A multi-wavelength wireless optical communication system demodulation method based on compressed sensing

Abstract

This invention discloses a demodulation method for a multi-wavelength wireless optical communication system based on compressed sensing, comprising two parts: a transmitter and a receiver. For the transmitter, firstly, sparse mapping is performed on the information to ensure that each symbol has sparsity, satisfying the sparsity condition of compressed sensing. Then, the multiple signals, combined with pilot signals, are modulated onto multiple lasers. Finally, the optical signals emitted by the multiple lasers are coupled and emitted. For the receiver, firstly, the signal gain matrix is ​​estimated using the least squares method. Then, the model is solved using a compressed sensing reconstruction algorithm. Finally, the demodulation of the signal is completed by comparing the model solution with a threshold. Using this easily implemented demodulation method for a multi-wavelength wireless optical communication system, optical signals can be directly received using photoelectric sensors, and multiple optical signals can be demodulated using the randomness of the channel, solving the problem of high reception difficulty in traditional schemes.

Description

Technical Field

[0001] This invention relates to the field of optical communication, and in particular to a demodulation method for a multi-wavelength wireless optical communication system based on compressed sensing. Background Technology

[0002] Wireless optical communication is a communication technology that integrates wireless and optical communication, providing a rich alternative spectrum resource to the increasingly scarce radio frequency resources. With its high bandwidth, low radiation, and interference resistance, wireless optical communication is receiving increasing attention both domestically and internationally, becoming one of the important research directions in the future field of communications.

[0003] With the maturity of wavelength division multiplexing (WDM) technology, many researchers have adopted WDM to improve the stability and channel capacity of wireless optical systems, enabling multi-wavelength wireless optical communication. However, multi-wavelength wireless optical communication suffers from high reception difficulty and receiver complexity. Current traditional multi-wavelength wireless optical communication demodulation schemes first require coupling the signal light into an optical fiber, then connecting the fiber to a WDM multiplexer for beam splitting, and finally demodulating the multiple optical signals output by the multiplexer. Because coupling the signal light into the optical fiber is difficult and lacks robustness, it is essential to propose an easily implementable demodulation method for multi-wavelength wireless optical communication to avoid this problem. Summary of the Invention

[0004] To address the problems in existing multi-wavelength wireless optical communication demodulation schemes, this invention provides a demodulation method for multi-wavelength wireless optical communication systems based on compressed sensing. This method uses photoelectric sensors to directly receive signal light, and compared to traditional schemes, the photoelectric sensor array approach solves the problem of high reception difficulty in traditional methods.

[0005] The technical solution to achieve the objective of this invention is:

[0006] A demodulation method for a multi-wavelength wireless optical communication system based on compressed sensing includes two parts: a transmitting end process and a receiving end process.

[0007] The sending process includes:

[0008] The transmitting end uses the combination of different laser states to carry information, and modulates the information onto multiple lasers with different wavelength specifications after mapping.

[0009] The code elements and pilot signals are combined and then emitted through a laser.

[0010] The output light from these multiple lasers is coupled into a single laser beam using a wavelength division multiplexer.

[0011] The laser signal output from the wavelength division multiplexer is amplified by an optical amplifier and then emitted through a collimating lens.

[0012] The mapping and modulation of information includes the following steps:

[0013] Step A101: The transmitting end uses the combination states of different lasers to carry information;

[0014] Suppose there are M lasers, and N of them are turned on when transmitting one symbol, there are a total of Given a combination of states, each symbol can transmit... bits of data, of which Indicates rounding down;

[0015] Step A102, let To reduce the impact of coherence between beams of different wavelengths, from From the combination of lasers, two are selected based on the principle of maximizing the wavelength spacing of the chosen lasers. B 1. Establish a mapping f such that the selected 2 B Each of the combined states of the lasers corresponds to a string of B bits of data, and the correspondence is one-to-one.

[0016] Step A103: Split the original data into groups of 1 bit each, with each group forming a symbol, and then perform the inverse mapping f using the established mapping f. -1 This code element is emitted via a laser;

[0017] The step of combining the code element and pilot signal and then emitting it through a laser includes the following steps:

[0018] Step A201: Combine several groups of symbols and several pilot signals into a frame, and the matrix P formed by the pilot signal combination... h Must meet Reversible; the pilot signal is used to estimate the optical-to-electrical conversion gain H, and the estimated H is used for signal demodulation calculation;

[0019] Step A202: Obtain the coherence time of the channel and sensor gain through simulation calculation, and find the minimum of the two as the minimum coherence time. During the minimum coherence time, the channel and sensor gain are considered to change steadily, and the frame composed of pilot and symbol is transmitted within the minimum coherence time.

