MIMO precoding-based mine ultraviolet non-line-of-sight link enhancement method

Abstract

The application discloses a mine ultraviolet non-line-of-sight link enhancement method based on MIMO precoding and belongs to the technical field of underground coal mine communication. The mine ultraviolet non-line-of-sight MIMO channel comprehensive information is acquired, the spatial subchannel correlation is calculated, and the layout of a transmitting-receiving array is determined; according to the spatial subchannel correlation and the achievable rate, a constellation set is determined, and different constellation diagrams are allocated to different spatial multiplexing data streams; data symbols are generated based on the constellation diagrams, the time reversal frequency domain precoding weight is generated according to the ultraviolet non-line-of-sight channel impulse response between the units of the transmitting-receiving array, and the precoded signal is mapped to the transmitting array; the spatial subchannel is estimated and the data symbols are detected at the receiving end, the time reversal frequency domain precoding weight, the transmitting end precoding parameter and the receiving end precoding parameter are jointly updated under the ultraviolet power and human exposure safety constraints according to the estimation error, the achievable rate and the outage probability, and are used for the next frame transmission. The steps are simple, convenient to use and have wide practicability.

Description

Technical Field

[0001] In the field of underground communication technology in coal mines, this specifically relates to the application and link enhancement optimization of MIMO precoding technology in ultraviolet light communication. Background Technology

[0002] In the confined spaces of underground coal mines, ultraviolet (UV) communication has become a key candidate technology for emergency communication and sensor data transmission due to its advantages such as resistance to electromagnetic interference and the elimination of complex wiring. Currently, existing UV communication technologies used underground primarily rely on non-line-of-sight propagation modes, i.e., signal transmission is achieved through the reflection and scattering of UV light.

[0003] However, the confined space and complex working conditions in coal mines, along with factors such as numerous equipment obstructions and random air turbulence, result in complex and variable ultraviolet light propagation paths. On the one hand, complex paths lead to severe path loss, causing a significant attenuation of signal strength; on the other hand, multipath propagation generates inter-symbol interference and pulse delay spread, directly resulting in limited bandwidth and low transmission rates for ultraviolet communication systems, making it difficult to meet the needs of high-definition video monitoring and real-time transmission of large-capacity data underground.

[0004] Existing optimization schemes for ultraviolet (UV) communication, such as power adjustment and simple modulation method improvements for traditional SISO (Single-Input Single-Output) systems, cannot effectively solve the problems of strong channel correlation and poor spatial diversity and multiplexing performance caused by multipath propagation in underground coal mines. While some attempts at applying MIMO (Multiple-Input Multiple-Output) technology to UV communication can improve transmission capacity through multi-antenna arrays, they do not fully consider the special constraints of underground coal mines: First, large-size optical transceiver arrays are unsuitable for deployment underground, and the limited range of array element spacing and angle adjustment leads to excessively high spatial subchannel correlation and a reduction in the number of independent data substreams; second, there is a lack of robust precoding designs for the time-varying and space-varying environments underground, resulting in large channel estimation errors and a high risk of communication interruptions; third, the symbol detection accuracy of complex superimposed constellations under weak light conditions is insufficient, further restricting system transmission performance. The core reason for these problems is that existing technologies have not incorporated the characteristics of the underground coal mine environment to construct suitable channel models, precoding mechanisms, and detection algorithms, resulting in the coverage and transmission stability of UV non-line-of-sight links failing to meet practical application requirements. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this paper proposes a method for enhancing non-line-of-sight ultraviolet (UV) links in mines based on MIMO precoding. This method aims to solve problems such as severe path loss, significant inter-symbol interference, strong channel correlation, low transmission rate, and poor stability in existing UV communication in coal mines.

[0006] To achieve the above technical objectives, this invention discloses a method for enhancing non-line-of-sight ultraviolet light links in mines based on MIMO precoding, characterized by the following steps: S1. Acquire measurement data, environmental parameters and array constraints of the mine ultraviolet non-line-of-sight MIMO channel, calculate spatial sub-channel correlation, and determine the transceiver array layout within the adjustable range of the array. S2. Determine the constellation diagram set based on spatial sub-channel correlation and reachability rate, and assign different constellation diagrams to different spatial multiplexed data streams; S3. Generate data symbols based on the constellation diagram, generate time-inverted frequency domain precoding weights according to the ultraviolet non-line-of-sight channel impulse response between transceiver array units, and map the precoded signal to the transmitting array. S4. The receiver estimates the spatial sub-channel and detects data symbols. Based on the estimation error, achievable rate, and interruption probability, and under the constraints of ultraviolet light power and human exposure safety, it jointly updates the time-inverted frequency domain precoding weights, the transmitter precoding parameters, and the receiver precoding parameters for the next frame transmission.

