where Λ(ω) represents the quantity within the integral in (16). From (18), either

$\begin{matrix} y (ω) = 0 & (19) \\ or \\ \frac{G_{n} (ω) {\langle Q (ω) \rangle}^{2}}{{G_{c} (ω) y^{2} (ω) + G_{n} (ω)}^{2}} - \frac{1}{λ^{2}} = 0, & (20) \end{matrix}$

which gives

$\begin{matrix} y^{2} (ω) - \frac{\sqrt{G_{n} (ω)} (λ \langle Q (ω) \rangle - \sqrt{G_{n} (ω)})}{G_{c} (ω)} & (21) \end{matrix}$

provided y²(ω)>0. See T. Kooij cited above incorporated by reference herein.

SUMMARY OF THE INVENTION

One or more embodiments of the present invention provide a method and an apparatus for transmitter-receiver design that enhances the desired signal output from the receiver while minimizing the total interference and noise output at the desired decision making instant. Further the method and apparatus of an embodiment of the present invention can be used for transmit signal energy-bandwidth tradeoff. As a result, transmit signal energy can be used to tradeoff for “premium” signal bandwidth without sacrificing performance level in terms of the output Signal to Interference plus Noise power Ratio (SINR). The two designs—before and after the tradeoff—will result in two different transmitter-receiver pairs that have the same performance level. Thus a design that uses a certain energy and bandwidth can be traded off with a new design that uses more energy and lesser bandwidth compared to the old design. In many applications such as in telecommunications, since the available bandwidth is at premium, such a tradeoff will result in releasing otherwise unavailable bandwidth at the expense of additional signal energy. The bandwidth so released can be used for other applications or to add additional telecommunications capacity.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 shows a diagram of a system, apparatus, and/or method including a transmitter, a transmitter output filter, a receiver, a target, interference, noise, and a switch;

FIG. 2A shows a prior art graph of a prior art transmitter signal versus time, wherein the transmitter signal is output from a transmitter, such as in FIG. 1;

FIG. 2B shows a prior art graph of a possible frequency spectrum of a known transmitter output filter, such as in FIG. 1;

FIG. 3A shows a graph of target transfer function magnitude response versus frequency;

FIG. 3B shows a graph of target transfer function magnitude response versus frequency;

FIG. 3C shows a graph of noise power spectrum divided by target transfer function magnitude response versus frequency;

FIG. 3D shows a graph of noise power spectrum divided by target transfer function magnitude response versus frequency;

FIG. 4A shows graphs of three different target transfer function magnitude responses versus frequency;

FIG. 4B shows a graph of noise power spectrum versus frequency;

FIG. 4C shows a graph of clutter power spectrum versus frequency;

FIG. 4D shows a graph of noise power spectrum divided by target transfer function magnitude response versus frequency;

FIG. 4E shows a graph of transmitter threshold energy versus bandwidth;

FIG. 4F shows a graph of signal to inference plus noise ratio (SINR) versus bandwidth;

FIG. 5A shows graphs of three different target transfer function magnitude responses versus frequency;

FIG. 5B shows a graph of noise power spectrum versus frequency;

FIG. 5C shows a graph of clutter power spectrum versus frequency;

FIG. 5D shows a graph of noise power spectrum divided by target transfer function magnitude response versus frequency;

FIG. 5E shows a graph of transmitter threshold energy versus bandwidth;

FIG. 5F shows a graph of signal to inference plus noise ratio (SINR) versus bandwidth;

FIG. 6A shows a graph of signal to interference plus noise ratio versus energy for a resonant target shown in FIG. 5A (solid line);

FIG. 6B shows a graph of signal to interference plus noise ratio versus energy for a low pass target shown in FIG. 5A (dashed line);

FIG. 6C shows a graph of signal to interference plus noise ratio versus energy for a flat target shown in FIG. 5A (dotted line);

FIG. 7 shows a graph of signal to interference plus noise ratio versus energy and the Bandwidth-Energy swapping design;

FIG. 8A shows a graph of the transform of the transmitter signal versus frequency corresponding to the design point A in FIG. 7;

FIG. 8B shows a graph of the transform of the transmitter signal versus frequency corresponding to the design point B in FIG. 7;

FIG. 8C shows a graph of the transform of the transmitter signal versus frequency corresponding to the design point C in FIG. 7; and

FIG. 9 is a graph of realizable bandwidth savings versus operating bandwidth.

DETAILED DESCRIPTION OF THE DRAWINGS

FIG. 1 shows a diagram of a system, apparatus, and/or method 1, including a transmitter 10, a transmitter output filter 12, a target 14, interference 16, noise 18, a summation block 20, receiver 22, and a switch 24. The present invention, in one or more embodiments, provides a new method and apparatus, by selecting a particular transmit signal f(t), to be output from transmitter 10, and a type of receiver or receiver transfer function for receiver 22 in accordance with criteria to be discussed below.

The transmitter 10 transmits an output signal f(t) at its output 10a and supplies this signal to the transmitter output filter 12. As remarked earlier, for design purposes, the transmitter output filter 12 can be lumped together with the target transfer function as well as the interference spectrum. The transmit signal f(t) passes through the airwaves and interacts with a target 14 and interference 16. The target-modified as well as the clutter-modified (or interference modified) versions of the transmit signal f(t) are supplied to the summation block 18 along with receiver noise 18. The summation block 18 may simply be used for description purposes to indicate that the target modified, clutter modified, and noise signals combine together. A combination signal is supplied to receiver 22 at its input 22a. The receiver 22 applies a transfer function H(ω) (which will be determined and/or selected by criteria of an embodiment of the present invention, to be described below) and a modified combination signal is provided at a receiver output 22b. The output is accessed at time t=t₀by use of switch 24.