[0020] The receiving end process includes:

[0021] Use a lens to focus the light;

[0022] The light focused by the lens is shone onto the PIN photoelectric sensor array, and the laser signal is received by the rectangular photoelectric sensor array.

[0023] The output of the photoelectric sensor array is sampled at the same transmission frequency as the transmitter to obtain a frame of signal sampling data;

[0024] The photoelectric conversion gain matrix is ​​estimated using the received pilot sampling data;

[0025] The original data is demodulated using the estimated photoelectric conversion gain matrix.

[0026] The receiving end's processing flow includes the following steps:

[0027] Step B101: A rectangular photoelectric sensor array is used to receive the laser signal. The number of sensors U is much smaller than the number of lasers M, and the following conditions must be met. Where c is a constant related to M and N. The value of c is determined through simulation experiments. The output signal of the photoelectric sensor array is sampled to obtain the sampled data of a frame of signal.

[0028] Step B102: The combined state of the laser at the transmitting end is represented by a vector x. Each element in vector x is either 0 or 1. A value of 1 indicates that the corresponding laser is on, and a value of 0 indicates that the corresponding laser is off. The output of each sensor is regarded as the weighted sum of all elements in vector x and the superposition of Gaussian noise. Each weight is the total gain of its corresponding wavelength input and output.

[0029] Step B103: After representing the outputs of all sensors using the method described in step B102, the result is written in matrix form as follows:

[0030] D = HP + ε

[0031] Where D is the sampled data, H is the gain matrix, P is the combined state vector of the transmitting laser, and ε is Gaussian noise; after representing the pilot signal output using the method in step B102, it can be written in matrix form as follows:

[0032] D h =HP h +ε h

[0033] Where D h Here are the sampled data for the pilot signal section, where H is the gain matrix and P is the sampled data. h Let ε be the combined state vector of the transmitting laser corresponding to the pilot signal. h Gaussian noise generated by the sensor array receiving pilot signals;

[0034] The total gain matrix H is estimated by solving this equation;

[0035] Step B104: Write the input and output of the data signal in the form D = HP + ε. Due to the sparsity of the column vectors in matrix P, the equation conforms to the compressed sensing model. Therefore, the equation can be solved using the reconstruction algorithm for solving the compressed sensing model. Substitute the estimated H into the equation to obtain P = P′.

[0036] Step B105: Compare each element estimated in P′ with a threshold to obtain a new matrix. If the elements in P′ are greater than this threshold, then The corresponding element is 1 if it is not in the range, and 0 otherwise. This threshold is selected through simulation experiments.

[0037] Step B106, will Demodulation of the signal can be achieved by mapping f to data information.

[0038] Advantages or beneficial effects of the present invention:

[0039] This invention provides an effective and easily implemented demodulation method for a multi-wavelength wireless optical communication system. The method utilizes a photoelectric sensor array to receive multi-wavelength wireless optical signals, constructs a compressed sensing model at the receiving end, solves the model, and demodulates the signal by processing the solution results. This method avoids the use of fiber optic coupling schemes, which are difficult to deploy and costly, thus saving significant resources. Attached Figure Description

[0040] Figure 1 This is a flowchart of a demodulation method for a multi-wavelength wireless optical communication system based on compressed sensing, as described in an embodiment of the present invention.

[0041] Figure 2 The figure shows the simulation results of the relationship between bit error rate and signal-to-noise ratio under different light intensity fluctuation variances in the example. Detailed Implementation

[0042] The present invention will now be described in detail with reference to the accompanying drawings and embodiments. Examples of the embodiments are shown in the accompanying drawings, and the embodiments described in the drawings are exemplary and intended to explain the present invention, but should not be construed as limiting the present invention.

[0043] Example:

[0044] like Figure 1 As shown, the demodulation method for a multi-wavelength wireless optical communication system based on compressed sensing includes two parts: the transmitting end process and the receiving end process.

[0045] The sending process includes:

[0046] The information is mapped and modulated onto multiple DFB lasers with different wavelength specifications;

[0047] The code elements and pilot signals are combined and then emitted through a laser.

[0048] The output light from these multiple lasers is coupled into a single laser beam using a wavelength division multiplexer.

[0049] The laser signal output from the wavelength division multiplexer is amplified by the EDFA optical amplifier and then emitted through a collimating lens.

[0050] The mapping and modulation of information includes the following steps:

[0051] Step a101: The transmitting end uses the combination states of different lasers to carry information;

[0052] There are 20 lasers at the transmitting end. Eight of these lasers are activated when transmitting one symbol, for a total of... Given these combinations of states, the number of bits that each symbol can transmit is:

[0053]

[0054] in Indicates rounding down;

[0055] Step a102, in order to reduce the influence of coherence between beams of different wavelengths, from From the combination of lasers, two are selected based on the principle of maximizing the wavelength spacing of the chosen lasers. 16 1. Establish a mapping f such that the selected 2 16 Each of the combined states of the lasers corresponds to a 16-bit data string, and the correspondence is one-to-one.

