An OSIC detection method for underwater optical Massive MIMO communication system under transmission and reception misalignment

Abstract

The present invention discloses an OSIC detection method for an underwater optical Massive MIMO communication system with misaligned transmission and reception. After estimating the gain of each sub-channel using the LS channel estimation algorithm, the offset direction of the transceiver can be determined based on the magnitude of each DC gain in the channel gain matrix. The detection column order and the order within the column are then determined based on the offset direction, and interference elimination is performed to complete the detection of each signal. The present invention sorts according to the offset direction, preferentially detecting signals without interference, and then detecting signals with interference, greatly reducing the error propagation caused by the sorting, making the subsequent stage more accurate in subtracting the detected signal, and making the remaining signal used in the subsequent stage have less interference, thereby improving the bit error rate performance, so as to solve the problem of increased bit error rate of the UWOC system caused by misalignment of the transceiver and offset and diffusion of the imaging spot in the underwater imaging optical Massive MIMO communication system.

Description

Technical Field

[0001] The present invention relates to the technical field of underwater optical Massive MIMO communication systems, and in particular to an OSIC (Ordered Successive Interference Cancellation) detection method under misaligned transmission and reception in an underwater optical Massive MIMO communication system. Background Art

[0002] my country boasts a vast ocean area. Activities such as underwater disaster warning, resource exploration, and environmental and pollution monitoring require the use of underwater communication technologies to transmit data to the surface in real time or near real time, and then forward it to shore stations or satellites. However, both underwater acoustic and radio frequency communications, currently suitable for long-distance underwater communication, have shortcomings. Acoustic communication suffers from narrow bandwidth and high latency, while radio frequency communication suffers from rapid propagation attenuation. Given that the underwater environment attenuates blue-green light with wavelengths between 450nm and 550nm relatively little, underwater wireless optical communication technologies based on the blue-green bands could serve as a powerful supplement to underwater communications. With the increasing development of underwater data transmission, MIMO systems based on light source arrays can fully utilize spatial division multiplexing gain, improving system capacity and anti-interference capabilities. However, due to the rapidly growing demand for underwater transmission, traditional small-scale optical MIMO communication systems are no longer sufficient to meet the needs of ocean information transmission.

[0003] In order to expand the channel capacity, improve the bit error rate performance of the communication system, and increase the transmission distance of the communication system, it is considered to introduce large-scale MIMO communication technology into the underwater optical communication system, such as Figure 1 As shown. However, changes in the underwater communication link environment (such as the movement of the transmitter and / or receiver caused by the underwater vehicle's autonomy, ocean currents and other turbulence sources, and the change of the refractive index of underwater materials due to water depth, temperature and salinity) will cause the link of underwater wireless optical communication to be misaligned, which in turn causes the underwater optical Massive MIMO communication system to be misaligned in transmission and reception. Under the misalignment of the underwater optical Massive MIMO communication system in transmission and reception, the separated light spots formed by the imaging lens group on the detection surface will not be able to fall accurately on the detector array, and relative horizontal and / or horizontal offset will occur (such as Figure 2 As shown in Figure 3, the interference between optical paths becomes larger, the correlation between sub-channels becomes stronger, and the signal detection becomes difficult and complicated, which in turn increases the bit error rate of the underwater wireless optical communication system.

[0004] Because underwater optical Massive MIMO communication systems suffer from significantly more severe transmission and reception misalignment than conventional MIMO communication systems, conventional MIMO signal detection algorithms cannot successfully detect signals in these systems. For example, the traditional OSIC detection algorithm used in conventional MIMO communication systems prioritizes detection based on the power of received signals and then performs signal demodulation based on the detection order. While this algorithm can successfully detect signals in conventional MIMO communication systems with minimal transmission and reception misalignment and a relatively small MIMO scale, it does so in underwater optical Massive MIMO communication systems with significant transmission and reception misalignment and a very large MIMO scale. The receiving detector will simultaneously receive multiple light spots, resulting in the detector with the highest received optical power also receiving the greatest interference between light beams. Therefore, the traditional OSIC detection algorithm cannot be applied to underwater optical Massive MIMO communication systems. Summary of the Invention

[0005] The present invention aims to solve the problem that signal detection is difficult and complicated due to misaligned transmission and reception in an underwater optical Massive MIMO communication system, and provides an OSIC detection method under misaligned transmission and reception in an underwater optical Massive MIMO communication system.