[0007] Furthermore, the specific steps of S1 are as follows: S1.1 Initialize the mine ultraviolet non-line-of-sight MIMO system. The mine ultraviolet non-line-of-sight MIMO system includes a transmitter and a receiver. The transmitter includes multiple ultraviolet light transmitting array units, and the receiver includes multiple ultraviolet light receiving array units. The transmitter and receiver transmit signals through the mine ultraviolet non-line-of-sight channel. S1.2 Initialize the transmitter and receiver parameters of the mine ultraviolet non-line-of-sight MIMO system, and set initial values ​​for the time-reversed frequency domain precoding weight, transmitter precoding parameters, and receiver precoding parameters; wherein, the initial value is a preset initial value or the optimization result of the previous transmitted frame; S1.3 Collect current environmental parameters in the coal mine, measurement data of the mine's ultraviolet non-line-of-sight MIMO channel, and transceiver array parameters. The current environmental parameters include at least one of turbulence intensity, number of turbulence, turbulence size, object reflection intensity, reflection order, and number of reflecting objects. The transceiver array parameters include at least one of the following: geometry of ultraviolet light transmitting and receiving array units, array unit spacing, spatial orientation angle, beam divergence angle, and receiving field of view. S1.4 Calculate the spatial sub-channel correlation between each transmit array unit and each receive array unit under different transmit / receive orientation combinations based on channel measurement data, obtain the spatial sub-channel correlation corresponding to the current transmit / receive array layout, and obtain the average value, maximum value and median based on the spatial sub-channel correlation to characterize the spatial sub-channel correlation level under the corresponding transmit / receive orientation combination. S1.5 Input the current environmental parameters, transceiver array parameters, and candidate transceiver array layout parameters into the spatial sub-channel correlation prediction model to obtain the predicted spatial sub-channel correlation index corresponding to each candidate transceiver array layout parameter; wherein, the spatial sub-channel correlation prediction model is a regression prediction model with the current environmental parameters, transceiver array parameters, and candidate transceiver array layout parameters as input and the predicted channel correlation as output. S1.6 Within the physical adjustment range of the spacing between transmitting array elements, the spacing between receiving array elements, the spatial orientation angle of the transmitting array, and the spatial orientation angle of the receiving array, the transceiver array layout is determined from the candidate transceiver array layout parameters with the goal of predicting channel correlation below a preset threshold.

[0008] Furthermore, the specific steps of S2 are as follows: S2.1 In the confined space of a coal mine, the spatial orientation angle of the ultraviolet light transmitting array unit and the ultraviolet light receiving array unit at the transmitting and receiving ends is limited by the physical adjustment range. Based on the spatial sub-channel correlation and achievable rate under the transceiver array layout, the transmitting end determines a set of constellation diagrams for different spatial multiplexed data streams. The set of constellation diagrams includes multiple constellation diagrams respectively assigned to different spatial multiplexed data streams. The different constellation diagrams differ in at least one of the following: the number of constellation points, the position of constellation points, the transmission probability of constellation points, the relative offset between different constellation diagrams, and the rotation angle. S2.2 The transmitting end performs constellation mapping on different spatially multiplexed data streams according to the constellation diagram set, so that the receiving end can distinguish the signal source of different spatially multiplexed data streams based on the differences between different constellation diagrams, in order to compensate for the insufficient data stream differentiation capability caused by the inability to fully reduce the spatial sub-channel correlation by simply adjusting the layout of the transceiver array.

[0009] Furthermore, the specific steps of S3 are as follows: S3.1 The sending end converts the service data to be sent into multiple spatial multiplexed data streams, and according to the constellation diagram set, assigns different constellation diagrams to different spatial multiplexed data streams for constellation mapping, and generates data symbols corresponding to each spatial multiplexed data stream; S3.2. For the received signal of the ultraviolet non-line-of-sight channel in the mine, which is formed by the reflection path of macroscopic objects and the scattering path of microscopic particles, resulting in problems such as weak light reception, multipath propagation, pulse delay spread and inter-symbol interference, the transmitting end performs frequency domain spread on the data symbols generated in S3.1 after constellation mapping and before time-reversed frequency domain precoding. The reflection path of macroscopic objects includes roadway walls, coal walls, rock walls and equipment surfaces, and the scattering path of microscopic particles includes dust, water mist and gas molecules. S3.3 The frequency domain spreading specifically includes: for each spatially multiplexed data stream, the transmitting end allocates the data symbols corresponding to the spatially multiplexed data stream to the frequency domain positions of multiple OFDM subcarriers according to a preset subcarrier interval, to obtain the frequency-domain spread data symbols corresponding to the spatially multiplexed data stream; the frequency-domain spread data symbols are used as the input for subsequent time-inverted frequency domain precoding; S3.4 The transmitting end determines the time-reversal frequency domain precoding weights applied to the corresponding OFDM subcarriers based on the ultraviolet non-line-of-sight channel impulse response between each ultraviolet transmitting array unit and each ultraviolet receiving array unit. S3.5. The transmitting end uses the time-reversed frequency domain precoding weight to perform frequency domain precoding on the frequency-spread data symbols obtained in S3.3 according to OFDM subcarriers to obtain the frequency-domain precoded signal; wherein, the frequency domain precoding is used to enable the frequency-spread data symbols to combine the multipath propagation characteristics of the mine ultraviolet non-line-of-sight channel to form time focus and spatial focus at the target ultraviolet light receiving array unit; S3.6 The transmitting end performs optical OFDM modulation on the frequency domain precoded signal to obtain a time domain non-negative real signal that satisfies the intensity modulation direct detection constraint, and weights and maps the time domain non-negative real signal to the corresponding ultraviolet light transmitting array unit according to the transmitting end precoding parameters. S3.7 The transmitting end determines the cyclic prefix length and training sequence length according to the transmitting end parameters initialized in S1; the transmitting end adds a cyclic prefix to the time-domain non-negative real signal and inserts the training sequence to obtain the ultraviolet light signal to be transmitted; wherein, the cyclic prefix is ​​formed by copying a portion of the sampling points at the tail of each OFDM time-domain symbol in the time-domain non-negative real signal to the front of the OFDM time-domain symbol, which is used to provide a protection interval for the delay spread caused by non-line-of-sight multipath propagation of ultraviolet light in the mine and to reduce inter-symbol interference between adjacent OFDM symbols; the training sequence is a known symbol sequence predetermined by the transmitting end and the receiving end, which is used by the receiving end to estimate the spatial sub-channel between each ultraviolet light transmitting array unit and each ultraviolet light receiving array unit.