[0056] Step a103: Split the original data into groups of 16 bits each, with each group forming a code element, and then perform the inverse mapping f using the established mapping f. -1 This code element is emitted through a laser.

[0057] The step of combining the code element and pilot signal and then emitting it through a laser includes the following steps:

[0058] Step a201: Combine several groups of symbols and several pilot signals into a frame, and the matrix P formed by the pilot signal combination... h Must meet Reversible, pilot signals are used to estimate the optical-to-electrical conversion gain;

[0059] Step a202: The coherence time of the channel and sensor gain is obtained by simulation calculation, and the minimum of the two is the minimum coherence time. The channel and sensor gain are considered to change steadily within the minimum coherence time. The frame composed of pilot and symbol is transmitted within the minimum coherence time.

[0060] In this embodiment, M=20, N=9, and one frame has 20 pilots and 10 symbols. After combining the data and pilots, one frame P0 is:

[0061]

[0062] The receiving end process includes:

[0063] Use a lens to focus the light;

[0064] The light focused by the lens is shone onto the PIN photoelectric sensor array, and the laser signal is received by the rectangular photoelectric sensor array.

[0065] The output of the photoelectric sensor array is sampled at the same transmission frequency as the transmitter to obtain a frame of signal sampling data;

[0066] The photoelectric conversion gain matrix is ​​estimated using the received pilot sampling data;

[0067] The original data is demodulated using the estimated photoelectric conversion gain matrix.

[0068] The receiving end's processing flow includes the following steps:

[0069] Step b101: A rectangular photoelectric sensor array is used to receive the laser signal. The number of sensors is 16. The output signal of the photoelectric sensor array is sampled to obtain the sampled data of one frame of signal.

[0070] Step b102: The combined state of the laser at the transmitting end is represented by a vector x. Each element in vector x is either 0 or 1. A value of 1 indicates that the corresponding laser is on, and a value of 0 indicates that the corresponding laser is off. The output of each sensor is regarded as the weighted sum of all elements in vector x and the superposition of Gaussian noise. Each weight is the total gain of its corresponding wavelength input and output.

[0071] Step b103: After representing the outputs of all sensors using the method in step b102, the result is written in matrix form as follows:

[0072] D = HP + ε

[0073] Where D is the sampled data, H is the gain matrix, P is the combined state vector of the laser at the transmitting end, and ε is Gaussian noise;

[0074] In this embodiment, after the frame signal P0 passes through the channel, the received 16-row, 30-column sampled data D0 is:

[0075]

[0076] After representing the pilot signal output using the method in step two, it can be written in matrix form as follows:

[0077] D h =HP h +ε h

[0078] Where D h Here are the sampled pilot signal data, H is the gain matrix, and P... h Let ε be the combined state vector of the transmitting laser corresponding to the pilot signal. h Gaussian noise generated by the sensor array receiving pilot signals;

[0079] In this embodiment, after the frame signal P0 passes through the channel, it receives 16 rows and 30 columns of sampled pilot data D. h0 for:

[0080]

[0081] The total gain matrix H is estimated by solving this equation;

[0082] This embodiment minimizes the cost function. The estimate of H is obtained as follows:

[0083]

[0084] In this embodiment, the estimated value of H obtained from the frame signal P0 is:

[0085]

[0086] Step b104: Write the input and output of the data signal in the form of D = HP + ε. Due to the sparsity of the column vectors in matrix P, the equation conforms to the compressed sensing model. Therefore, the equation can be solved using the reconstruction algorithm for solving the compressed sensing model. Substitute the estimated H into the equation to obtain P = P′.

[0087] In this embodiment, the OMP algorithm is used to solve the compressed sensing model, and the preliminary estimate P0′ of the 10 symbols in the frame signal P0 is obtained as follows:

[0088]

[0089] Step b105: Compare each element in the estimated P with a threshold to obtain a new matrix. If the number of elements in P is greater than this threshold, then The corresponding element is 1 if it is not in the range, and 0 otherwise. This threshold is selected through simulation experiments.

[0090] In this embodiment, a threshold of 0.5 is selected. Processing P0′ at this threshold yields an estimate of ten symbols out of ten symbols in the frame signal P0. for:

[0091]

[0092] Step b106, will Demodulation of the signal can be achieved by mapping f to data information.

[0093] Numerical simulations of the system were performed using simulation software. A Gamma-Gamma channel model was adopted, and the simulation covered the transmission of 50,000 symbols. The relationship between the system's channel ratio and bit error rate was analyzed under different light intensity fluctuation variances. The light intensity fluctuation variance describes the strength of turbulence in the Gamma-Gamma channel; a larger variance indicates stronger turbulence. Simulation results are as follows: Figure 2 As shown, from Figure 2 As can be seen, this scheme can still achieve an acceptable bit error rate when the light intensity fluctuation variance σ = 0.6.