FIG. 2A shows a prior art graph of a prior art transmitter output signal f(t) versus time. The signal used here is arbitrary.

FIG. 2B shows a prior art graph of a frequency spectrum of the transmitter output filter 12 of FIG. 1.

FIG. 3A shows a typical graph of a target transfer function magnitude response for target 14 versus frequency; target as appearing in (14)-(21).

FIG. 3B shows a typical graph of target transfer function magnitude response for target 14 versus frequency; target as appearing in (14)-(21).

FIG. 3C shows a graph of noise power spectrum divided by target transfer function magnitude response versus frequency; as in right side of equation (23).

FIG. 3D shows a graph of noise power spectrum divided by target transfer function magnitude response versus frequency; as in right side of equation (23).

FIG. 4A shows graphs of three different target transfer function magnitude responses versus frequency; target as appearing in (14)-(21).

FIG. 4B shows a graph of noise power spectrum versus frequency as appearing in equations (14)-(23).

FIG. 4C shows a graph of clutter power spectrum versus frequency as appearing in equations (14)-(23).

FIG. 4D shows a graph of noise power spectrum divided by target transfer function magnitude response versus frequency as in right side of equation (23).

FIG. 4E shows a graph of transmitter threshold energy versus bandwidth using equation (26).

FIG. 4F shows a graph of signal to inference plus noise ratio versus bandwidth using equations (27)-(31).

FIG. 5A shows graphs of three different target transfer function magnitude responses versus frequency; target as appearing in (14)-(21).

FIG. 5B shows a graph of noise power spectrum versus frequency as appearing in equations (14)-(23).

FIG. 5C shows a graph of clutter power spectrum versus frequency as appearing in equations (14)-(23).

FIG. 5D shows a graph of noise power spectrum divided by target transfer function magnitude response versus frequency as in right side of equation (23).

FIG. 5E shows a graph of transmitter threshold energy versus bandwidth using equation (26).

FIG. 5F shows a graph of signal to inference plus noise ratio versus bandwidth using equations (27)-(31).

FIG. 6A shows a graph of signal to interference plus noise ratio versus energy for a resonant target shown in FIG. 5A (solid line) using equations (34)-(35).

FIG. 6B shows a graph of signal to interference plus noise ratio versus energy for a low pass target shown in FIG. 5A (dashed line) using equations (34)-(35).

FIG. 6C shows a graph of signal to interference plus noise ratio versus energy for a flat target shown in FIG. 5A (dotted line) using equations (34)-(35).

FIG. 7 shows a graph of signal to interference plus noise ratio versus energy; generated using equations (39), (48), and (51).

FIG. 8A shows a graph of the transform of the transmitter signal versus frequency corresponding to the design point A in FIG. 7 generated using (42).

FIG. 8B shows a graph of the transform of the transmitter signal versus frequency corresponding to the design point B in FIG. 7 generated using (42).

FIG. 8C shows a graph of the transform of the transmitter signal versus frequency corresponding to the design point C in FIG. 7 generated using (42) for a third energy condition.

FIG. 9 is a realizable bandwidth savings versus operating bandwidth generated using equation (60).

Define Ω₊ as shown in FIGS. 3C and 3D to represent the frequencies over which y²(ω) in equation (21) is strictly positive, and let ω_orepresent the complement of Ω₊. As shown in FIGS. 3C and 3D, observe that the set Ω₊ is a function of the noise and target spectral characteristics as well as the constraint constant λ. In terms of Ω₊, we have

$\begin{matrix} {\langle F (ω) \rangle}^{2} = {\begin{matrix} y^{2} (ω), & ω \in Ω_{+} \\ 0, & ω \in Ω_{o} . \end{matrix} & (22) \end{matrix}$

From (21), y²(ω)>0 over Ω₊ gives the necessary condition

$\begin{matrix} λ \geq \max_{ω \in Ω_{+}} \frac{\sqrt{G_{n} (ω)}}{\langle Q (ω) \rangle} & (23) \end{matrix}$

and the energy constraint in (15) when applied to (21) gives

$\begin{matrix} E = \frac{1}{2 π} \int_{Ω_{+}} y^{2} (ω) \partial ω = \frac{λ}{2 π} \int_{Ω_{+}} \frac{\sqrt{G_{n} (ω)} \langle Q (ω) \rangle}{G_{c} (ω)} \partial π - \frac{1}{2 π} \int_{Ω_{+}} \frac{G_{n} (ω)}{G_{c} (ω)} \partial ω & (24) \end{matrix}$

or, for a given value of E, we have

$\begin{matrix} λ = \frac{E + \frac{1}{2 π} \int_{Ω_{+}} \frac{G_{n} (ω)}{G_{c} (ω)} \partial ω}{\frac{1}{2 π} \int_{Ω_{+}} \frac{\sqrt{G_{n} (ω)} \langle Q (ω) \rangle}{G_{c} (ω)} \partial ω} •λ (E) . & (25) \end{matrix}$

Clearly, λ(E) in (25) must satisfy the inequality in (23) as well. This gives rise to the concept of transmitter energy threshold that is characteristic to this design approach.