[0010] Furthermore, the specific steps of S4 are as follows: S4.1 The receiver estimates the spatial sub-channels between each ultraviolet light transmitting array unit and each ultraviolet light receiving array unit based on the training sequence, and obtains the current channel estimation result; S4.2. In view of the time-varying and spatially varying environment caused by mining processes, ventilation airflow and equipment movement in coal mines, the channel estimation error is determined by combining the current channel estimation results, historical channel measurement data and current environmental parameters; the channel estimation error is used to characterize the time reversal focus offset and spatial multiplexing data stream interference changes caused by changes in the mine's ultraviolet non-line-of-sight channel. S4.3 The receiver performs weighted merging of the signals received by multiple ultraviolet light receiving array units according to the receiver precoding parameters to obtain the received signal corresponding to each spatial multiplexed data stream; wherein, the receiver precoding parameters are used to determine the merging weight of the signals received by different ultraviolet light receiving array units. S4.4 The receiving end determines the achievable rate of each spatial multiplexed data stream and the transmission interruption probability of the mine ultraviolet non-line-of-sight MIMO link based on the channel estimation error and the received signal corresponding to each spatial multiplexed data stream. S4.5 Under the constraints of transmit power, instantaneous intensity of ultraviolet light, average intensity of ultraviolet light, safety of human exposure to ultraviolet light, achievable rate, and transmission interruption probability, with the goal of reducing transmit power and meeting the achievable rate requirements of each spatially multiplexed data stream, the time-reversed frequency domain precoding weight, the transmit end precoding parameters, and the receive end precoding parameters are used as joint optimization variables for robust optimization. Among them, robust optimization is used to coordinately adjust the time-reversed frequency domain precoding weights, transmitter precoding parameters, and receiver precoding parameters when channel estimation errors, ultraviolet light emission security restrictions, and transmission interruption probability constraints coexist. The time-reversed frequency domain precoding weights are used to adjust the time and spatial focusing of the mine ultraviolet non-line-of-sight multipath signal at the target receiving array unit. The transmitter precoding parameters are used to adjust the weighted mapping of the time-domain non-negative real signal to the ultraviolet light transmitting array unit and satisfy the ultraviolet light emission security restrictions. The receiver precoding parameters are used to adjust the diversity merging of the signals in the receiving array unit and reduce the interference between spatially multiplexed data streams and the risk of transmission interruption. S4.6. The robustly optimized time-reversed frequency domain precoding weights, transmitter precoding parameters, and receiver precoding parameters are used for mine ultraviolet non-line-of-sight MIMO link transmission in the next transmission frame or the next time slot.

[0011] Beneficial effects: (I) Optimization of Ultraviolet MIMO Channel Correlation and Cooperative Design of Array / Constellation in Coal Mines: To address the problem of excessively high correlation of MIMO spatial sub-channels (few independent data sub-streams) caused by narrow spacing and limited array size, a progressive solution of "environmental analysis - layout optimization - constellation enhancement" was invented. 1. Channel correlation modeling and prediction: Based on actual downhole channel measurement data, establish the numerical relationship between environmental parameters such as turbulence intensity, number / order of reflecting objects, and beam divergence angle and channel correlation (quantified by a specific formula). Then, use generative adversarial network (GAN) and diffusion model to train the correlation predictor, which can quickly predict the channel state under different environments. 2. Constrained Space Array Layout Optimization: Under the constraints of limited array cell spacing and angle adjustment range, reinforcement learning and particle swarm optimization algorithms are used to find the transceiver array layout that "minimizes channel correlation" without relying on large-size arrays; 3. Superimposed constellation modulation enhancement: If the correlation still does not meet the standard after layout optimization, the number, position, probability and relative offset / rotation of constellation points are adjusted, and geometric shaping and probabilistic shaping are combined to improve the receiver's ability to distinguish different signal sources, further reduce the correlation, and enhance spatial multiplexing and diversity performance.

[0012] (ii) Robust precoding mechanism for time-reversed MIMO adapted to downhole constraints: To address the challenges of time- and space-varying environments in underground mines (due to breached channel reciprocity and large estimation errors) and power / human safety limitations (with upper limits on UV exposure), a robust transmission scheme based on "channel estimation - precoding optimization - constraint adaptation" was invented. 1. Joint optimization of precoding under multiple constraints, namely time-reversed precoding matrix, transmit precoding matrix and receive precoding matrix: Under the constraints of "total transmit power limit, instantaneous / average intensity safety limit of ultraviolet light, and transmission interruption probability ≤ threshold", the time-reversed precoding matrix, transmit precoding matrix and receive precoding matrix are jointly optimized; 2. Efficient solution of non-convex problems: The Berstein inequality is used to transform the difficult-to-handle chance constraint of "interruption probability" into a deterministic constraint, and then the problem is solved by continuous convex approximation and alternating optimization algorithms, or by reinforcement learning ("environment = channel data, reward = power / rate loss, action = precoding matrix selection") to adapt to complex time-varying environments, ensuring high stable transmission under low power and high security standards. Attached Figure Description

[0013] Figure 1 This is a flowchart of the method for enhancing non-line-of-sight ultraviolet light links in mines based on MIMO precoding, as described in this invention.

