KR20040049589A - Method of analyzing the inner loop power control in a cellular system - Google Patents

Abstract

PURPOSE: A method for interpreting closed inner-loop power control of a mobile communication system is provided to mathematically model components, and to mathematically obtain a transmission power probability distribution function through different operating methods, then to mathematically obtain a reception power probability distribution function, thereby remarkably improving an operating speed of a simulation. CONSTITUTION: In consideration of a channel gain correlation, a system mathematically calculates a transmission power probability distribution function by using features of a target Doppler frequency and a channel gain. In case of AWGN(Additive White Gaussian Noise) or high-speed fading, the system obtains a transition matrix in different methods, and mathematically calculates a transmission power probability distribution function by using a vector. The system operates the transmission power probability distribution function with differently obtained values, and mathematically calculates a reception power probability distribution function.

Description

METHOD OF ANALYZING THE INNER LOOP POWER CONTROL IN A CELLULAR SYSTEM}

The present invention relates to a method for analyzing an internal closed power control of a mobile communication system. Particularly, after mathematically modeling members of an internal power control circuit among closed power control circuits, a transmission power probability distribution function is obtained by using a different calculation method according to a condition. The present invention relates to a closed internal power control analysis method of a mobile communication system that can obtain a received power probability distribution function.

In a mobile communication environment, power control is required to maintain a constant call quality of each user while maintaining a call regardless of changing signal power while reducing interference with other users.

Power control in a wideband code division multiple access (W-CDMA) system can be largely divided into open-loop power control and closed-loop power control. Open power control is a power control method that determines transmission power by estimating attenuation in the signal transmission path from the received signal power, and is used in limited cases such as call setup or when high accuracy is not required.

However, closed power control maintains a constant call quality by allowing the receiver to measure the call quality and send an "up" command to the transmitter if the measured quality is lower than the target quality and a "down" command to the transmitter. Do it. Thus, the closed power control circuit has a feedback loop.

Closed power control is further divided into inner-loop power control and outer-loop power control. The internal power control circuit maintains a certain level of signal-to-interference ratio (SIR). In addition, the external power control circuit adjusts a target SIR by measuring a bit error rate (BER) or a frame error rate (FER) of a received symbol to maintain a certain level of BER or FER to maintain a constant call quality.

Since the present invention relates to internal power control, the general internal power control circuit will be described. This example targets the reverse direction of the WCDMA (terminal to base station), and the internal power control cycle of the WCDMA is 1500 Hz in slot units.

1 is a block diagram of an internal power control circuit, in which a transmitter transmits a signal by multiplying a gain value controlled by a power control circuit as shown, and the signal reaches a receiver after multiplying by a channel gain in a wireless channel. do.

The internal power control circuit operates through a number of functions shown, where a transmit power probability distribution function for the input power ( ) And the received power probability distribution function for power passing through the uplink fading channel ( ) Determines the characteristics of the internal power control circuit.

The SIR estimator of the receiver estimates the SIR of the received signal, compares this value with the target SIR, and generates a transmitted power control (TPC) command bit as a result. However, since the TPC bit is also transmitted to the transmitter through the radio channel, a bit error occurs during the transmission process. Thus, the variables that determine the reception performance after performing internal power control are the accuracy of the SIR estimator, the bit error rate of the TPC command bits (ε), the power control staff size (Δg) of the transmitter, and the delay time (ξ) of the feedback loop. And so on.

Summarizing the multiple functions and corresponding variables shown, g (t) is the transmit power function for time t, ω (t) is the channel power gain at time t, and TPC _T is the TPC command bit transmitted at the receiver. Where TPC _R is the TPC command bit received at the transmitter, β (ω) is the SIR estimator characteristic function, α (ω) is the characteristic function of the SIR estimator and TPC command bit bit error rate, and δ _o is the target SIR.