[0006] To solve the above problems, the present invention is achieved through the following technical solutions:

[0007] A method for detecting OSIC under misaligned transmission and reception in an underwater optical Massive MIMO communication system comprises the following steps:

[0008] Step 1: Use the LS channel estimation method to estimate the DC gain of each sub-channel of the received electrical signal sent by the detector array to obtain a channel gain matrix;

[0009] Step 2: Take each column of the current channel gain matrix in turn, and compare the channel gains of the subchannels in each row of the current column with the channel gains of the subchannels in each row of other columns of the current channel gain matrix. If the channel gains of the subchannels in each row of the current column are less than the channel gains of the subchannels in each row of all other columns of the current channel gain matrix, then delete the current column from the current channel gain matrix, select the current column as the detection column, and go to step 3. Otherwise, repeat step 2.

[0010] Step 3: sort the channel gains of the sub-channels in each row of the detection column from small to large, thereby obtaining the intra-column detection sorting of the sub-channels in each row of the detection column;

[0011] Step 4: For the jth selected detection column:

[0012] If j=1, then according to the intra-column detection order of the sub-channels in each row of the detection column selected for the jth time, the original signals of the sub-channels in each row are sequentially sent to the subsequent signal demodulation process for signal demodulation;

[0013] If j≠1, first subtract the product of the original signal of each row sub-channel of the detection column selected for the j-1th time and the channel gain from the original signal of each row sub-channel of the detection column selected for the jth time to obtain the residual signal of each row sub-channel of the detection column selected for the jth time; then, according to the intra-column detection order of each row sub-channel of the detection column selected for the jth time, the residual signal of each row sub-channel is sequentially sent to the subsequent signal demodulation process for signal demodulation.

[0014] In the above step 1, the LS channel estimation method inserts a pilot signal every 4 data points into each sub-channel signal, and after estimating the gain at the pilot position of the signal, linear interpolation is used to obtain the DC gain of all sub-channels.

[0015] Compared with the prior art, the present invention takes into account the serious problem of misalignment in reception and transmission caused by changes in the underwater communication link environment and the extremely large scale of MIMO in underwater optical Massive MIMO communication systems. On the basis of the traditional OSIC detection algorithm, an improved OSIC detection algorithm (I-OSIC) based on the order of minimum interference is proposed. The order of minimum interference in underwater optical Massive MIMO communication systems is difficult to obtain, and the traditional OSIC detection algorithm based on the order of received signal power does not take the interference into consideration. The detector at the receiving end will receive multiple light spots at the same time, and the signal with the largest received optical power also has the largest interference between light beams. The present invention proposes a method for channel and offset direction estimation to obtain the order of minimum interference, that is, after estimating the gain of each sub-channel using the LS channel estimation algorithm, the offset direction of the transmitting and receiving end can be determined according to the size of each DC gain in the channel gain matrix, and then the detection column order and the order within the column are determined according to the offset direction, and interference elimination is performed to complete the detection of each signal. The present invention sorts according to the offset direction, preferentially detects signals without interference, and then detects signals with interference, which greatly reduces the error propagation caused by sorting, makes the subtraction of the detected signal in the latter stage more accurate, and makes the remaining signal used in the subsequent stage have less interference, thereby improving the bit error rate performance, and solving the problem of increased bit error rate of the UWOC system caused by misalignment of the transceiver and the offset and diffusion of the imaging spot in the underwater imaging light Massive MIMO communication system. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 Schematic diagram of the optical part of the underwater optical Massive MIMO communication system;

[0017] Figure 2 Schematic diagram of the distribution of imaging spots on the detector array under misalignment of transmission and reception;

[0018] Figure 3 This is the schematic diagram of the underwater optical Massive MIMO communication system;

[0019] Figure 4 This is a comparison chart of the bit error rate performance of the present invention and the existing signal detection algorithm under the condition of small misalignment;

[0020] Figure 5 This is a comparison chart of the bit error rate performance of the present invention and the existing signal detection algorithm under the condition of large misalignment. DETAILED DESCRIPTION

[0021] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to specific examples.