Transmitter Threshold Energy

From (23)-(25), the transmit energy E must be such that λ(E) obtained from (25) should satisfy (23). If not, E must be increased to accommodate it, and hence it follows that there exists a minimum threshold value for the transmit energy below which it will not be possible to maintain |F(ω)|²>0. This threshold value is given by

$\begin{matrix} E_{\min} = (\max_{ω \in Ω_{-}} \frac{\sqrt{G_{n} (ω)}}{\langle Q (ω) \rangle}) \frac{1}{2 π} \int_{Ω_{+}} \frac{\sqrt{G_{n} (ω)} \langle Q (ω) \rangle}{G_{c} (ω)} \partial ω - \frac{1}{2 π} \int_{Ω_{+}} \frac{G_{n} (ω)}{G_{c} (ω)} \partial ω & (26) \end{matrix}$

and for any operating condition, the transmit energy E must exceed E_min. Clearly, the minimum threshold energy depends on the target, clutter and noise characteristics as well as the bandwidth under consideration. With E>E_min, substituting (20)-(21) into the SINR_maxin (14) we get

$\begin{matrix} \begin{matrix} {SINR}_{\max} = \frac{1}{2 π} \int_{Ω_{+}} \frac{{\langle Q (ω) \rangle}^{2} y^{2} (ω)}{λ \sqrt{G_{n} (ω)} \langle Q (ω) \rangle} \partial ω \\ = \frac{1}{2 π} \int_{Ω_{-}} \frac{\langle Q (ω) \rangle}{λ (E) \sqrt{G_{n} (ω)}} \\ \frac{\sqrt{G_{n} (ω)} (λ (E) \langle Q (ω) \rangle - \sqrt{G_{n} (ω)})}{G_{c} (ω)} \partial ω \\ = \frac{1}{2 π} \int_{Ω_{-}} (\langle Q (ω) \rangle - \frac{\sqrt{G_{n} (ω)}}{λ (E)}) \frac{\langle Q (ω) \rangle}{G_{c} (ω)} \partial ω . \end{matrix} & (27) \end{matrix}$

Finally making use of (25), the output SINR_maxbecomes

$\begin{matrix} \begin{matrix} {SINR}_{1} = \frac{1}{2 π} \int_{Ω_{+}} \frac{{\langle Q (ω) \rangle}^{2}}{G_{c} (ω)} \partial ω - \frac{{(\frac{1}{2 π} \int_{Ω_{+}} \frac{\langle Q (ω) \rangle \sqrt{G_{n} (ω)}}{G_{c} (ω)} \partial ω)}^{2}}{E + \frac{1}{2 π} \int_{Ω_{+}} \frac{G_{n} (ω)}{G_{c} (ω)} \partial ω} \\ = a - \frac{c}{λ (E)} \\ = a - \frac{c^{2}}{E + b} \\ = \frac{aE + (ab - c^{2})}{E + b} \end{matrix} & (28) \\ where \\ a = \frac{1}{2 π} \int_{Ω_{+}} \frac{{\langle Q (ω) \rangle}^{2}}{G_{c} (ω)} \partial ω, & (29) \\ b = \frac{1}{2 π} \int_{Ω_{+}} \frac{G_{n} (ω)}{G_{c} (ω)} \partial ω, & (30) \\ and \\ c = \frac{1}{2 π} \int_{Ω_{+}} \frac{\sqrt{G_{n} (ω)} \langle Q (ω) \rangle}{G_{c} (ω)} \partial ω . & (31) \end{matrix}$

Notice that ab−c²≧0. (This was published in Waveform Diversity and Design conference, Kauai, Hi., January 2006).

The optimization problem in (14)-(15) can be restated in term of Ω₊ as follows: Given Q(ω), G_c(ω), G_n(ω) and the transmit energy E, how to partition the frequency axis into an “operating band” Ω₊ and a “no show” band Ω_oso that λ₊ obtained from (25) satisfies (23) and SINR_maxin (27)-(28) is also maximized. In general maximization of SINR_maxin (27)-(28) over Ω₊ is a highly nonlinear optimization problem for arbitrary Q(ω), G_c(ω) and G_n(ω).

In what follows a new approach to this problem is presented.

An Embodiment of the Present Invention—Desired Band Approach

One approach in this situation is to make use of the “desired frequency band” of interest B_othis is usually suggested by the target response Q(ω) (and the transmitter output filter) to determine the operating band Ω₊. The desired band B_ocan represent a fraction of the total available bandwidth, or the whole bandwidth itself. The procedure for determining Ω₊ is illustrated in FIGS. 3A-3C and FIGS. 3B-3D for two different situations. In FIGS. 3A-3D, the frequency band B_orepresents the desired band, and because of the nature of the noise and clutter spectra, it may be necessary to operate on a larger region Ω₊ in the frequency domain. Thus the desired band B_ois contained always within the operating band Ω₊. To determine Ω₊, using equation (23) we project the band B_oonto the spectrum √{square root over (G_n(ω))}/|Q(ω)| and draw a horizontal line corresponding to

$\begin{matrix} λ_{B_{o}} = \max_{ω \in B_{o}} \frac{\sqrt{G_{n} (ω)}}{\langle Q (ω) \rangle} & (32) \end{matrix}$

as shown there. Define Ω₊(B_o) to represent the frequency region where

$\begin{matrix} ω \in Ω_{+} (B_{o}) : \frac{\sqrt{G_{n} (ω)}}{\langle Q (ω) \rangle} \leq λ_{B_{o}} = \max_{ω \in B_{o}} \frac{\sqrt{G_{n} (ω)}}{\langle Q (ω) \rangle} . & (33) \end{matrix}$

This procedure can give rise to two situations as shown in FIG. 3A and FIG. 3B. In FIG. 3A, the operating band Ω₊(B_o) coincides with the desired band B_oas shown in FIG. 3C, whereas in FIG. 3B, the desired band B_ois a subset of Ω₊(B_o) as seen from FIG. 3D.