In practice, in order to design power control-related components such as SIR estimators and to optimally adjust the related variables, it is necessary to clearly know the characteristics of the power control circuit as the variables change. From the point of view of the communication system, what is important as a characteristic of the power control circuit is the Tx power probability density function (Tx power PDF) of the transmitting end and the Rx power probability density function (Rx power) of the receiving end after power control. PDF). From the former, the power consumption of the transmitter can be calculated, and from the latter, the received bit error rate and the capacity of the system can be calculated.

Therefore, in order to obtain the power probability distribution function and the received power probability distribution function of the transmission after the internal power control, an empirical method or a software simulation using the power control circuit in hardware and changing various variable values is used. How to do it.

That is, after implementing the power control circuit as shown in FIG. 1 substantially or in software, a large number of signal samples are generated to perform power control directly or software, and then the transmit power probability distribution function is recorded from the recorded transmit / receive power samples. And the received power probability distribution function.

As described above, in order to obtain the transmit / receive power probability distribution function for the conventional internal power control analysis, if the actual hardware is implemented and the results of experiments with a large number of sample signals are used, the implementation is costly and time consuming, and a large amount of maintenance is performed after the implementation. Is difficult to modify to the optimum condition, and when using a simulation method using software, many combinations of variables are required, and a lot of computation time is required, and reliability and accuracy are low.

In view of the above problems, the present invention mathematically models the components constituting the internal power control, and mathematically obtains a transmission power probability distribution function through a different calculation method according to various cases, and thus, the received power probability distribution function. It is an object of the present invention to provide a method for analyzing the closed internal power control of a mobile communication system that can be obtained mathematically.

1 is an internal power control configuration diagram.

2 is a flow chart of calculating the transmit power probability distribution function of the present invention.

3 is a flowchart illustrating a calculation of a reception power probability distribution function of the present invention.

4 is an SIR estimator model.

5 is a bit error rate model of TPC command bits.

6A and 6B are transition diagrams of the transmission power probability distributions.

7 is a conceptual diagram of a partial constant approximation method.

8 is a structure of {π (ω _n )} ^D ;

Fig. 9 is a comparison diagram between the analysis method of the transmission power probability distribution function and the simulation test result in AWGN.

10 is a comparison diagram of a method of analyzing a transmission power probability distribution function and a simulation experiment result at high speed fading;

Fig. 11 is a comparison diagram of a method of analyzing a transmission power probability distribution function and simulation results when considering correlation.

12 is a comparison diagram of a method of analyzing a received power probability distribution function and simulation results when considering correlation.

In the present invention for achieving the above object, when considering the channel gain correlation, the transmission power probability distribution function is mathematically calculated using the characteristics of the target Doppler frequency and the channel gain, and the channel gain correlation is not considered. Obtaining a transition matrix in a different method for additive white Gaussian noise (AWGN) or fast fading, and mathematically calculating a transmit power probability distribution function using an eigenvector; And calculating the received power probability distribution function mathematically by calculating the calculated transmission power probability distribution function with a value obtained differently according to each case.

The mathematically calculating the transmission power probability distribution function may include determining a depth D from a Doppler frequency of a target environment in consideration of channel gain correlation, and calculating an average transition matrix according to depth from a characteristic of channel gain. ), Then Calculating a transmission power probability distribution function by numerically obtaining an eigenvector when eigenvalue of 1 is 1; Calculating a transmission power probability distribution function from an analytic solution of an eigenvector after obtaining a transition matrix from an SIR estimator and a characteristic function of a TPC bit error rate in an AWGN without considering channel gain correlation; In the case of fast fading without considering channel gain correlation, after calculating the SIR estimator and TPC bit error rate characteristic and the average transition matrix considering the fading channel characteristics, the transmission power probability distribution function is calculated from the analytic solution of the eigenvectors. It is characterized by consisting of steps.