[0022] In an underwater large-scale optical Massive MIMO communication system, the transmitter first modulates the binary bit data stream to be transmitted, and then uses a light source array composed of light sources to convert the modulated electrical signal into multiple transmit optical signals and send them to the receiver. At the receiver, the multiple transmit optical signals first pass through the imaging lens, forming separate light spots on the detection surface. The detector array composed of detectors converts them into multiple receive electrical signals, which are then detected and demodulated to restore them to binary bit data.

[0023] OSIC detection is based on a group of linear receivers for signal detection and interference cancellation. Each receiver detects one of the parallel data streams and can successfully subtract the detected signal component from the received signal at each stage, so that the remaining signal used in the subsequent stage has less interference. Since incorrect judgments in the previous stage will cause error propagation, the detection order will significantly affect the overall performance of OSIC detection. Considering the serious misalignment of transmission and reception in underwater optical Massive MIMO communication systems and the extremely large scale of MIMO, the present invention proposes an improved OSIC detection algorithm based on sorting with minimal interference. After using the LS channel estimation algorithm to estimate the gain of each sub-channel, the offset direction of the transmitting and receiving end can be determined according to the size of each DC gain in the channel gain matrix. Then, the detection column order and the order within the column are determined according to the offset direction, and interference cancellation is performed to complete the detection of each signal.

[0024] Specifically, the present invention proposes an OSIC detection method for an underwater optical Massive MIMO communication system under misaligned transmission and reception, which specifically includes the following steps:

[0025] Step 1: Use the LS channel estimation method to estimate the DC gain of each sub-channel of the received electrical signal y sent by the detector array to obtain a channel gain matrix.

[0026] The receiving end performs LS channel estimation based on the pilot interval during the OFDM modulation process performed by the transmitting end. In a preferred embodiment of the present invention, the pilot interval of the OFDM modulation performed by the transmitting end is 4, so the sub-channel signal received by the receiving end is a signal with a pilot inserted every 4 data. In addition, in the LS channel estimation process, a linear interpolation algorithm is first used to interpolate all OFDM symbol frequency point data to obtain the OFDM symbol frequency response, and then the OFDM symbol frequency response is statistically averaged to obtain the channel gain. According to the absorption, scattering and attenuation of underwater Massive MIMO channels and the requirement to reduce pilot overhead, the LS channel estimation method of the present invention treats each sub-channel signal as a signal with a pilot inserted every 4 data, and after performing gain estimation at the pilot position of the signal, linear interpolation is used to obtain the DC gain of all sub-channels. The channel gain matrix of LS channel estimation is:

[0027]

[0028] Where Y is the received signal, X is the pilot signal, and X H is the conjugate transposed matrix of X, X -1 is the inverse matrix of X. It is directly estimated by the pilot, so only the channel gain of the pilot insertion position can be obtained. The channel gain of other non-pilot subcarrier positions can be estimated by linear interpolation, that is, After linear interpolation, we get

[0029] Step 2: Take each column of the current channel gain matrix in turn, and compare the channel gains of the sub-channels in each row of the current column with the channel gains of the sub-channels in each row of other columns of the current channel gain matrix;

[0030] If the channel gain of each row subchannel of the current column is less than the channel gain of each row subchannel of all other columns in the current channel gain matrix, the current column is considered to be the column with the most serious offset in the current channel gain matrix. In this case, the current column is deleted from the current channel gain matrix and selected as the detection column, and the process goes to step 3.

[0031] Otherwise, it is considered that the current column is not the column with the most serious offset in the current channel gain matrix, and step 2 is repeated to reselect the column with the most serious offset in the current channel gain matrix.

[0032] Step 3: Sort the channel gains of the sub-channels in each row of the detection column from small to large, thereby obtaining the intra-column detection sorting of the sub-channels in each row of the detection column.