Knowing Ω₊(B_o), one can compute λ=λ(E) with the help of equation (25) over that region, and examine whether A so obtained satisfies (23). If not, the transmitter energy E is insufficient to maintain the operating band Ω₊(B_o) given in (33), and either E must be increased, or Ω₊(B_o) must be decreased (by decreasing B_o) so that (23) is satisfied. Thus for a given desired band B_o(or an operating band Ω₊(B_o)), as remarked earlier, there exists a minimum transmitter threshold energy E_B_o, below which it is impossible to maintain |F(ω)²>0 over that entire operating band.

Threshold Energy

From equations (24) and (32), we obtain the minimum transmitter threshold energy in this case to be the following

$\begin{matrix} \begin{matrix} E_{B_{o}} = \frac{λ_{B_{o}}}{2 π} \int_{Ω_{+} (B_{o})} \frac{\sqrt{G_{n} (ω)} \langle Q (ω) \rangle}{G_{c} (ω)} \partial ω - \\ \frac{1}{2 π} \int_{Ω_{-} (B_{o})} \frac{G_{n} (ω)}{G_{c} (ω)} \partial ω \\ = λ_{B_{o}} c_{o} - b_{o} > 0. \end{matrix} & (34) \end{matrix}$

With E≧E_B_o, the SINR_maxa in (28) can be readily computed. In particular with E=E_B_o, we get

$\begin{matrix} {SINR}_{1} = {SINR}_{1} (B_{o}) = a_{o} - \frac{c_{o}^{2}}{E_{B_{o}} + b_{o}} . & (35) \end{matrix}$

Here a_o, b_oand c_oare as given in (29)-(31) with Ω₊ replaced by Ω₊(B_o). Eq. (35) represents the performance level for bandwidth B_ousing its minimum threshold energy. From (21), we also obtain the optimum transmit signal transform corresponding to energy E_B_oto be

$\begin{matrix} \begin{matrix} {\langle F (ω) \rangle}^{2} = {\begin{matrix} \frac{\sqrt{G_{n} (ω)} (λ_{B_{o}} \langle Q (ω) \rangle - \sqrt{G_{n} (ω)})}{G_{c} (ω)}, & ω \in Ω_{+} (B_{o}) \\ 0, & ω \in Ω_{o} \end{matrix} \\ = {\begin{matrix} \sqrt{G_{n} (ω)} (\begin{matrix} (\max_{ω \in B_{o}} \frac{\sqrt{G_{n} (ω)}}{\langle Q (ω) \rangle}) - \\ \frac{\sqrt{G_{n} (ω)}}{\langle Q (ω) \rangle} \end{matrix}) \frac{\langle Q (ω) \rangle}{G_{c} (ω)}, & ω \in Ω_{+} (B_{o}) \\ 0, & ω \in Ω_{o} \end{matrix} . \end{matrix} & (36) \end{matrix}$

To summarize, to maintain a given desired band B_o, there exists an operating band Ω₊(B_o)≧B_oover which |F(ω)|²<0 and to guarantee this, the transmit energy must exceed a minimum threshold value E_B_ogiven by (34).

FIGS. 4A-F shows the transmitter threshold energy E in (34) and the corresponding SINR in (35) as a function of the desired bandwidth B_ofor various target, clutter, and noise spectra. Target to noise ratio (TNR) is set at 0 dB, and the clutter to noise power ratio (CNR) is set at 20 dB here. The total noise power is normalized to unity. The desired bandwidth B_ois normalized with respect to the maximum available bandwidth (e.g., carrier frequency).

In FIGS. 4A-F, the noise and clutter have flat spectra and for the highly resonant target (solid line), the required minimum energy threshold and the SINR generated using (34)-(35) reach a saturation value for small values of the bandwidth. In the case of the other two targets, additional bandwidth is required to reach the maximum attainable SINR. This is not surprising since for the resonant target, a significant portion of its energy is concentrated around the resonant frequency. Hence once the transmit signal bandwidth reaches the resonant frequency, it latches onto the target features resulting in maximum SINR at a lower bandwidth.

FIGS. 5A-F show results for a new set of clutter and noise spectra as shown there; the transmitter threshold energy E in (34) and the corresponding SINR in (35) as a function of the desired bandwidth B_oshow similar performance details.

From FIG. 5F, in the case of the resonant target (solid curve) the SINR reaches its peak value resulting in saturation even when B_ois a small fraction of the available bandwidth. This is because in that case, the transmit waveform is able to latch onto the dominant resonant frequency of the target. On the other extreme, when the target has flat characteristics (dotted curve), there are no distinguishing frequencies to latch on, and the transmitter is unable to attain the above maximum SINR even when B_ocoincides with the total available bandwidth. For a low pass target (dashed curve), the transmitter is indeed able to deliver the maximum SINR by making use of all the available bandwidth.