[0094] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A demodulation method for a multi-wavelength wireless optical communication system based on compressed sensing, characterized in that, It includes two parts: the sending process and the receiving process. The sending process includes: The transmitting end uses the combination of different laser states to carry information, and modulates the information onto multiple lasers with different wavelength specifications after mapping. The code elements and pilot signals are combined and then emitted through a laser. The output light from these multiple lasers is coupled into a single laser beam using a wavelength division multiplexer. The laser signal output from the wavelength division multiplexer is amplified by an optical amplifier and then emitted through a collimating lens. The receiving end process includes: Use a lens to focus the light; The light focused by the lens is shone onto the PIN photoelectric sensor array, and the laser signal is received by the rectangular photoelectric sensor array. The output of the photoelectric sensor array is sampled at the same transmission frequency as the transmitter to obtain a frame of signal sampling data; The photoelectric conversion gain matrix is ​​estimated using the received pilot sampling data; The original data is demodulated using the estimated photoelectric conversion gain matrix; The sending process includes the following steps for mapping and modulating information: Step A101: The transmitting end uses the combination states of different lasers to carry information; Suppose there are M lasers, and N of them are turned on when transmitting one symbol, there are a total of In this combination of states, each symbol is transmitted bits of data, of which Indicates rounding down; Step A102, let In order to reduce the influence of coherence between beams of different wavelengths, from From the combination of lasers, the selection should be based on the principle of maximizing the wavelength spacing of the selected lasers. Let f be a set of elements, such that the selected elements are... Each of the combined states of the lasers corresponds to a string of B bits of data, and the correspondence is one-to-one. Step A103: Split the original data into groups of 1B bits each, with each group forming a symbol, and then perform the inverse mapping of the established mapping f. This code element is emitted through a laser.

2. The demodulation method for a multi-wavelength wireless optical communication system based on compressed sensing according to claim 1, characterized in that, In the transmitting process, the step of combining the symbol and pilot signal and transmitting it through a laser includes the following steps: Step A201: Combine several groups of symbols and several pilot signals into a frame, and the matrix formed by the pilot signal combination... Must meet reversible; Pilot signals are used to estimate the optical-to-electrical conversion gain H, and the estimated H is used for signal demodulation calculations. Step A202: Obtain the coherence time of the channel and sensor gain through simulation calculation, and find the minimum of the two as the minimum coherence time. During the minimum coherence time, the channel and sensor gain are considered to change steadily, and the frame composed of pilot and symbol is transmitted within the minimum coherence time.

3. The demodulation method for a multi-wavelength wireless optical communication system based on compressed sensing according to claim 1, characterized in that, The receiving end's processing flow includes the following steps: Step B101: Use a rectangular photoelectric sensor array to receive the laser signal. The number of sensors U is less than the number of lasers M, and it must meet the following requirements. , where c is a constant related to M and N. The value of c is determined through simulation experiments. The output signal of the photoelectric sensor array is sampled to obtain the sampled data of a frame of signal. Step B102: The combined state of the laser at the transmitting end is represented by a vector x. Each element in vector x is either 0 or 1. A value of 1 indicates that the corresponding laser is on, and a value of 0 indicates that the corresponding laser is off. The output of each sensor is regarded as the weighted sum of all elements in vector x and the superposition of Gaussian noise. Each weight is the total gain of its corresponding wavelength input and output. Step B103: After representing the outputs of all sensors using the method described in step B102, the result is written in matrix form as follows: D=HP+ε Where D is the sampled data, H is the gain matrix, P is the combined state vector of the laser at the transmitting end, and ε is Gaussian noise; After representing the pilot signal output using the method in step B102, it can be written in matrix form as follows: D h =HP h +ε h Where D h Here are the sampled pilot signal data, H is the gain matrix, and P... h Let ε be the combined state vector of the transmitting laser corresponding to the pilot signal. h Gaussian noise generated by the sensor array receiving pilot signals; The total gain matrix H is estimated by solving this equation; Step B104, write the input and output of the data signal as Since the column vectors in matrix P are sparsity-dependent, this equation conforms to the compressed sensing model. Therefore, it can be solved using a reconstruction algorithm for solving the compressed sensing model. Substituting the estimated H into the equation yields the solution. ; Step B105 will estimate Each element in the matrix is ​​compared with a threshold to obtain a new matrix. ,if If the element in the array is greater than this threshold, then The corresponding element is 1 if it is not in the range, and 0 otherwise. This threshold is selected through simulation experiments. Step B106, will Demodulation of the signal can be achieved by mapping f to data information.

Images