The mathematically calculating the received power probability distribution function may include: in the case of AWGN not considering channel gain correlation, the transmit power probability distribution function becomes a received power probability distribution function as it is; For fast fading without considering channel gain correlation, the average shifting matrix according to the channel power characteristics ( ) And multiply the received power interpolation matrix I _E and the transmit power probability distribution function P by Obtaining a received power probability distribution function; When considering the correlation, obtain the depth (D) from the Doppler frequency of the target environment and use the average matrix ( ), And multiplying the transmission power probability distribution function P to obtain a reception power probability distribution function.

An embodiment of the present invention configured as described above will be described in detail with reference to the accompanying drawings.

First, the present invention is characterized by a method for calculating the transmission power probability distribution function and calculating the received power probability distribution function using the same, and the order thereof is shown in FIGS. 2 and 3.

2 is a flowchart illustrating a process of calculating a transmission power probability distribution function according to the present invention. As shown in the drawing, a transmission power probability distribution function is divided into a case in which there is a correlation between channel gains between slots and a case in which there is no correlation.

If there is no correlation, it is divided into AWGN (Additive White Gaussian Noise) and fast fading.

AWGN assumes that the gain of the channel does not change over time, and fast fading assumes that the channel gain is independent between power control cycles. In the case where there is a correlation, it is an intermediate step between no change of the channel and the rapid change, and the degree of change of the channel is represented by the Doppler frequency. The internal power control cycle is in slots, which is 1500 Hz in typical cases.

Therefore, the calculation method is incorrect because the characteristic values of the internal power control member that can reflect the corresponding characteristics are applied according to the characteristics of the above cases.

The detailed analysis method will then be described in more detail by mathematically modeling the internal power control members and using them to derive the calculation formula for each case.

FIG. 3 is a flowchart illustrating a calculation of a received power probability distribution function. As in FIG. 2, different calculation processes are listed according to three cases. However, as a characteristic, it can be seen that the appropriate characteristic matrices are obtained and multiplied by the transmission power probability distribution function P obtained in FIG. 2.

To interpret the internal power control circuit, its components must be modeled by equations.

SIR estimator modeling

4 models the SIR estimator in the form of two parts. The SIR estimator is divided into the soft decision and hard decision parts of the SIR estimation.

The soft decision portion estimates the SIR of the signal component power (ω) of the input signal, and the hard decision portion compares the estimated SIR value (δ) with the target SIR (δ _o ) to increase / decrease the command. Will occur). Using the symbols, the SIR estimator characteristic function is defined as follows.

E.g, The characteristic function of the SIR estimator that produces the estimated SIR is expressed as follows. If the SIR estimator has more complex characteristics, it can be expressed in the form of a table rather than an equation.

However, for the convenience of analysis, it is desirable to calculate the soft-determinant characteristics in the SIR estimator with the Q-functions above, approximated by Gaussian. Where σ2 is the variance of the SIR estimator modeled as Gaussian.

TPC Command Bit Error Rate Modeling and SIR Estimator Characteristic Functions Considered Together

5 illustrates TPC command bit error rate modeling, in which the path from the receiver to the transmitter through which the TPC bits are delivered should also exhibit fading characteristics. However, as shown, it is modeled as a classic binary symmetric channel. Therefore, the probability of the TPC applied to the transmitter can be expressed as follows.

Finally, the characteristic function of the TPC applied by the transmitter is expressed as follows.

Now, we have modeled the SIR characteristic function considering the TPC command bit error rate together with the mathematical modeling of the SIR estimator.

Next, to calculate the transition equation that can be the core of the present invention.

Transition Equation of Transmission Power Probability Distribution Function

The transmit power may be formulated as a discrete time-discrete valued random process that proceeds while increasing / decreasing by the power control step size Δg according to the TPC. This consideration of the TPC command bit is the basis of power control, and it is processed digitally to be processed as discrete values. Such discrete signals are intended to enable computation by a computer. Thus, consideration of the subject of computation is also an important part of the present invention.

First, in order to obtain an equation for a continuous signal, assuming that channel gain between slots is independent (when not considering correlation between slots), a transition equation for a given channel gain can be expressed as follows.