As FIG. 3B shows, Ω₊(B_o) can consist of multiple disjoint frequency bands whose complement Ω_orepresents the “no show” region. Notice that the “no show” region Ω_oin the frequency domain in (36) for the optimum transmit signal can be controlled by the transmit energy E in (25). By increasing E, these “no show” regions can be made narrower and this defines a minimum transmitter threshold energy E_∞ that allows Ω₊(B_o) to be the entire available frequency axis. To determine E_∞, let λ_∞ represent the maximum in (23) over the entire frequency axis. Thus

$\begin{matrix} λ_{\infty} = \max_{\langle ω \rangle < \infty} \frac{\sqrt{G_{n} (ω)}}{\langle Q (ω) \rangle}, & (37) \end{matrix}$

and let a_∞, b_∞, c_∞ refer to the constants a, b, c in (29)-(31) calculated with Ω₊ representing the entire frequency axis. Then from (24)

$\begin{matrix} E_{\infty} = λ_{\infty} c_{\infty} - b_{\infty} = \frac{λ_{\infty}}{2 π} \int_{- \infty}^{+ \infty} \frac{\sqrt{G_{n} (ω)} \langle Q (ω) \rangle}{G_{c} (ω)} \partial ω - \frac{1}{2 π} \int_{- \infty}^{+ \infty} \frac{G_{n} (ω)}{G_{c} (ω)} \partial ω > 0 & (38) \end{matrix}$

represents the minimum transmit energy (threshold) required to avoid partitioning in the frequency domain. With E_∞ as given by (38), we obtain SINR_maxto be (use (28))

$\begin{matrix} {SINR}_{1} (\infty) = a_{\infty} - \frac{c_{\infty}}{λ_{\infty}} = a_{\infty} - \frac{c_{\infty}^{2}}{E_{\infty} + b_{\infty}} > 0 & (39) \\ and \\ {\langle F (ω) \rangle}^{2} = \frac{\sqrt{G_{n} (ω)} (λ_{\infty} \langle Q (ω) \rangle - \sqrt{G_{n} (ω)})}{G_{c} (ω)}, \langle ω \rangle < \infty . & (40) \end{matrix}$

Clearly by further increasing the transmit energy in (39) beyond that in (38) we obtain

$\begin{matrix} {SINR}_{1} -> a_{ \inf} = \frac{1}{2 π} \int_{- \infty}^{- \infty} \frac{{\langle Q (ω) \rangle}^{2}}{G_{c} (ω)} \partial ω . & (41) \end{matrix}$

It follows that to avoid any restrictions in the frequency domain for the transmit signal, the transmitter energy E must exceed a minimum threshold value E_∞ given by (38) and (39) represents the maximum realizable SNR. By increasing E beyond E_∞, the performance can be improved upto that in (41).

In general from (34) for a given desired bandwidth B_o, the transmit energy E must exceed its threshold value E_B_o. With E>E_B_oand λ(E) as in (25), the corresponding optimum transmit signal transform is given by (see (21) (22))

$\begin{matrix} {\langle F (ω) \rangle}^{2} = {\begin{matrix} \frac{\sqrt{G_{c} (ω)} (λ (E) \langle (ω) \rangle - \sqrt{G_{n} (ω)})}{G_{c} (ω)}, & ω \in Ω_{+} (B_{o}) \\ 0, & ω \in Ω_{o} \end{matrix} & (42) \end{matrix}$

and clearly this signal is different from the minimum threshold energy one in (36). From (28), the performance level SINR₁(E,B_o) corresponding to (42) is given by (35) with E_B_oreplaced by E. Thus

$\begin{matrix} {SINR}_{1} (E, B_{o}) = a_{o} - \frac{c_{o}^{2}}{E + b_{o}} > {SINR}_{1} (B_{o}) . & (43) \end{matrix}$

From (43), for a given bandwidth B_o, performance can be increased beyond that in (35) by increasing the transmit energy. Hence it follows that SINR₁(B_o) represents the minimum performance level for bandwidth B_othat is obtained by using its minimum threshold energy. It is quite possible that this improved performance SINR₁(E,B_o) can be equal to the minimum performance level corresponding to a higher bandwidth B₁>B_o. This gives rise to the concept of Energy-Bandwidth tradeoff at a certain performance level. Undoubtedly this is quite useful when bandwidth is at premium.

FIGS. 5E-5F exhibit the transmit threshold energy and maximum output SINR₁(B_o) as a function of the desired bandwidth B_o. Combining these figures using (35), an SINR vs. transmit threshold energy plot can be generated as in FIGS. 6A-C for each target situation.

For example, FIG. 6A-C corresponds to the three different target situations considered in FIG. 5 with clutter and noise spectra as shown there. Notice that each point on the SINR-Energy threshold curve for each target is associated with a specific desired bandwidth. Thus for bandwidth B₁, the minimum threshold energy required is E₁, and the corresponding SINR equals SINR₁(B₁) in (35). Let A represent the associated operating point in FIG. 6. Note that the operating point A corresponding to a bandwidth B₁has different threshold energies and different performance levels for different targets. From (35), each operating point generates a distinct transmit waveform. As the bandwidth increases, from (39), SINR→SINR₁(∞).