[0033] Step 4: For the jth selected detection column:

[0034] If j=1 (i.e. the first selected detection column), then according to the column detection order of each row of sub-channels of the jth selected detection column, the original signals of each row of sub-channels are sequentially sent to the subsequent signal demodulation process for signal demodulation. Since the first selected detection column is the column with the most serious deviation in the entire received electrical signal (such as Figure 2 ), which does not generate interference, so the original signal is directly used for signal demodulation;

[0035] If j≠1 (i.e., it is not the detection column selected for the first time), then the original signal of each row sub-channel of the detection column selected for the jth time is first subtracted from the product of the original signal of each row sub-channel of the detection column selected for the j-1th time and the channel gain to obtain the residual signal of each row sub-channel of the detection column selected for the jth time; then, according to the intra-column detection order of each row sub-channel of the detection column selected for the jth time, the residual signal of each row sub-channel is sequentially sent to the subsequent signal demodulation process for signal demodulation. Since the detection column selected subsequently is not the column with the most serious offset in the entire received electrical signal (such as Figure 2 The original signal will be affected by the adjacent column signal (such as Figure 2 Therefore, it is necessary to subtract the interference of the corresponding rows of the adjacent columns from the original signal before signal demodulation.

[0036] In the later stage of detection, the interference caused by the detected adjacent signals is subtracted from the original signal of each sub-channel, so that the subsequent receiver contains less interference in the detection stage.

[0037] The performance of the present invention is described below through a specific example.

[0038] Figure 3 The schematic diagram of the underwater optical Massive MIMO communication system includes optical and electrical parts.

[0039] (1) Optical part:

[0040] The transmitting end uses a 64×64 light source array, and the receiving end uses a 64×64 detector array. The imaging lens at the receiving end is composed of a convex lens and a concave lens. The main optical axes of the convex lens and the concave lens coincide, and the convex lens is located at the front end of the concave lens. The convex lens converges the multi-path optical signals sent from the transmitting end onto the concave lens, and the concave lens separates the light spots of the converged optical signals to obtain multi-path optical signals. The front and rear end surfaces of the convex lens and the concave lens of the imaging lens are both parabolic surfaces. By designing the paraboloids of the convex lens and the concave lens, the interference caused by imaging aberrations can be effectively reduced, making it better suitable for underwater environments. In order to achieve precise adjustment of the lens surface while simplifying the lens design, the curvature radius and the cone coefficient of the lens parabola are set to zero while considering the first-order and second-order coefficients of the lens parabola. The surface vector height expression of the simplified lens parabola is as follows:

[0041] z=αr 2 +βr 4

[0042] Where z is the sag of the lens parabola, r is the radial coordinate of the axisymmetric lens surface, α is the first-order coefficient of the lens parabola, and β is the second-order coefficient of the lens parabola. The imaging lens is optimized by the above formula. By increasing the focal length F and the lens diameter D of the lens, a higher optical gain is obtained while reducing the correlation of the system channel gain matrix. In this embodiment, the first-order coefficient and the second-order coefficient of the parabola of the front face of the convex lens are α = 0.02, β = 5 × 10 -6 The first-order coefficient and second-order coefficient of the rear end parabola of the convex lens are α=0.01, β=1×10 -6 The aperture (mm) of the convex lens is 30 mm and the thickness is 15 mm; the first-order coefficient and second-order coefficient of the parabola of the front face of the concave lens are α = -0.03, β = -1×10 -6 The first-order coefficient and second-order coefficient of the parabola of the rear end face of the concave lens are α=0.03, β=1×10 -6 The aperture (mm) of the concave lens is 20mm and the thickness is 2mm. Since the convex surfaces of a biconvex lens are all in the same direction, the coefficients on the front and back ends have the same sign. However, since the convex surfaces of a biconcave lens are both inward, the convex directions of the front and back ends are opposite, so the coefficients on the front and back ends of the concave lens have opposite signs.

[0043] (2) Electrical part:

[0044] The signal modulation process at the transmitter includes serial-to-parallel conversion, 4QAM mapping, Hermitian symmetry, IFFT, CP addition, parallel-to-serial conversion, and wave elimination. The signal demodulation process at the receiver includes serial-to-parallel conversion, CP removal, FFT, data subcarrier extraction, 4QAM demodulation, and parallel-to-serial conversion. Furthermore, the transmitter performs analog-to-digital conversion on the modulated signal before sending it to the light source array. The receiver performs digital-to-analog conversion on the signal received by the detector array before performing signal detection.

[0045] At the transmitting end, the binary bit stream undergoes serial-to-parallel conversion, 4QAM mapping, Hermitian symmetry, IFFT, cyclic prefix addition, parallel-to-serial conversion, and clipping, modulating it into an ACO-OFDM signal. This signal is then converted to an optical signal through analog-to-digital conversion and then fed into a light source array for transmission. At the receiving end, a detector array receives the signal, converts it into an optical signal, and then converts it into an analog-to-digital signal. This signal is then sent to the signal detection method of the present invention for detection. The detected signal is then demodulated back to the original binary bit stream through serial-to-parallel conversion, cyclic prefix removal, FFT, data subcarrier extraction, 4QAM demodulation, and parallel-to-serial operations.

[0046] The underwater optical Massive MIMO communication method using the improved OSIC detection algorithm (I-OSIC) of the present invention has the following process:

[0047] When a binary bit stream enters the system, the signal modulation section at the transmitter first performs serial-to-parallel conversion to convert the serial bit stream into a parallel bit stream, applying 4QAM modulation to each parallel signal. After Hermitian symmetry and IFFT operations, the 4QAM constellation corresponding to each data channel is mapped to real numbers. This is then converted to an OFDM signal through digital-to-analog conversion. A cyclic prefix is ​​added to each signal before parallel-to-serial conversion. To ensure that the subsequent optical signal is suitable for underwater channel transmission, the signal processing section also performs clipping. After clipping, the signal undergoes analog-to-digital conversion and is then sent to a 64×64 light source array, which converts the electrical signal into an optical signal for transmission. At the receiver, a 64×64 detector array receives the optical signal and digitizes it using analog-to-digital conversion. The digital signal is then transmitted to the LS channel estimation section. The LS channel estimation section performs LS channel estimation based on the pilot signal inserted by the transmitter. The resulting channel gain matrix is ​​then transmitted to the subsequent OSIC signal detection section. The OSIC signal detection section calculates the DC gain of each column in the channel gain matrix and correlates misalignment with DC gain. If the overall channel gain of a column is smaller than that of the other columns, and the values ​​within the columns are similar, this column is identified as the first column to be detected. Specifically, the first column in the direction opposite to the light spot offset is the first column to be detected. Within the first column, the first signal is selected from the top to the bottom as the first signal to be detected. Within this column, interference cancellation is performed sequentially and then detected. The second to last columns are then determined in the direction of the light source offset, thus determining the detection order for OSIC signal detection. The OSIC signal detection section utilizes a set of linear receivers, each detecting one of the parallel data streams. After detection, the detected data is subtracted from the received data, reducing interference in the subsequent detection stages. This method results in an OSIC-detected received signal. After OSIC signal detection, the receiving end uses the opposite operations of the transmitting end, namely serial / parallel conversion, CP removal, FFT, data subcarrier extraction, 4QAM demodulation, and parallel / serial conversion to demodulate the modulated signal and restore it to the original binary bit stream.

[0048] The relative offset between the transmitting and receiving ends of the underwater optical Massive MIMO communication system can be divided into eight directions. Taking a single rightward offset as an example, the algorithm proposed in this scenario can also be applied to systems with single offset directions such as upward, downward, and left. A specific comparison of the bit error rate performance of the OSIC detection method of the present invention and the existing algorithm under misalignment conditions can be seen. Figure 4 and Figure 5 . Figure 4 This is a simulation diagram of the system bit error rate when the horizontal offset error is small. Figure 5This is a simulation diagram of the system bit error rate with a large horizontal offset error. As the relative offset error between the transmitting and receiving ends increases, the channel correlation also increases. At this time, the SVD precoding detection algorithm has limited improvement on the system bit error performance in a poor channel environment. The ZF detection algorithm will weight and amplify the additive noise. The MMSE detection algorithm is improved on the basis of the ZF detection algorithm and requires noise variance estimation, but the improvement on the system bit error rate is also relatively limited, especially in the case of large offsets. These shortcomings also make the ZF detection algorithm and the MMSE detection algorithm have lower bit error rate performance than the OSIC detection algorithm under the same signal-to-noise ratio environment. Compared with the traditional OSIC algorithm (OSIC) based on the power sorting of the receiving end, the improved OSIC detection algorithm (I-OSIC) based on the interference minimum sorting of the present invention estimates the light spot offset direction according to the channel gain matrix, selects the first column in the opposite direction of the offset direction as the first column to be detected, so that there is no interference from other light spots when detecting the first column. After subtracting the detected signal, the column with interference is detected, and the bit error rate performance is improved under the same signal-to-noise ratio conditions. In addition, the time complexity of the SVD detection algorithm is O(min(m 2 n,mn 2 ), the time complexity of ZF detection algorithm and MMSE detection algorithm is O(k 3 ), the time complexity of the traditional OSIC detection algorithm is O(k 2 ), where k represents the number of transmitting antennas, and m and n represent the number of rows and columns of the matrix, respectively. Since the present invention is improved on the basis of the traditional OSIC detection algorithm, its time complexity is also relatively small.

[0049] In summary, the OSIC detection method under the misalignment of transmission and reception of the underwater optical Massive MIMO communication system proposed in the present invention combines the application of channel estimation, channel coding, signal detection and other technologies to meet the needs of improving the communication rate and bit error rate indicators of the UWOC system under the conditions of misalignment of the transmitting and receiving ends of the underwater imaging optical system, enhanced signal interference caused by imaging spot offset and diffusion, and misalignment of the communication link. At the same time, it avoids the problems of the existing technical solutions being greatly affected by the misalignment of the light source and the poor bit error rate performance after misalignment.

[0050] It should be noted that although the embodiments of the present invention described above are illustrative, they are not intended to limit the present invention. Therefore, the present invention is not limited to the above-mentioned specific embodiments. Without departing from the principles of the present invention, any other embodiments obtained by those skilled in the art under the guidance of the present invention are deemed to be within the protection of the present invention.

Claims

1. A method for detecting OSIC under misaligned transmission and reception in an underwater optical Massive MIMO communication system, characterized by: The steps are as follows: Step 1: Use the LS channel estimation method to estimate the DC gain of each sub-channel of the received electrical signal sent by the detector array to obtain a channel gain matrix; Step 2: Take each column of the current channel gain matrix in turn, and compare the channel gains of the subchannels in each row of the current column with the channel gains of the subchannels in each row of other columns of the current channel gain matrix. If the channel gains of the subchannels in each row of the current column are less than the channel gains of the subchannels in each row of all other columns of the current channel gain matrix, then delete the current column from the current channel gain matrix, select the current column as the detection column, and go to step 3. Otherwise, repeat step 2; Step 3: sort the channel gains of the sub-channels in each row of the detection column from small to large, thereby obtaining the intra-column detection sorting of the sub-channels in each row of the detection column; Step 4: For the jth selected detection column: If j=1, then according to the intra-column detection order of the sub-channels in each row of the detection column selected for the jth time, the original signals of the sub-channels in each row are sequentially sent to the subsequent signal demodulation process for signal demodulation; If j≠1, first subtract the product of the original signal of each row sub-channel of the detection column selected for the j-1th time and the channel gain from the original signal of each row sub-channel of the detection column selected for the jth time to obtain the residual signal of each row sub-channel of the detection column selected for the jth time; then, according to the intra-column detection order of each row sub-channel of the detection column selected for the jth time, the residual signal of each row sub-channel is sequentially sent to the subsequent signal demodulation process for signal demodulation.

2. The method for detecting OSIC under misaligned transmission and reception in an underwater optical Massive MIMO communication system according to claim 1, wherein: The LS channel estimation method in step 1 inserts a pilot signal every 4 data points into each sub-channel signal, and after estimating the gain at the pilot position of the signal, linear interpolation is used to obtain the DC gain of all sub-channels.

Images