Since the distribution of the transmission power represented by the continuous signal may have only an integer squared value of the power control step size Δg, the transmission power probability distribution function may be easily changed into the discrete form as follows.

here,

Pn is a quantized transmit power probability distribution function in the nth slot. The equation converted to the discrete form as described above can be calculated using a computer, it is necessary to set the range for the calculation. The transmit power distribution only considers the range -L≤l≤L. If you make L large enough, you can ignore the error.

6 illustrates a transition equation obtained as described above, and as illustrated, FIG. 6A illustrates a transition equation in the form of a continuous signal, and FIG. 6B illustrates a transition equation in the form of a discrete signal. As shown, after the conversion to the discrete form, there will be no change in operation, but it will be understood that the variables have been simplified.

Matrix representation of the transition expression

The transmit power probability distribution function may be expressed as a vector moving in a direction as shown in FIG. 6, and thus a transition matrix may be obtained. Therefore, after obtaining the transition matrix, the simplified transition matrix equation between slots is obtained according to the consideration of correlation.

First, the transmission power probability distribution function is expressed as a vector as follows.

Now, the transition matrix for channel power gain ω _n is defined as follows.

However, the following conditions are added by definition of the transition matrix regardless of the calculated value.

In the transmission power probability distribution function calculation method of the present invention, the calculation method is different according to the channel gain correlation between slots. Therefore, it is necessary to simplify the transition equation obtained as described above according to the channel gain correlation. Through this, the calculation process of the present invention can be simplified.

First, in the case of AWGN / fast fading without considering the correlation, the transition equation between slots is expressed as follows. Since the correlation is not taken into account, there is no channel change or the channel gain is independent so that the transition matrix can be used as it is.

In the case of considering the degree of correlation, since the channel change becomes variable according to the Doppler frequency, the equation according to the depth D of the Doppler frequency should be obtained. However, in view of the fluctuation of the channel power, this continuous signal should be approximated to a partial constant, and the partial constant approximation method used in this case is shown in FIG.

As shown in FIG. 7, the channel gain is calculated by approximating a partial constant. This method approximates the channel gain by S / H (sample-and-hold) method, which changes the sampling period according to the Doppler frequency, and reduces the error in transient by observing only the probability distribution function at the sampling point. As an approximation, in this case the transition equation is expressed as By varying the depth D of the Doppler frequency, the change in the Doppler frequency can be reflected.

Now, the basic transition equations for the case of considering the correlation and the case of not considering the correlation can be seen.

Representation of Mean Transfer Expression

In the present invention, when fast fading is used, an average transition equation is used (see FIG. 2). Therefore, it is necessary to summarize the mean transition equation, and therefore, it will be understood depending on whether or not the correlation is considered here.

First, in the case of not considering the correlation (AWGN / fast fading), taking the average of the channel gain on both sides of the inter-slot transition equation is as follows.

here,

If the correlation is considered, the Doppler frequency needs to be considered.

As described above, the transition matrices were briefly arranged according to the situation, and their averages were also examined. Now, let's summarize how to compute the transition matrix in each case.

Calculation of the Transition Matrix

First, in the case of AWGN that does not consider the correlation, ω _n = 1 and ∀ _n since Channel _n does not change, so π (1) = π.

Then, for fast fading without considering correlation, the gain of the channel is independent, so for an SIR estimator approximated by Gaussian, if we have a channel gain probability distribution function, we can calculate π as the average of π (ω _n ). do.

In this case, a method of obtaining a transition matrix element for one path fading and a transition matrix element for multipath fading will be described. In this case, the transition matrix element may be obtained in a close-form.

The transition matrix element for one path fading is calculated as follows.

For M different paths of different sizes, it can be calculated as follows. If the gain for each path is τ _k , then the probability distribution function of the channel gain is given by

here,

Using this, the transition matrix element for multipath fading can be obtained as follows.

Now, considering the correlation, in this case, the average transition matrix with a transition Doppler frequency depth of D Is not obtained in close-form form, but can be easily calculated given the channel gain distribution function. Since is a matrix in which the unique region is at the center as shown in FIG. Again, it can be calculated without large calculations.

As described above, transition matrices for various cases can be obtained. Then, the transmission power probability distribution function used in the present invention will be calculated in earnest.

Calculation of the transmit power probability distribution function

In the determinant of the form P = A · P, P is the eigenvector corresponding to the eigenvalue of A = 1. Therefore, the transmit power probability distribution function is π or This can be obtained by normalizing the eigenvectors of (by definition, the sum of the probability distribution functions is 1).

Referring to FIG. 2, the transmission power probability distribution function is calculated by applying various numerical values already calculated in that order.

First, let's look at the case of not considering the correlation.

Since the above-mentioned various equations should be listed in a complicated manner, the following definition is used to simplify the expression.

If this determinant is previously developed into a system of equations according to the definition of π, it is as follows.

L _o is defined to exclude the trivial equation, Indicates the maximum value of l. If you release it and solve it in order, you get the next solution.

By averaging to satisfy the physical meaning of the probability distribution function, the transmission power probability distribution function can be obtained as follows.

Then, the case of considering the degree of correlation will be described.

The eigenvectors corresponding to eigenvalues of 1 are not obtained in closed-form form, but can be obtained numerically. The eigenvectors obtained numerically are averaged to obtain the transmission power probability distribution function.

Calculation of the reception power probability distribution function

The distribution of received power is such that discrete channel power is multiplied by a continuous channel gain value to have a continuous value. To calculate the received power distribution using a computer, And quantize with (-R ≦ l ≦ R).

Receive power is obtained by multiplying the transmit power by the channel gain value. If the correlation is not taken into account, the channel gain and the transmission power are independent, so that each average value is obtained and multiplied to obtain a reception power distribution. In the case of considering the correlation, the reception power distribution is obtained by multiplying the channel gain by the value after the transmission power progresses by the depth slot.

The following matrices are defined for expressions using matrices. First, the unit matrix for easily representing other matrices is defined as follows.

Then, the transmission power expressed by the power of (Δg) is ( Defines a matrix to be converted into a unit of received power. Generally ( Since this is less than (Δg), this transformation is similar to the interpolation or expansion operation in digital systems, so we call this matrix an expansion matrix. m is (Δg) and ( ) Ratio.

The operation of multiplying the channel gain value is expressed by multiplying the shifting matrix as follows. s is the channel gain Quantized by the power of).

First, the case of not considering the correlation is as follows.

Since is easily obtained by quantizing the channel gain probability distribution function, the received power probability distribution function can be easily obtained from the above-described transmission power probability distribution function.

Next, the case of considering the correlation is as follows.

therefore, By calculating the received power probability distribution function.

Practical application of the present invention

Now, the above-described various equations will be substituted into the flowchart of the present invention to explain the actual operation process.

In the process of calculating the transmission power probability distribution function, as shown in FIG. 2, it is determined whether to consider the correlation of channel gain.

If we consider the channel gain correlation, we determine D from the Doppler frequency of the target environment and Obtain The transmission power probability distribution function is obtained by numerically obtaining the eigenvectors at.

If the channel gain correlation is not considered, the calculation is divided into two cases.

In the case of AWGN which does not consider the correlation of channel gain, the transition matrix is obtained from the SIR estimator and the TPC bit error rate, and then the transmission power probability distribution function is obtained directly from the analytic solution of the eigenvectors.

In the case of fast fading without considering the correlation of channel gains, in addition to the SIR estimator and the TPC bit error rate characteristic function, after calculating the average transition matrix considering the characteristics of the fading channel, the transmission power probability distribution is also obtained from the analytic solution of the eigenvector. Get the function

In the process of obtaining the reception power probability distribution function, as shown in FIG. 3, it is first determined whether to consider a correlation of channel gains.

If the correlation is considered, the depth D is first determined from the Doppler frequency of the target environment. Next, the average according to the channel power characteristics Obtaining The received power probability distribution function is obtained by multiplying the transmission power probability distribution function (P).

If the channel gain correlation is not considered, the calculation is divided into two cases.

In the case of AWGN which does not consider the correlation, the transmit power probability distribution function P becomes the received power probability distribution function.

In the case of fast fading without considering the correlation, the average shift matrix according to the channel power characteristics ( ), Obtain the received power probability distribution function by

The received bit error rate P _e is obtained by performing the following calculation using the Q-function from the received power probability distribution function.

Comparison of the results obtained with the present invention and simulation

Now, assuming that the SIR estimator is a Gaussian estimator with variance 1 (σ ² = 1.0) in order to prove the effect of the results obtained through the analysis method of the present invention, the analysis method of the present invention is compared with the simulation result.

FIG. 9 is a comparison between the calculated transmission power probability distribution function in AWGN without considering the correlation by analyzing and simu. The variable var in each experiment is the variance of the SIR estimator modeled as Gaussian.

FIG. 10 is a comparison between calculation and simulation of a transmission power probability distribution function in fast fading without considering correlation.

11 is a comparison between the transmission power probability distribution function calculated by an analysis method and a simulation in consideration of correlation. The analytical method uses the depth of the Doppler frequency (D) as a variable and the Doppler frequency in the simulation.

12 illustrates a method of calculating the received power probability distribution function in the same manner when considering the correlation.

It can be seen that the results are almost the same as shown. Therefore, while showing similar results to the existing simulation, the computational speed is dramatically faster, can be used for performance analysis and design of the system, as well as can be applied to performance optimization.

As described above, the closed internal power control analysis method of the present invention provides a mathematical model of the components constituting the internal power control, and then mathematically obtains a transmission power probability distribution function through different calculation methods according to various cases. By using this method, the received power probability distribution function can be mathematically obtained, which can greatly improve the computational speed of the simulation and can be used for analysis, design and optimization of communication systems. It is also useful for system analysis.

Claims (3)

When considering the channel gain correlation, the transmission power probability distribution function is mathematically calculated using the characteristics of the target Doppler frequency and the channel gain, and the additive white Gaussian noise (AWGN) or fast fading of the channel gain correlation is not considered. Each case obtaining a transition matrix in a different manner and mathematically calculating a transmission power probability distribution function using an eigenvector; And calculating the received power probability distribution function mathematically by calculating the calculated transmission power probability distribution function with a value obtained differently according to each case. The method of claim 1, wherein mathematically calculating the transmission power probability distribution function comprises determining a depth D from a Doppler frequency of a target environment in consideration of channel gain correlation, and averaging according to depth from a characteristic of channel gain. Transition matrix ( ), Then Calculating a transmission power probability distribution function by numerically obtaining an eigenvector when eigenvalue is 1; Calculating a transmission power probability distribution function from an analytic solution of an eigenvector after obtaining a transition matrix from an SIR estimator and a characteristic function of a TPC bit error rate in an AWGN without considering channel gain correlation; In the case of fast fading without considering channel gain correlation, the average transition matrix considering the characteristics of the SIR estimator and TPC bit error rate and the characteristics of the fading channel is obtained, and then the transmission power probability distribution function is calculated from the analytic solution of the eigenvectors. Closed internal power control analysis method of a mobile communication system, characterized in that consisting of a step. The method of claim 1, wherein mathematically calculating the received power probability distribution function comprises: in the case of AWGN not considering channel gain correlation, the transmit power probability distribution function becomes a received power probability distribution function as it is; For fast fading without considering channel gain correlation, the average shifting matrix according to the channel power characteristics ( ) And multiply the received power interpolation matrix I _E and the transmit power probability distribution function P by Obtaining a received power probability distribution function; When considering the correlation, obtain the depth (D) from the Doppler frequency of the target environment and use the average matrix ( Obtaining a received power probability distribution function by multiplying the transmit power probability distribution function (P).