Monotonic Property of SINR

The threshold energy and SINR associated with a higher bandwidth is higher. To prove this, consider two desired bandwidths B₁and B₂with B₂>B₁. Then from (32) we have

$\begin{matrix} λ_{2} = \max_{ω \in B_{2}} \frac{\sqrt{G_{n} (ω)}}{\langle Q (ω) \rangle} > λ_{1} = \max_{ω \in B_{1}} \frac{\sqrt{G_{n} (ω)}}{\langle Q (ω) \rangle}, & (44) \end{matrix}$

and from FIG. 3, the corresponding operating bandwidths Ω₊(B_o) and Ω₊(B₂) satisfy

Ω₊(B₂)≧Ω₊(B₁). (45)

From (34) (or (24)), the minimum threshold energies are given by

$\begin{matrix} E_{i} = \frac{1}{2 π} \int_{Ω_{+} (B_{i})} \sqrt{G_{n} (ω)} (λ_{i} - \frac{\sqrt{G_{n} (ω)}}{Q (ω)}) \frac{\langle Q (ω) \rangle}{G_{c} (ω)} \partial ω, i = 1, 2 & (46) \end{matrix}$

and substituting (44) and (45) into (46) we get

E₂>E₁. (47)

Also from (27), the performance levels at threshold energy SINR₁(B_i) equals

$\begin{matrix} {SINR}_{1} (B_{i}) = \frac{1}{2 π} \int_{Ω_{+} (B_{i})} (\langle Q (ω) \rangle - \frac{\sqrt{G_{n} (ω)}}{λ_{i}}) \frac{\langle Q (ω) \rangle}{G_{c} (ω)} \partial ω & (48) \end{matrix}$

and an argument similar to (44)-(45) gives

SINR
₁(B₂)≧SINR₁(B₁) (49)

for B₂>B₁. Thus as FIGS. 5A-F-FIGS. 6A-C show, SINR₁(B_i) is a monotonically nondecreasing function of both bandwidth and energy. FIG. 7 illustrates this SINR-energy relation for the target with flat spectrum shown in FIG. 5A. In FIG. 7, the two operating points A and B are associated with bandwidths B₁and B₂, threshold energies E₁and E₂, and performance levels SINR₁(B₁) and SINR₁(B₂) respectively. Since

B₂>B₁E₂≧E₁and SINR₁(B₂)≧SINR₁(B₁). (50)

The distinct transmit waveforms |F₁(ω)|²and |F₂(ω)|²associated with these operating point A and B are given by (36) and they are shown in FIGS. 8A and 8B.

Consider the operating point A associated with the desired bandwidth B₁. If the transmit energy E is increased beyond the corresponding threshold value E₁with bandwidth held constant at B₁, the performance SINR₁(E, B₁) increases beyond that at A since from (43)

$\begin{matrix} {SINR}_{1} (E, B_{1}) = a_{1} - \frac{c_{1}^{2}}{E + b_{1}} \geq a_{1} - \frac{c_{1}^{2}}{E_{1} + b_{1}} = {SINR}_{1} (B_{1}) & (51) \end{matrix}$

and it is upper bounded by a₁. Here a₁corresponds to the SINR performance for bandwidth B₁as the transmit energy E→∞. Note that a₁, B₁and c₁are the constants in (29)-(31) with Ω₊ replaced by Ω₊(B₁). The dashed curve Aa₁in FIG. 7 represents SINR₁(E, B₁) for various values of E. From (42), each point on the curve Aa₁generates a new transmit waveform as well.

Interestingly the dashed curves in FIG. 7 cannot cross over the optimum performance (solid) curve SINR(B_i). If not, assume the performance SINR₁(E,B₁) associated with the operating point A crosses over SINR(B_i) at some E₁′>E₁. Then from (47), there exists a frequency point B₁′>B₁with threshold energy E₁′ and optimum performance SINR₁(B₁′). By assumption,

SINR
₁(E₁′,B₁)>SINR₁(B₁′). (52)

But this is impossible since SINR₁(B₁′) corresponds to the maximum SINR realizable at bandwidth B₁′ with energy E₁′, and hence performance at a lower bandwidth B₁with the same energy cannot exceed it. Hence (52) cannot be true and we must have

SINR
₁(E₁′,B₁)≦SINR₁(B₁′), (53)

i.e., the curves Aa₁, Ba₂, etc. does not cross over the optimum performance curve ABD.

In FIG. 7, assume that the saturation performance value

a
₁
≧SINR
₁(B₂), (54)

i.e., the maximum performance level for bandwidth B₁is greater than of equal to the performance level associated with the operating point B with a higher bandwidth B₂and a higher threshold energy E₂. Draw a horizontal line through B to intersect the curve Aa₁at C, and drop a perpendicular at C to intersect the x-axis at E₃. From (51) with E=E₃we get

SINR
₁(E₃,B₁)=SINR₁(B₂). (55)

Thus the operating point C on the curve Aa₁is associated with energy E₃, bandwidth B₁and corresponds to a performance level of SINR₁(B₂) associated with a higher bandwidth. Notice that

E₃>E₂>E₁, and B₁<B₂. (56)

In other words, by increasing the transmit energy from E₁to E₃while holding the bandwidth constant at B₁, the performance equivalent to a higher bandwidth B₂can be realized provided B₂satisfies (54). As a result, energy-bandwidth tradeoff is possible within reasonable limits. The transmit waveform |F₃(ω)|²associated with the operating point C is obtained using (42) by replacing E and B_othere with E₃and B₁respectively. and it is illustrated in FIG. 8C. In a similar manner, the waveforms corresponding to the operating points A and B in FIG. 7 can be obtained using equation (42) by replacing the energy-bandwidth pair (E,B₀) there with (E₁,B₁) and (E₂,B₂) respectively. These waveforms are shown in FIG. 8A and FIG. 8B respectively A comparison with FIGS. 8A and 8B show that the waveform at C is different from those associated with operating point A and B.

It is important to note that although the transmit waveform design |F₃(ω)|²and |F₁(ω)|²correspond to the same bandwidth (with different energies E₃and E₁), one is not a scaled version of the other. Changing transmit energy from E₁to E₃unleashes the whole design procedure and ends up in a new waveform |F₃(ω)|²that maintains a performance level associated with a larger bandwidth B₂.