[0014] Figure 2 This invention presents a model of an ultraviolet MIMO system based on time reversal and superimposed constellations. Detailed Implementation

[0015] The embodiments of the present invention will be further described below with reference to the accompanying drawings: This invention discloses a method for enhancing non-line-of-sight ultraviolet light links in mines based on MIMO precoding, as follows: This application addresses the problems in underground ultraviolet (UV) non-line-of-sight (MIMO) MIMO communication in coal mines, including strong spatial subchannel correlation, severe multipath delay spread, significant inter-symbol interference, large channel estimation errors, and limitations on UV transmission power and human safety. It proposes a link enhancement method that collaboratively designs array layout, target constellation diagram, time-reversed frequency domain precoding, and multi-constraint robust precoding. This method adapts the above processing steps to the characteristics of underground UV non-line-of-sight channels in coal mines, thereby improving the coverage, transmission reliability, and anti-interference capabilities of UV non-line-of-sight links.

[0016] 1. Environmental factors drive the determination of MIMO array layout and target constellation set; system model as follows: Figure 2 As shown, the transmitting end maps the original data stream to a new modulated point through data stream mapping, and then generates a transmit precode through serial-to-parallel conversion, frequency domain spread, time-inverse precoding, Hermitian symmetry, inverse Fourier transform, single polarization, and parallel-to-serial conversion. This transmit precode is then sent to the receiving end through the ultraviolet light channel via the CP and training sequence. The receiving end obtains the receive precode by de-CP channel estimation, and then recovers the data stream through Fourier transform, channel equalization, parallel-to-serial conversion, and symbol detection to obtain the original data stream.

[0017] Before formal transmission, the current environmental parameters in the coal mine, the parameters of the ultraviolet MIMO system, and channel measurement data are acquired. The current environmental parameters include at least one of turbulence intensity, number of turbulence, turbulence size, object reflection intensity, reflection order, and number of reflecting objects; the ultraviolet MIMO system parameters include at least one of the following: transceiver array geometry, array cell spacing, spatial orientation angle, beam divergence angle, and receiving field of view.

[0018] Based on channel measurement data, the spatial sub-channel correlation between each transmit array unit and each receive array unit under different transmit / receive orientation combinations is calculated, and the average, maximum, and median of the spatial sub-channel correlation are obtained. The aforementioned spatial sub-channel correlation is used to characterize the degree of independence between different MIMO sub-channels; the lower the correlation, the stronger the distinguishability between different spatially multiplexed data streams.

[0019] The current environmental parameters, ultraviolet MIMO system parameters, and candidate MIMO transceiver array layout parameters are input into a pre-trained spatial sub-channel correlation prediction model to obtain the predicted channel correlation corresponding to each candidate MIMO transceiver array layout parameter. Within the physical adjustment range of transceiver array cell spacing and spatial orientation angle, the candidate MIMO transceiver array layout parameters are iteratively optimized with the goal of reducing predicted channel correlation to obtain the target MIMO transceiver array layout parameters.

[0020] Specifically, let the set of candidate MIMO transceiver array layouts be: , in, Indicates the first Candidate transceiver array layouts Indicates the first Spacing between transmit array elements under candidate layouts Indicates the spacing between the receiving array elements. Indicates the orientation angle of the transmitting array space. Indicates the spatial orientation angle of the receiving array; These represent the physical adjustment range of the corresponding parameters.

[0021] For the Each candidate transceiver array layout is used to construct an input feature vector, which combines the current mine environment parameters, the ultraviolet MIMO system parameters, and the candidate transceiver array layout parameters. , in, This represents the current mine environment parameter vector. This represents the parameter vector of the current ultraviolet MIMO system. Indicates the first A vector of candidate transceiver array layout parameters. The current mine environment parameter vector is represented as: , In the formula, Indicates turbulence intensity. Indicates the amount of turbulence. Indicates turbulence size, Indicates the intensity of an object's reflection. Indicates the reflection order. This represents the number of reflecting objects. The current ultraviolet MIMO system parameter vector is represented as: , In the formula, and These represent the geometric parameters of the transmitting array and the receiving array, respectively. Indicates the beam divergence angle at the transmitting end. Indicates the field of view of the receiving end. and These represent the spatial orientation reference parameters for the transmitting array and the receiving array, respectively.

[0022] No. The vector representation of the candidate transceiver array layout parameters is as follows: , The spatial subchannel correlation prediction model is expressed as follows: , in, This represents a regression prediction model trained using historical ultraviolet non-line-of-sight MIMO channel measurement data from underground coal mines. Indicates the parameters of the prediction model. Indicates the first The predicted spatial sub-channel correlation index vector corresponding to each candidate transceiver array layout, and ,in This represents the predicted average correlation of spatial sub-channels. This indicates the predicted maximum correlation of the spatial sub-channels. This represents the median of the predicted spatial sub-channel correlation.

[0023] The training objective of the prediction model is expressed as: , in, This indicates the number of historical training samples. Indicates the first The input feature vector of each training sample This represents the real spatial sub-channel correlation index calculated based on historical channel measurement data. This represents the regularization coefficient.

