TWI420848B - Channel Estimation Method for Relay System with Orthogonal Frequency Division Multiplexing - Google Patents

Description

Channel estimation method for orthogonal frequency division multiplexing relay system

The present invention relates to a channel estimation method, and more particularly to a channel estimation method for a relay system that is applied after amplification by orthogonal frequency division multiplexing in a wireless communication system.

With the needs of the times, mobile communication technology is constantly pursuing the progress of quality and quantity. In the next generation of mobile communication systems, the relay system is used to expand the coverage of the system and improve the overall transmission volume, which can effectively control The transmission efficiency and the transmission loss between the base station and the mobile station can significantly reduce the transmission power of the mobile station and effectively extend the life of the mobile station battery. In addition, properly deploying the relay system at the edge of the coverage area of the base station or the area where the shadowing effect is severe can make the base station's data transmission rate for users in different locations within its coverage area tend to be consistent ( Uniform Data Rate Coverage). Furthermore, the relay system is also applied to the implementation of Transmit Diversity. Due to the transmission characteristics of the wireless channel, the destination can receive the signal transmitted from the direct path (source end - destination end), and From the signal of the relay path (source-relay system-destination), the two signals from different paths (which carry the same information) can be combined at the destination to obtain the diversity of transmission, thus effectively Improve the signal quality of the destination. In addition, the relay system can also be applied to the Cooperative Diversity technology in the wireless communication environment to combat shadowing and multi-path fading, especially when the source The terminal, the relay system, and the destination end are all in the state of a single antenna.

Orthogonal Frequency Division Multiplexing (OFDM) technology is an efficient modulation method, which not only effectively increases the bandwidth efficiency (Bandwidth Efficiency), but also effectively avoids symbol interference between signals; In addition, orthogonal frequency division multiplexing technology has been applied in various wireless and wired communication systems. For example, the IEEE 802.16e specification, 3GPP-LTE and WiMAX are all based on orthogonal frequency division multiplexing technology.

The operation mode of the relay system can be roughly divided into two types: Decode and Forward and Amplify and Forward. After receiving the signal transmitted by the source end, the relay system transmitted after re-encoding will decode and re-encode, and then forward it to the destination end; and the amplified relay system will be amplified by the receiving signal at the source end. Transferred to the destination end, the amplified relay system does not require re-encoded active components compared to the re-encoded transmission method, which is not only low cost but also has the advantages of easy installation and small size.

Currently in the literature, the channel estimation method for the amplified transmission relay system based on orthogonal frequency division multiplexing is mainly as follows: CS Patel and GL Stuber in IEEE Trans. Wireless Commun., Vol 6, pp. 2348 -2356, 2007 "Channel estimation for amplify and forward relay based cooperation diversity systems" which exposes a channel estimation method for Linear Minimum Mean Square Error (LMMSE), and Fand Liu, Zhe Chen , "Xin Zhang and Dacheng Yang" in "International Conference on Wireless Communications, Networking and Mobile Computing, Oct. 2008, pp. 1-4." "Channel estimation for amplify and forward relay in OFDM system" which discloses a low rank minimum The Low Rank MMSE channel estimation method is based on the Singular Value Decomposition (SVD) and avoids the operation of the inverse matrix of the Channel Correlation Matrix.

However, the methods of channel estimation existing in the literature are all the results of directly estimating the composite channel, that is, the overall channel estimation method from the source to the destination through the relay system, and it is not possible to estimate the source to the relay system individually. And the individual channels of the relay system to the destination end, and the difficulty in calculating the individual channels is that the signals transmitted by the relay system are amplified, and the channels and noise observed at the destination end are no longer Gaussian. In addition, the channel estimation method existing in the prior art is developed under the statistical characteristics of the Multipath Intensity Profile (MIP) of the known wireless channel. In order to obtain the statistical characteristics of the channel, an additional system is required. resource of.

As mentioned earlier, in order to obtain the transmission diversity, the destination must combine the signals transmitted from the direct path (source end-destination end) and relay path (source end-relay system-destination end) respectively to obtain the transmission. Diversity, thus effectively improving the signal quality of the destination. However, to achieve this combination of optimization, the destination must have an individual channel estimate from the source end to the relay system and the relay system to the destination end.

The main object of the present invention is to solve the problem that the prior art cannot separately estimate the channel from the source end to the channel of the relay system and the channel from the relay system to the destination end.

To achieve the above objective, the present invention provides a channel estimation method for a quadrature frequency division multiplexing relay system, which is based on an expected value maximization algorithm and estimates a source end in a wireless communication system to a destination end. a relay system and a channel impulse response (Channel Impulse Responses) of the relay system to the destination end, the estimation method includes the following steps:

S1: establishing a system model, considering that the path of the relay system is a two-stage path, and the system is set in a time division duplexing state, respectively, to the source end to the relay system and the The relay system establishes a signal model to the destination end, wherein the signal sent by the source terminal is:

N is the size of the inverse discrete Fourier transform, and d _k is the data uploaded by the kth channel subcarrier, which are independent of each other between the respective channels. Let { h _l ^r , l =0,1,..., L _r -1} be the coefficient of the impulse response of the channel from the source to the relay system, and the signal received on the relay system Can be established as:

among them, Is the channel gain for the kth subcarrier, and In order to relay the noise on the system side, after transmitting back to the frequency domain by Fourier transform, the transmission signal of the time domain is then converted back by using the inverse Fourier transform. , n =0,1,..., N -1}, then the signal received by the destination is:

Where { h _l ^d , l =0,1,..., L _d -1} is the coefficient of the channel impulse response between the relay system and the destination end, For the noise of the relay system, α _{k is} the amplification gain of the relay system at the kth subcarrier, and The interference value on the kth subcarrier for the noise of the relay system side.

S2: setting an expected value function, and setting an expected value function through an expected value setting unit;

S3: performing maximization processing to maximize processing through a maximizing processing unit;

S4: Performing iteration, iterating between step S2 and step S3 until the set number of iterations is reached.

It can be seen from the above description that the present invention estimates the channel impulse response of the source end to the relay system and the relay system to the destination end on the orthogonal frequency division multiplexing system based on the expectation value maximization algorithm. The coefficient, in the relay system transmitted after amplification, can estimate the channel impulse response coefficient of the two channels at the destination end. The invention does not need to additionally estimate the statistical characteristics of the multi-path strength statistical graph of the wireless channel, and can effectively reduce the waste of resources.

The detailed description and the technical content of the present invention are as follows with reference to the drawings: Referring to FIG. 1, FIG. 2 and FIG. 3, FIG. 1 is a schematic diagram of a system channel according to a preferred embodiment of the present invention, and FIG. 2 is a schematic diagram of the system. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 3 is a schematic diagram of a unit configuration according to a preferred embodiment of the present invention. As shown in the figure, the present invention is a channel estimation method for a quadrature frequency division multiplexing relay system. Based on the expected value maximization algorithm, the destination end 30 respectively performs a channel of a source terminal 10 to a relay system 20 in the wireless communication system and a channel of the relay system 20 to a destination end 30. Channel estimation, the estimation method includes the following steps:

S1: Establish a system model 40 according to the inverse discrete Fourier transform, and respectively establish a signal model for the source terminal 10 to the relay system 20 and the relay system 20 to the destination terminal 30, wherein the signal sent by the source terminal 10 is :

The signal received on the relay system 20 can be established as:

After translating back to the frequency domain by Fourier transform, the transmission signal of the time domain is then converted back by using the inverse Fourier transform. , n =0,1,..., N -1}, then the signal received by the destination terminal 30 is:

among them, and The channel impulse response of the source terminal 10 to the relay system 20 and the relay system 20 to the destination terminal 30, respectively, α _{k is} the amplification gain of the relay system 20 at the kth subcarrier, according to equation (3) The effect of the source end 10 to the relay system 20 and the two channels of the relay system 20 to the destination end 30 can be clearly seen: the signal seen by the destination end 30 includes the source end 10 to the The relay system 20 and the individual channel effects of the relay system 20 to the destination end 30.

S2: The system model 40 is vector matrixed, and the equation (3) vector is matrixed by a vector matrix unit 41 to facilitate calculation. The result of the rewriting is:

The superscript symbols ^H and ^T represent Hermitian Transpose and Transpose, respectively, and U represents a discrete Fourier transform matrix. Is an N × L matrix in which the ( p , q )th element is e ^{- j 2π( p -1)( q -2) / N} . h ^r and h ^d are vectors of the channel impulse response of the source terminal 10 to the relay system 20 and the relay system 20 to the destination end 30 of L _r ×1 and L _d ×1, respectively, w ^d is the purpose The N × 1 vector composed of the end 30 noise. A , D and W ^r are N × N diagonal diagonal matrices, and their kth diagonal elements are α _k , d _k and . And defined here with , Representing the signal part, and On behalf of the relay system 20, the noise is transmitted to the noise portion caused by the destination terminal 30.

S3: setting the expected value function, inputting the result of equation (4) to an expected value setting unit 42 to set an expected value function, the expected value setting unit 42 is based on the following equation (5):

Where E _m [‧] represents the expected value operation for m, p (‧) represents a Probability Density Function (PDF), and Φ={ h ^r , h ^d } is the set of parameters to be estimated, For the set of parameters estimated after j iterations, y ^d is the result of (4), called Incomplete Data, and m is Missing Data. The combination of { y ^d , m} is chosen to be Complete Data, which is calculated by Equation (5) and obtains an expected value function.

S4: Simplify and bring into the probability density function, according to the Chain Rule of Probability, use a simplification unit 43 to remove the number of terms independent of Φ in equation (5), and simplify equation (5) to The simplification unit 43 is coupled to a probability density function unit 44 that provides a probability density function and inputs to the simplification unit 43; the simplification unit 43 finally outputs a reduced equation (6) :

Where Ω _m is the space vector of m .

S5: Calculate the expected value function of the closed form. It is difficult for the equation (6) to integrate to obtain a closed form representation. However, in each iteration, if it is directly optimized for (6) Processing, its complexity is unacceptable. Therefore, we use an expectation value function closed form processing unit 45 to reduce equation (6) to an equivalent closed form representation of the expected value function, and equation (6) is rewritten as the following equivalent equation (7):

Wherein, the invention defines: N × N matrix Z (Φ) is U ^H A diag { Mean vector of h ^d }, N ×1 for , and N × N Covariance Matrix K ( )for .

S6: performing maximization processing, maximizing processing according to a maximum value processing unit 46 for the expected value function (7) expressed by the closed form, wherein the maximizing processing unit 46 performs maximization processing according to the formula (8):

Ω is the space vector of Φ.

It should be particularly noted that the present invention solves by using Conditional Conditional Maximization (ECM), which updates only one parameter to be estimated ( h ^r or h ^d ) in Φ at a time, where:

Equation (9) and S _r (in equation (10) ) and S _d ( ) are defined as:

Where U _L contains the first L row vectors of the discrete Fourier transform matrix U , and diag{ a } represents the diagonalized matrix with the vector a as the diagonal element.

S7: performing an iterative judgment, the result of which will be passed through the maximization processing unit 46 Bringing back to the expected value function closed form processing unit 45 performs the next iteration to find the best convergence solution, and after satisfying the set number of iterations, ends the channel estimation.

Please refer to FIG. 4 , which is a schematic diagram of the result of the simulation result root mean square error (MSE) according to a preferred embodiment of the present invention, and please cooperate with FIG. 1 to use a fixed gain amplifier system in this embodiment. And establish a simulation system under Binary Phase Shift Keying (BPSK) and orthogonal frequency division multiplexing systems, assuming that the channel is Rayleigh Fading, the source end 10 to the relay system 20 and the number of impulse response coefficients L _{r and} L _{d of} the relay system 20 to the destination end 30 are both set to 4; in addition, the error correction code is a code rate of 1/2 and a memory length (Memory) ) is a Convolutional Code of 4, and the signal to noise ratio (SNR) of the two channels is equal. As shown in FIG. 4, the root mean square error of the source end to the channel 51 of the relay system is In the case of low signal-to-noise ratio, the performance is significantly worse than that of the relay system to the destination end channel 52. However, in the case where the signal-to-noise ratio is gradually increased, the difference between the two is gradually reduced, and the average value 53 is as shown in the figure. 4 is shown. In addition, the error floor is also found to be 4×10 ⁻³ in FIG. 4 , and it can be seen that the source terminal 10 can be connected to the relay system 20 and the relay system 20 by the present invention. Channels to the destination end 30 are channel estimates, respectively.

Figure 5 is a graph showing the results of the number of iterations of the channel estimation. As shown in the figure, as the number of iterations increases, the result gradually converges, and after the third iteration, the result has approached a fixed value. . Referring to FIG. 6 again, when the number of iterations is gradually increased, the performance is very close to the ideal value. For example, when the bit error rate (BER) is 1×10 ⁻³ , the simulation result of the present invention is compared with The result of the ideal channel only loses a signal-to-noise ratio of 0.5 dB.

Please refer to FIG. 7 , which is built on the 16th-order phase key shift modulation (16-PSK). It can be found from FIG. 7 that the channel estimation method of the present invention is based on Maximum Ratio Combining (MRC). After the diversity combining 60 results, the signal-to-noise ratio is 10 dB better than the conventional technique 61 when the root mean square error is also 6 × 10 ^-3 . Referring to FIG. 8, in the performance of the bit error rate, the result of the diversity combining 60 is also significantly improved compared with the prior art 61, and thus the present invention has practical reference value.

In summary, the present invention is based on an expectation maximization algorithm, and under the relay system 20 transmitted after the orthogonal frequency division multiplexing, the destination terminal 30 estimates the source terminal 10 to the relay system. 20 and two individual channel impulse responses of the relay system 20 to the destination end 30. The results of the two separate channel estimations allow the system to understand the strength of the signal transmission of the individual transmission channels, and determine the quality of the channel based on the channel state information, and then determine the amount of data to be transmitted. In addition, the present invention does not require additional calculation of the statistical characteristics of the multipath strength statistics of the delivery channel, which can effectively reduce the waste of resources. Furthermore, the present invention is applied to the diversity combination of the maximum synthesis ratios, and has a significant improvement in root mean square error and bit error rate as compared with the prior art. Therefore, the present invention is highly progressive and conforms to the requirements of the invention patent application, and the application is made according to law, and the praying office grants the patent as soon as possible.

The present invention has been described in detail above, but the foregoing is only a preferred embodiment of the present invention, and is not intended to limit the scope of the invention. That is, the equivalent changes and modifications made by the scope of the present application should remain within the scope of the patent of the present invention.

S1-S7. . . step

10. . . Source

20. . . Relay system

30. . . Destination

40. . . System model

41. . . Vector matrix unit

42. . . Expectation value setting unit

43. . . Simplification unit

44. . . Probability density function unit

45. . . Expected value function closed form processing unit

46. . . Maximize processing unit

51. . . The source end to the relay system channel

52. . . The relay system to the destination channel

53. . . average value

60. . . Diversity combination

61. . . Conventional technology

1 is a schematic diagram of a system channel in accordance with a preferred embodiment of the present invention.

2 is a flow chart showing the steps of a preferred embodiment of the present invention.

3 is a schematic diagram of a unit configuration of a preferred embodiment of the present invention.

4 is a schematic diagram showing the results of root mean square error of simulation results in accordance with a preferred embodiment of the present invention.

Figure 5 is a schematic diagram showing the average iteration result of a channel according to a preferred embodiment of the present invention.

FIG. 6 is a schematic diagram showing the result of bit error rate according to a preferred embodiment of the present invention.

Figure 7 is a schematic illustration of a comparison of a preferred embodiment of the present invention.

Figure 8 is a diagram showing the error rate of the comparison result of a preferred embodiment of the present invention.

S1-S7. . . step

Claims (8)

A channel estimation method for a quadrature frequency division multiplexing relay system, which is applied to a relay system transmitted after amplification, based on an expectation maximization algorithm, and respectively estimates a source in a wireless communication system at a destination end Ending to a channel of a relay system and a channel of the relay system to a destination end, the estimating method comprises the steps of: establishing a system model, respectively, the source end to the relay system and the relay system to the The destination end establishes a channel, and vector matrix is performed by a vector matrix unit to establish a vector matrix for the development of the channel estimation algorithm. The signal received by the destination is: Setting an expectation function, performing an expectation setting through an expectation setting unit; performing a maximization process, maximizing processing by a maximizing processing unit; and performing iteration, bringing the result after the maximizing processing unit back to the expected value The setting unit performs iteration and re-adjusts the expected value until the set number of iterations is reached. The channel estimation method for the orthogonal frequency division multiplexing relay system according to claim 1, wherein the source signal model in the system model is: x _n = (1/ ) ,0 n N -1. The channel estimation method for the orthogonal frequency division multiplexing relay system described in claim 2, wherein the signal model received by the relay system in the system model is: The channel estimation method of the orthogonal frequency division multiplexing relay system according to claim 1, wherein the vector matrix unit simulates the channel as: A channel estimation method for a quadrature frequency division multiplexing relay system as described in claim 4, wherein: The channel estimation method of the orthogonal frequency division multiplexing relay system according to claim 1, wherein the expected value setting unit is according to the equation Q' ( Φ | ) = E _m [log p ( m,y ^d |Φ)| y ^d , ] The expected value is set, y ^d is an incomplete data, and m corresponds to y ^d and is a lost data, and the two are combined to become a complete data. The channel estimation method of the orthogonal frequency division multiplexing relay system according to claim 1, wherein the method further includes a simplification and a probability density function step, which is performed via a simplification unit and according to a probability chaining rule Simplification is performed, and a probability density function unit is connected to the simplification unit to provide a probability density function to the simplification unit. The channel estimation method for the orthogonal frequency division multiplexing relay system according to claim 7, wherein the closed value processing unit is connected to the simplification unit by using an expected value function, and the data after the simplification unit is performed. Closed form processing.