The question of how much bandwidth tradeoff can be achieved at an operating point is an interesting one. From the above argument, equality condition in (54) gives the upper bound on how much effective bandwidth increment can be achieved by increasing the transmit energy. Notice that for an operating point A, the desired bandwidth B₁gives the operating bandwidth Ω₊(B₁) and from (29) the performance limit

$\begin{matrix} a_{1} = \frac{1}{2 π} \int_{Ω_{-} (B_{1})} \frac{{\langle Q (ω) \rangle}^{2}}{G_{c} (ω)} \partial ω & (57) \end{matrix}$

for bandwidth B₁can be computed. Assume B₂>B₁, and from (48) SINR₁(B₂) the minimum performance at B₂also can be computed, and for maximum bandwidth swapping the nonlinear equation

a
₁
=SINR
₁(B₂) (58)

must be solved for B₂. Then

ΔB(B₁)=B₂−B₁ (59)

represents the maximum bandwidth enhancement that can be realized at B₁. This is illustrated in for the target situation in FIG. 7. Notice that the maximum operating bandwidth if finite in any system due to sampling considerations and after normalization, it is represented by unity. Hence the upper limit in (59) must be min(1, B₂). This gives

ΔB=min(1,B₂)−B₁ (60)

and this explains the linear nature of ΔB for larger value of B_i. In that case, bandwidth can be enhanced by 1−B₁only.

The design approach described in this section requires the knowledge of the target characteristics in addition to the clutter and noise spectra.

Although the invention has been described by reference to particular illustrative embodiments thereof, many changes and modifications of the invention may become apparent to those skilled in the art without departing from the spirit and scope of the invention. It is therefore intended to include within this patent all such changes and modifications as may reasonably and properly be included within the scope of the present invention's contribution to the art.