[0024] Based on the predicted spatial sub-channel correlation index, construct the first... Evaluation function for candidate transceiver array layouts: , in, , , Let be the weighting coefficients, and satisfy: , The final selected target MIMO transceiver array layout is as follows: , And prioritize candidate transceiver array layouts that meet the following conditions: , in, This represents the preset spatial sub-channel correlation threshold. When no threshold is met... When selecting candidate transceiver array layouts, The smallest candidate transceiver array layout is selected as the current target MIMO transceiver array layout, and the distinguishability between different spatially multiplexed data streams is further adjusted through the target constellation diagram set.

[0025] When the predicted channel correlation corresponding to the target MIMO transceiver array layout parameters is still greater than a preset threshold, the target constellation map set is further determined. Specifically, the constellation map parameters used by the transmitter for different spatially multiplexed data streams are adjusted. The constellation map parameters include at least one of the following: number of constellation points, constellation point positions, constellation point transmission probability, relative offset between different constellation maps, and rotation angle. The target constellation map set is obtained with the goal of improving the achievable rate and reducing spatial sub-channel correlation. The target constellation map set is used by the transmitter to perform constellation mapping on different spatially multiplexed data streams, enabling the receiver to distinguish the signal source of different spatially multiplexed data streams based on the differences in the constellation maps.

[0026] 2. Transmitter link enhancement based on target constellation diagram set and time-reversed frequency domain precoding: 1) At the transmitting end, according to the principle of superimposed constellations, for Spatial multiplexing data stream Different constellation diagrams are assigned, and different target constellation diagrams are used for different spatially multiplexed data streams to improve the signal source differentiation capability under strongly correlated ultraviolet MIMO subchannel conditions in underground coal mines.

[0027] 2) Based on the principle of frequency domain time reversal technology, for each data stream... Each symbol is frequency-domain spread, that is: , In the formula, the expansion matrix Used to separate each symbol by intervals Assigned to On each subcarrier, and Its function is equivalent to up-conversion; Let the first The data symbol vector obtained after spatial multiplexing data streams are mapped by constellation is: , in, , Indicates the number of spatially reused data streams. Indicates the first The number of data symbols contained in a spatially multiplexed data stream within a frequency domain spread cell.

[0028] Let the total number of OFDM subcarriers be The preset subcarrier interval is , No. The starting subcarrier offset of each spatially multiplexed data stream is Then the first In the spatial multiplexing data stream, the first The OFDM subcarrier numbers mapped to each data symbol are: , in, Furthermore, the initial subcarrier offsets of different spatially multiplexed data streams satisfy... This reduces the overlap of subcarrier positions after frequency domain spread of multiplexed data streams from different spaces.

[0029] Definition of the first The frequency domain spread matrix corresponding to each spatially multiplexed data stream is: , its first Line 1 Column elements are represented as: , Then the first The data symbol vector after frequency domain spread of a spatially multiplexed data stream is: ,Right now: , in, Indicates the first The spatially multiplexed data stream in the first Frequency-domain spread symbols on OFDM subcarriers.

[0030] Since each data symbol is assigned to only one preset subcarrier position, the frequency domain spread matrix satisfies: ,in, This indicates the conjugate transpose. express An identity matrix of order 1. This formula is similar to the aforementioned expanded matrix. Consistent constraints are used to ensure that the frequency domain spreading process does not change the number of original data symbols and to facilitate frequency domain despreading at the receiver based on the same subcarrier positions.

[0031] Furthermore, the first The first ultraviolet light transmitting array unit to the first The channel impulse response between the ultraviolet light receiving array elements is denoted as . Then in the first The frequency domain channel response on each OFDM subcarrier is: , in, Indicates the channel impulse response length. Indicates the first The first ultraviolet light transmitting array unit to the first The first ultraviolet light receiving array unit in the... Frequency domain channel response on OFDM subcarriers.

[0032] Correspondingly, the time-reversed frequency domain precoding weights can be expressed as: , in, Indicates the first The spatially multiplexed data stream in the first On the OFDM subcarrier, via the first The first ultraviolet light transmitting array unit and facing the first Time-reversed frequency-domain precoding weights for each ultraviolet light receiving array unit Indicates complex conjugation. This represents the regularization coefficient used to avoid zero denominators.

[0033] Therefore, the first The frequency-domain spread symbols of a spatially multiplexed data stream, after time-reversed frequency-domain precoding, are represented as: , The above formula is consistent with the aforementioned "time-reversed precoding matrix". With data symbols The expression "multiplication for frequency domain precoding" is consistent, but the calculation form corresponding to OFDM subcarriers is further given.

[0034] 3) Obtain the time-reversed precoding matrix based on the channel state information. and with data symbols Multiplication is used for frequency domain precoding, i.e.: , Thus, the data symbols are transmitted via the first After the first light source is emitted, it can be spatiotemporally focused on the first light source. One photodetector.

[0035] 4) After optical OFDM modulation, a time-domain non-negative real signal is obtained. The precoding matrix is ​​used to weight and map the precoding data to the transmit array cells, and a cyclic prefix (CP) and a training sequence (TS) are added. , In the formula, To send the elements of the precoding matrix, i.e. .when At that time, the first The data stream does not pass through the first Each data stream is transmitted through at least one transmit array unit, i.e. .

[0036] 5) At the receiving end, from the received signal , After removing the CP, the spatial sub-channel is estimated using the training sequence. , ; 6) By receiving precoding, a set union is performed to obtain... Data streams ,Right now: In the formula, To receive the elements of the precoding matrix, i.e. ; 7) After OFDM demodulation, channel equalization, frequency domain despreading, and superimposed constellation symbol detection, the source symbol estimate is obtained. .

[0037] 3. Joint robust update of precoding parameters under multiple constraints: The receiver estimates the current spatial sub-channel based on the received channel estimation sequence and determines the channel estimation error range by combining it with historical channel measurement data. Based on the current channel estimation results, the channel estimation error range, and the transmission requirements of each spatial multiplexed data stream, the achievable rate of each spatial multiplexed data stream and the ultraviolet light in the coal mine are determined. Under constraints of transmit power, UV human exposure safety, achievable rate, and transmission interruption probability, a joint robust update is performed on the time-reversed frequency domain precoding weights, transmitter precoding parameters, and receiver precoding parameters. Specifically, the transmit power constraint limits the total transmit power of the UV transmitter array; the UV human exposure safety constraint limits the instantaneous and average UV intensity; the achievable rate constraint ensures the minimum transmission requirements of each spatially multiplexed data stream; and the transmission interruption probability constraint ensures the reliability of the UV non-line-of-sight link in underground coal mines.

[0038] 1) The corresponding robust precoding problem can be expressed as: , In the formula, Indicates the first The achievable rate of a data stream, Indicates the overall static power consumption of the system. To reach the rate threshold, This represents the maximum tolerable interruption probability.

[0039] 2) Based on the probability distribution type of channel estimation error and the numerical relationship of reachable rate, the interruption probability constraint is transformed into a deterministic constraint using Berstein-type inequalities, spherical boundary method and worst-case conditional risk value.

[0040] When considering channel estimation error, the first The actual channel matrix on each OFDM subcarrier is represented as follows: , in, This represents the channel matrix estimated by the receiver based on the training sequence. This represents the channel estimation error matrix.

[0041] Based on the channel estimation error range, construct the channel uncertainty set: ,in, Denotes the Frobenius norm. Indicates the first Upper bound of channel estimation error on OFDM subcarriers.

[0042] The upper bound of this error can be determined by the current channel estimation results and historical channel measurement data: , in, This indicates the number of historical channel measurement samples. Indicates the first The true channel frequency domain response from a historical sample This represents the corresponding channel estimation result. This represents the maximum tolerable probability of transmission interruption.

[0043] To uniformly represent the combined effect of time-reversed frequency-domain precoding weights and transmitter precoding parameters, the first... The spatially multiplexed data stream in the first The equivalent transmit weight vector on each OFDM subcarrier is: , in, Indicates by The time-inverted frequency domain precoding weight vector is composed of the transmitting array units. Indicates the transmission of the precoding matrix The corresponding number in the middle The transmission precoding vector of a spatially multiplexed data stream This indicates that corresponding elements are multiplied.

[0044] The receiving end receives the precoded vector After the merger, the first The spatially multiplexed data stream in the first The signal-to-interference-plus-noise ratio (SINR) on each OFDM subcarrier is expressed as: , in, Indicates noise power. This represents the signal-to-interference-plus-noise ratio (SIR) after considering interference between spatially multiplexed data streams.

[0045] When channel estimation error exists, according to the uncertainty set Construct the first The spatially multiplexed data stream in the first Conservative lower bound of SIR / Noise Ratio on OFDM subcarriers: , This formula is used to give the first... Each spatially multiplexed data stream can guarantee the achievement of a lower bound on the signal-to-interference-plus-noise ratio. Therefore, the first... The robust reachable lower bound of the spatially multiplexed data stream is: in, Indicates the first The set of OFDM subcarriers occupied by a spatially multiplexed data stream after frequency domain spread. This indicates the OFDM subcarrier spacing.

[0046] Constrain the original opportunity: , Transformed into the following sufficient condition of certainty: , This ensures that within the channel estimation error range, the system's total achievable rate is below a threshold. The probability does not exceed the maximum tolerable interruption probability. .

[0047] 3) Based on the above deterministic transformation, the problem is solved using methods such as continuous convex approximation, alternating optimization, and sequential parametric convex approximation, yielding a feasible solution to the original problem. The specific solution steps are as follows: Step 1: Initialization. Based on the optimization results of the previous transmitted frame or preset initial values, initialize the time-inverted frequency domain precoding weights, the transmitting end precoding matrix, and the receiving end precoding matrix: And set the maximum number of iterations. Convergence threshold and iterative index .

[0048] Step 2: Channel estimation error calculation; estimate the current channel matrix based on the training sequence. The upper bound of the error was determined by combining historical channel measurement data. The channel uncertainty set is obtained. .

[0049] Step 3: Fix Update time-inverted frequency domain precoding weights .

[0050] With the precoding matrices at the transmitting and receiving ends fixed, the objective function from the original optimization problem is used as the optimization objective. Under constraints of transmitted power, instantaneous ultraviolet light intensity, average ultraviolet light intensity, and deterministic rate, the update is performed. .

[0051] Step 4: Fix Update the precoding matrix at the sending end. .

[0052] With the time-reversed frequency domain precoding weights and the receiver precoding matrix fixed, the objective function in the original optimization problem is used as the optimization objective, and the algorithm is updated. .

[0053] And it meets the original transmission power constraints, ultraviolet light human exposure safety constraints, and: , Step 5: Fix Update the receiver precoding matrix , When the time-reversed frequency domain precoding weights and the transmitter precoding matrix are fixed, update This increases the robust reachability lower bound of each spatially multiplexed data stream and reduces interference between spatially multiplexed data streams.

[0054] Step 6: Continuous convex approximation processing, for the above update process... The resulting nonconvex term, in the first... A first-order approximation is performed near the next iteration point, yielding: ,in, This represents the variable currently being optimized. Indicates the first The approximate lower bound of the achievable rate is constructed under the next iteration.

[0055] Step 7: Convergence judgment. If the change in the original objective function value between two consecutive iterations satisfies: , or If the iteration stops, the output is: , in, Indicates the first The objective function value of the robust precoding problem at the next iteration. If the convergence condition is not met, then let Return to step 3.

[0056] The final precoding parameter combination includes time-reversed frequency domain precoding weights, transmitter precoding parameters, and receiver precoding parameters, which are used for ultraviolet non-line-of-sight link transmission in the next time slot or the next transmission frame. This reduces transmission power and enhances the coverage capability of ultraviolet non-line-of-sight links in coal mines while ensuring transmission reliability and ultraviolet security constraints.

[0057] The above description is merely one embodiment of the present invention and is not intended to limit the present invention. Any minor modifications, equivalent substitutions, and improvements made to the above embodiment based on the technical essence of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for enhancing non-line-of-sight ultraviolet light links in mines based on MIMO precoding, characterized in that, The steps are as follows: S1. Obtain measurement data, environmental parameters and array constraints of the mine ultraviolet non-line-of-sight MIMO channel, calculate spatial sub-channel correlation, and determine the transceiver array layout within the adjustable range of the array. S2. Determine the constellation diagram set based on spatial sub-channel correlation and reachability rate, and assign different constellation diagrams to different spatial multiplexed data streams; S3. Generate data symbols based on the constellation diagram, generate time-inverted frequency domain precoding weights according to the ultraviolet non-line-of-sight channel impulse response between the transceiver array units, and map the precoded signal to the transmitting array. S4. The receiver estimates the spatial sub-channel and detects data symbols. Based on the estimation error, achievable rate, and interruption probability, and under the constraints of ultraviolet light power and human exposure safety, it jointly updates the time-inverted frequency domain precoding weights, the transmitter precoding parameters, and the receiver precoding parameters for the next frame transmission.

2. The method for enhancing non-line-of-sight ultraviolet light links in mines based on MIMO precoding according to claim 1, characterized in that, The specific steps for S1 are as follows: S1.1 Initialize the mine ultraviolet non-line-of-sight MIMO system. The mine ultraviolet non-line-of-sight MIMO system includes a transmitter and a receiver. The transmitter includes multiple ultraviolet light transmitting array units, and the receiver includes multiple ultraviolet light receiving array units. The transmitter and receiver transmit signals through the mine ultraviolet non-line-of-sight channel. S1.2 Initialize the transmitter and receiver parameters of the mine ultraviolet non-line-of-sight MIMO system, and set initial values ​​for the time-reversed frequency domain precoding weight, transmitter precoding parameters, and receiver precoding parameters; wherein, the initial value is a preset initial value or the optimization result of the previous transmitted frame; S1.3 Collect current environmental parameters in the coal mine, measurement data of the mine's ultraviolet non-line-of-sight MIMO channel, and transceiver array parameters. The current environmental parameters include at least one of turbulence intensity, number of turbulence, turbulence size, object reflection intensity, reflection order, and number of reflecting objects. The transceiver array parameters include at least one of the following: geometry of ultraviolet light transmitting and receiving array units, array unit spacing, spatial orientation angle, beam divergence angle, and receiving field of view. S1.4 Calculate the spatial sub-channel correlation between each transmitting array unit and each receiving array unit under different transmit / receive orientation combinations based on channel measurement data, obtain the spatial sub-channel correlation corresponding to the current transmit / receive array layout, and obtain the average value, maximum value and median based on the spatial sub-channel correlation. S1.5 Input the current environmental parameters, transceiver array parameters, and candidate transceiver array layout parameters into the spatial sub-channel correlation prediction model to obtain the predicted spatial sub-channel correlation index corresponding to each candidate transceiver array layout parameter; wherein, the spatial sub-channel correlation prediction model is a regression prediction model with the current environmental parameters, transceiver array parameters, and candidate transceiver array layout parameters as input and the predicted channel correlation as output. S1.6 Within the physical adjustment range of the spacing between transmitting array elements, the spacing between receiving array elements, the spatial orientation angle of the transmitting array, and the spatial orientation angle of the receiving array, the transceiver array layout is determined from the candidate transceiver array layout parameters with the goal of predicting channel correlation below a preset threshold.

3. The method for enhancing non-line-of-sight ultraviolet light links in mines based on MIMO precoding according to claim 1, characterized in that, The specific steps for S2 are as follows: S2.1 In the confined space of a coal mine, the spatial orientation angle of the ultraviolet light transmitting array unit and the ultraviolet light receiving array unit at the transmitting and receiving ends is limited by the physical adjustment range. Based on the spatial sub-channel correlation and achievable rate under the transceiver array layout, the transmitting end determines a set of constellation diagrams for different spatial multiplexed data streams. The set of constellation diagrams includes multiple constellation diagrams respectively assigned to different spatial multiplexed data streams. The different constellation diagrams differ in at least one of the following: the number of constellation points, the position of constellation points, the transmission probability of constellation points, the relative offset between different constellation diagrams, and the rotation angle. S2.2 The transmitting end performs constellation mapping on different spatially multiplexed data streams according to the constellation diagram set, so that the receiving end can distinguish the signal source of different spatially multiplexed data streams based on the differences between different constellation diagrams, in order to compensate for the insufficient data stream differentiation capability caused by the inability to fully reduce the spatial sub-channel correlation by simply adjusting the layout of the transceiver array.

4. The method for enhancing non-line-of-sight ultraviolet light links in mines based on MIMO precoding according to claim 3, characterized in that, The specific steps for S3 are as follows: S3.1 The sending end converts the service data to be sent into multiple spatial multiplexed data streams, and according to the constellation diagram set, assigns different constellation diagrams to different spatial multiplexed data streams for constellation mapping, and generates data symbols corresponding to each spatial multiplexed data stream; S3.

2. For the received signal of the ultraviolet non-line-of-sight channel in the mine, which is formed by the reflection path of macroscopic objects and the scattering path of microscopic particles, resulting in problems such as weak light reception, multipath propagation, pulse delay spread and inter-symbol interference, the transmitting end performs frequency domain spread on the data symbols generated in S3.1 after constellation mapping and before time-reversed frequency domain precoding. The reflection path of macroscopic objects includes roadway walls, coal walls, rock walls and equipment surfaces, and the scattering path of microscopic particles includes dust, water mist and gas molecules. S3.3 The frequency domain spreading specifically includes: for each spatially multiplexed data stream, the transmitting end allocates the data symbols corresponding to the spatially multiplexed data stream to the frequency domain positions of multiple OFDM subcarriers according to a preset subcarrier interval, to obtain the frequency-domain spread data symbols corresponding to the spatially multiplexed data stream; the frequency-domain spread data symbols are used as the input for subsequent time-inverted frequency domain precoding; S3.4 The transmitting end determines the time-reversal frequency domain precoding weights for the corresponding OFDM subcarriers based on the ultraviolet non-line-of-sight channel impulse response between each ultraviolet transmitting array unit and each ultraviolet receiving array unit. S3.

5. The transmitting end uses the time-reversed frequency domain precoding weight to perform frequency domain precoding on the frequency-domain spread data symbols obtained in S3.3 according to OFDM subcarriers to obtain the frequency-domain precoded signal. S3.6 The transmitting end performs optical OFDM modulation on the frequency domain precoded signal to obtain a time domain non-negative real signal that satisfies the intensity modulation direct detection constraint, and weights and maps the time domain non-negative real signal to the corresponding ultraviolet light transmitting array unit according to the transmitting end precoding parameters. S3.7 The transmitting end determines the cyclic prefix length and training sequence length according to the transmitting end parameters initialized in S1; the transmitting end adds a cyclic prefix to the time-domain non-negative real signal and inserts the training sequence to obtain the ultraviolet light signal to be emitted; wherein, the cyclic prefix is ​​formed by copying a portion of the sampling points at the tail of each OFDM time-domain symbol in the time-domain non-negative real signal to the front of the OFDM time-domain symbol; the training sequence is a known symbol sequence predetermined by the transmitting end and the receiving end.

5. The method for enhancing non-line-of-sight ultraviolet light links in mines based on MIMO precoding according to claim 4, characterized in that, The specific steps for S4 are as follows: S4.1 The receiver estimates the spatial sub-channels between each ultraviolet light transmitting array unit and each ultraviolet light receiving array unit based on the training sequence, and obtains the current channel estimation result; S4.

2. In view of the time-varying and spatially varying environment caused by mining processes, ventilation airflow and equipment movement in coal mines, the channel estimation error is determined by combining the current channel estimation results, historical channel measurement data and current environmental parameters; the channel estimation error is used to characterize the time reversal focus offset and spatial multiplexing data stream interference changes caused by changes in the mine's ultraviolet non-line-of-sight channel. S4.3 The receiver performs weighted merging of the signals received by multiple ultraviolet light receiving array units according to the receiver precoding parameters to obtain the received signals corresponding to each spatial multiplexed data stream; S4.4 The receiving end determines the achievable rate of each spatial multiplexed data stream and the transmission interruption probability of the mine ultraviolet non-line-of-sight MIMO link based on the channel estimation error and the received signal corresponding to each spatial multiplexed data stream. S4.5 Under constraints of transmit power, instantaneous ultraviolet light intensity, average ultraviolet light intensity, ultraviolet light human exposure safety, achievable rate, and transmission interruption probability, with the goal of reducing transmit power while meeting the achievable rate requirements of each spatially multiplexed data stream, robust optimization is performed by using the time-reversed frequency domain precoding weights, transmit-end precoding parameters, and receive-end precoding parameters as joint optimization variables. S4.

6. The robustly optimized time-reversed frequency domain precoding weights, transmitter precoding parameters, and receiver precoding parameters are used for mine ultraviolet non-line-of-sight MIMO link transmission in the next transmission frame or the next time slot.

6. A computer device, characterized in that, It includes a processor and a memory, the processor being electrically connected to the memory, the memory being used to store instructions and data, and the processor being used to execute the MIMO precoding-based method for enhancing ultraviolet non-line-of-sight links in mines as described in any one of claims 1 / 5.

Images