Claims

1. A method comprising providing a transmitter and a receiver;selecting a desired bandwidth Bo for a transmit signal f(t);outputting the transmit signal f(t) from the transmitter towards a target and towards interference;wherein the target produces a target signal;and further comprising receiving a combination signal at the receiver, wherein the combination signal includes noise and the transmit signal f(t) modified by interacting with the target and the interference;wherein the receiver acts on the combination signal to form a receiver output signal;wherein the transmit signal f(t) is selected and the receiver is configured so that the ratio of the receiver output signal to interference plus noise power is maximized while maintaining the desired bandwidth Bo for the receiver output signal; andwherein the Fourier transform F(ω) of the transmit signal f(t) is given by:
2. The method of claim 1 further comprising selecting a receiver filter for the receiver such that the receiver filter has a Fourier transform Hopt(ω) is given by
3. The method of claim 1 wherein the transmit signal f(t) so generated has a minimum energy given by the threshold value Emin where
4. A method comprising providing a transmitter and a receiver;selecting a desired bandwidth B1 for a transmit signal f(t);selecting a desired energy E that exceeds a predetermined energy level Emin given below;outputting the transmit signal f(t) from the transmitter towards a target and towards interference;wherein the target produces a target signal;and further comprising receiving a combination signal at the receiver, wherein the combination signal includes noise and the transmit signal f(t) modified by interacting with the target and the interference;wherein the receiver acts on the combination signal to form a receiver output signal;wherein the transmit signal f(t) is selected and the receiver is configured so that the ratio of the receiver output signal to interference plus noise power is maximized while maintaining the desired bandwidth B1 for the receiver output signal; and
5. The method of claim 1 wherein the interference and the noise are part of a space based radar scene.
6. The method of claim 1 wherein the interference and the noise are part of an airborne based radar scene.
7. The method of claim 1 wherein the interference and the noise are part of a ground based radar scene.
8. The method of claim 1 wherein the interference and the noise are part of an underwater sonar scene.
9. The method of claim 1 wherein the transmit signal f(t), the target, the interference, and the noise are part of a cellular communication scene (ground based and/or space based) wherein the transmit signal f(t) represents the desired voice or data modulated signal, the target represents the channel, and the interference represents all interference signals.
10. A method comprising providing a transmitter and a receiver;outputting the transmit signal f(t) from the transmitter towards a target and towards interference;wherein the target produces a target signal;wherein the transmit signal f(t) has a transmit signal bandwidth, transmit signal energy, and a transmit signal waveform;and further comprising receiving a combination signal at the receiver, wherein the combination signal includes noise and the transmit signal f(t) modified by interacting with the target and the interference;wherein the receiver acts on the combination signal to form a receiver output signal;wherein the receiver output signal has a receiver output signal waveform;trading transmit signal bandwidth against transmit signal energy by redesigning the transmit signal and receiver output signal waveforms without sacrificing the performance level.
11. The method of claim 10 further comprising selecting an initial desired first bandwidth B1 for the transmit signal bandwidth and determining the minimum required energy E1 for the transmit signal according to
12. The method of claim 11 further comprising selecting a second bandwidth B2 larger than B1 so as to satisfy the condition
13. The method of claim 12 further comprising constructing a first transmit signal having a first transmit signal waveform;constructing a second transmit signal having a second transmit signal waveform;wherein the first transmit signal is different from the second transmit signal and the first transmit signal waveform is different from the second transmit signal waveform;wherein the first transmit signal has a prescribed bandwidth which is B2;wherein the first transmit signal has an minimum energy which is E2;wherein the second transmit signal has a bandwidth which is B1;wherein the second transmit signal has a prescribed energy level which is E3;
14. The method of claim 10 wherein the transmit signal f(t), the target, the interference, and the noise are part of a cellular communication scene where the transmit signal f(t) represents the desired voice or data modulated signal, the target represents the channel, and the interference represents all interference signals.
15. The method of claim 10 wherein the interference and the noise are part of a space based radar scene.
16. The method of claim 10 wherein the interference and the noise are part of an airborne based radar scene.
17. The method of claim 10 wherein the interference and the noise are part of a ground based radar scene.
18. The method of claim 10 wherein the interference and the noise are part of an underwater sonar scene.
19. An apparatus comprising a transmitter; anda receiver;wherein the transmitter is configured to transmit a transmit signal f(t) of a bandwidth Bo towards a target and towards interference;wherein the target produces a target signal;wherein the receiver is configured to receive a combination signal, wherein the combination signal includes noise and the transmit signal f(t) modified by interacting with the target and the interference;wherein the receiver is configured to act on the combination signal to form a receiver output signal;wherein the transmit signal f(t) is selected and the receiver is configured so that the ratio of the receiver output signal to interference plus noise power is maximized while maintaining the bandwidth Bo for the receiver output signal; andwherein the Fourier transform F(ω) of the transmit signal f(t) is given by:
20. The apparatus of claim 19 further comprising a receiver filter which is part of the receiver;and wherein the receiver filter is configured to have a Fourier transform Hopt(ω) which is given by
21. The apparatus of claim 19 wherein the transmit signal f(t) has a minimum energy given by the threshold value Emin where
22. An apparatus comprising a transmitter; anda receiver;wherein the transmitter is configured to output a transmit signal f(t) having a bandwidth B1 and an energy E that exceeds a predetermined energy level Emin given below;wherein the transmitter is configured to output the transmit signal f(t) towards a target and towards interference;wherein the target produces a target signal;wherein the receiver is configured to receive a combination signal at the receiver, wherein the combination signal includes noise and the transmit signal f(t) modified by interacting with the target and the interference;wherein the receiver is configured to act on the combination signal to form a receiver output signal;wherein the transmit signal f(t) is selected and the receiver is configured so that the ratio of the receiver output signal to interference plus noise power is maximized while maintaining the bandwidth B1 for the receiver output signal; and
23. The apparatus of claim 19 wherein the interference and the noise are part of a space based radar scene.
24. The apparatus of claim 19 wherein the interference and the noise are part of an airborne based radar scene.
25. The apparatus of claim 19 wherein the interference and the noise are part of a ground based radar scene.
26. The apparatus of claim 19 wherein the interference and the noise are part of an underwater sonar scene.
27. The apparatus of claim 19 wherein the transmit signal f(t), the target, the interference, and the noise are part of a cellular communication scene (ground based and/or space based) wherein the transmit signal f(t) represents the desired voice or data modulated signal, the target represents the channel, and the interference represents all interference signals.
28. An apparatus comprising a transmitter;and a receiver;wherein the transmitter is configured to output a transmit signal f(t) towards a target and towards interference;wherein the target produces a target signal;wherein the transmit signal f(t) has a transmit signal bandwidth, transmit signal energy, and a transmit signal waveform;wherein the receiver is configured to receive a combination signal at the receiver, wherein the combination signal includes noise and the transmit signal f(t) modified by interacting with the target and the interference;wherein the receiver is configured to act on the combination signal to form a receiver output signal;wherein the receiver output signal has a receiver output signal waveform;wherein the transmitter and the receiver are configured so that transmit signal bandwidth is traded off against transmit signal energy by redesigning the transmit signal and receiver output signal waveforms without sacrificing the performance level.
29. The apparatus of claim 28 further comprising wherein the transmitter is configured so that an initial desired first bandwidth B1 for the transmit signal bandwidth is selected and a minimum required energy E1 for the transmit signal is selected according to
30. The apparatus of claim 29 further comprising configuring the transmitter so that a second bandwidth B2 is selected, which is larger than B1 so as to satisfy the condition
31. The apparatus of claim 30 further comprising wherein the transmitter is configured so that it outputs a first transmit signal having a first transmit signal waveform;wherein the transmitter is configured so that it outputs a second transmit signal having a second transmit signal waveform;wherein the first transmit signal is different from the second transmit signal and the first transmit signal waveform is different from the second transmit signal waveform;wherein the first transmit signal has a prescribed bandwidth which is B2;wherein the first transmit signal has an minimum energy which is E2;wherein the second transmit signal has a bandwidth which is B1;wherein the second transmit signal has a prescribed energy level which is E3;
32. The apparatus of claim 28 wherein the transmit signal f(t), the target, the interference, and the noise are part of a cellular communication scene where the transmit signal f(t) represents the desired voice or data modulated signal, the target represents the channel, and the interference represents all interference signals.
33. The apparatus of claim 28 wherein the interference and the noise are part of a space based radar scene.
34. The apparatus of claim 28 wherein the interference and the noise are part of an airborne based radar scene.
35. The apparatus of claim 28 wherein the interference and the noise are part of a ground based radar scene.
36. The apparatus of claim 28 wherein the interference and the noise are part of an underwater sonar scene.

STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT

The present invention is based upon work supported and/or sponsored by the Air Force Research Laboratory (AFRL), Rome, N.Y. under contract No. FA8750-06-C-0202

APPARATUS AND METHOD FOR PROVIDING ENERGY - BANDWIDTH TRADEOFF AND WAVEFORM DESIGN IN INTERFERENCE AND NOISE

Information

Publication Number

Date Filed

Date Published

Inventors

CPC

US Classifications

International Classifications

Abstract

Description

Claims

STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT