54 results about "Minimum mean square error" patented technology

The invention relates to the technical field of communication, and discloses a network communication chip signal interference suppression processing method. The method comprises the following steps: acquiring a local oscillator phase jitter sequence in real time through an on-chip phase noise monitoring unit; estimating Doppler frequency shift and channel coherence time based on the sequence; dynamically adjusting the closed-loop bandwidth of the phase-locked loop; and generating a phase compensation signal, and performing time-varying channel minimum mean square error equalization in combination with a received baseband signal to eliminate inter-symbol interference. The system integrates a phase noise monitoring function module, a channel parameter estimation function module, a phase-locked loop bandwidth regulation function module, a phase compensation generation function module and an interference joint elimination function module. By means of dynamic bandwidth reconstruction and high-precision phase compensation, the bit error rate under high-speed movement is remarkably reduced, the 5G millimeter wave communication standard requirement is met, and the method has the advantages of being achieved on a whole chip, low in delay and high in robustness.

The invention relates to the technical field of agricultural machinery load control, in particular to a corn harvester operation load control system and method, and the system comprises a power extraction module, a speed regulation and control module, a trend analysis module, a parameter optimization module and a load closed-loop module. According to the method, frequency characteristics are extracted in a periodic and segmented mode through header power signals, load dynamic sensing is carried out, advancing speed adjustment is more real-time and prospective, response sensitivity and stability are improved in combination with PID control, equipment stably operates at multiple crop densities and plots, the torque increment of a main shaft is monitored in real time, and wavelet transformation multi-scale analysis is carried out; load fluctuation fine-grained identification is carried out, a roller gap and power output are dynamically adjusted to optimize parameters with a minimum mean square error, frequency and torque characteristic closed-loop comparison is constructed, control self-matching and continuous self-adjustment are carried out, error minimization and trend tracking feedback are fused, the intelligence of the harvester is enhanced, and load sudden change performance fluctuation and energy consumption loss are reduced.

The invention belongs to the technical field of laser vibration measurement, and provides a self-focusing laser Doppler vibration measurement method and system capable of resisting self-vibration of an optical fiber, and the method comprises the steps: adding a compensation light path which is used for sensing and generating reference light which is equal to an interference signal in amplitude and is opposite to the interference signal in real time in a Doppler vibration measurement light path, and taking the minimum mean square error as a self-adaptive matching criterion; interference phases obtained by synchronous demodulation of the compensation light path and the main light path are subjected to adaptive matching and cancellation in amplitude and phase dimensions, residual errors of vibration of the vibration measurement system and other environmental interferences are suppressed below noise of the vibration measurement system, and the purpose of balancing and weakening optical fiber vibration interferences in an instrument is achieved. The influence of the vibration factor of the instrument is reduced, and the interference influence is accurately eliminated through amplitude and phase matching by a demodulation algorithm corresponding to a compensation light path; meanwhile, by dynamically observing and updating the return light intensity threshold, low-frequency disturbance is resisted, and accurate automatic focusing is achieved.

The invention discloses a cognitive function screening system and method based on an MMSE (Minimum Mean Square Error) prediction model, and relates to the technical field of data analysis, the method comprises the following steps: collecting physical examination index data, removing missing records, adopting a multiple interpolation method for interpolation, and obtaining multiple sets of complete data sets; a continuous prediction model is constructed, MMSE continuous prediction values are obtained, and a sensitivity analysis report is generated; constructing a first-stage classification model, adaptively dividing an optimal threshold combination, and dividing a sample into a high-confidence region, a to-be-discriminated region and a low-confidence region; if the sample size of the to-be-discriminated region is higher than a preset training threshold value, constructing an enhanced feature set, and constructing a second-stage classification model; when the prediction probability reaches the optimal re-discrimination threshold value, the classification result in the first stage is corrected, and otherwise, the classification result is maintained; if not, maintaining the classification result; and integrating the classification results to obtain a final classification result of all the samples.

The invention discloses a channel estimation method of a time-varying channel in a time delay-Doppler domain pilot mode, which comprises the following steps of: superimposing a hybrid superimposed pilot and a data symbol in a time delay-Doppler domain, the pilot comprising a high-power center pilot and a low-power edge pilot; the signal is converted to a time domain through inverse symplectic Fourier transform and Hisenberg transform; a receiving end obtains a receiving signal through Wigner transform and Sextile Fourier transform; estimating integer time delay and Doppler taps by using a center pilot frequency, and constructing a sparse channel matrix; estimating fractional Doppler and channel gain through a sparse Bayesian learning algorithm; and detecting a data symbol based on a minimum root-mean-square error algorithm, and realizing accurate estimation of channel parameters through iterative interference elimination. According to the method, the peak-to-average power ratio and the algorithm complexity can be effectively reduced, and non-correlation channel components and cross interference are inhibited.

The application relates to a kind of OFDM-based broadband millimeter wave relay system hybrid beamforming design method, comprising: calculating the receiving signal of target node under the condition that each node is all digital processor;The receiving signal is processed using minimum mean square error criterion in target node, and MMSE matrix is calculated;Using the equivalence between maximizing sum rate and minimum weighted mean square error algorithm, the maximum sum rate problem is converted into minimum weighted mean square error problem according to MMSE matrix;Deep unfolding neural network is used to solve the minimum weighted mean square error problem, and all digital processor of each node is obtained;Each node's all digital processor is decomposed based on least square decomposition algorithm, and hybrid beamforming matrix of each node is calculated, the complexity of the algorithm is greatly reduced, and the running time of the system can be effectively improved.

The application belongs to the technical field of unmanned aerial vehicles, and discloses an unmanned aerial vehicle target searching method and system based on an electromagnetic map. A wireless signal receiving element carried by the unmanned aerial vehicle receives radio frequency signals from a target. Terrain information, scene parameters and the radio frequency signal strength received in the first step are used to construct an electromagnetic map through electromagnetic wave propagation theory. The accurate target position coordinates are estimated through the least mean square error estimation method according to the electromagnetic map obtained in the second step. The unmanned aerial vehicle is controlled to fly to the specified position coordinates. The above steps are repeatedly performed until the search is successful or fails. The application fully utilizes the prior rich real terrain information through the low-cost and low-energy hardware requirement of the unmanned aerial vehicle end signal strength receiving, realizes the accurate positioning and searching of the radio frequency target, and is significantly superior to the algorithm without using the electromagnetic map in terms of accuracy and time efficiency.

The invention provides an uplink MIMO-NOMA (Multiple Input Multiple Output-Non-Orthogonal Multiple Access) system optimization method with a maximized sum rate, which comprises the following steps: firstly, in order to suppress interference between a transmission cluster and a reflection cluster, providing an inter-cluster SIC method: clustering users according to transmission characteristics, adopting space division multiple access in each cluster, and performing inter-cluster SIC of NOMA transmission so as to reduce the interference between the clusters; secondly, a non-convex optimization problem aiming at maximizing the total rate of the system is formulated and solved by using an alternating optimization framework, in the framework, a receiving equalizer is optimized according to a minimum mean square error standard, and transmitting power control of a user is processed through successive convex approximation, so that sub-problems are easy to process; and finally, the beam forming of the active STAR-RIS adopts a sequential rotation design. According to the invention, the dual-fading effect is effectively overcome by adjusting the phase shift and amplitude of electromagnetic waves, an alternating optimization framework is also provided to solve the described non-convex problem, user power is controlled through continuous convex approximation, and active STAR-RIS beam forming is realized through sequential rotation.

The application relates to a weak underwater target azimuth detection pre-tracking method based on belief propagation, which comprises the following steps: converting a weak target tracking problem into solving a joint posterior probability density function (pdf) based on a Bayesian rule; factorizing the joint posterior pdf, constructing a corresponding factor graph through a graph model; solving the transmitted information in the factor graph by using a belief propagation (BP) algorithm, and converting multiple integrals into ordinary integrals by using a Goodman variable substitution principle; constructing a likelihood function by using a complex Wishart distribution to describe the relationship between original sonar data and target states, and simplifying the remaining pdf by using a Gaussian distribution to obtain a confidence approximation of a target state edge posterior pdf; and calculating the number of weak targets and corresponding states by using a minimum mean square error (MMSE) estimator. The application can simultaneously track multiple weak targets without a complex clustering algorithm, has fewer preset parameters, and can accurately and efficiently track target state information including an azimuth, an azimuth angular velocity and target intensity.

The application discloses a dynamic human body reconstruction method based on a neural radiation field, and comprises the following steps: S1, establishing a dynamic human body reconstruction network based on a neural radiation field, wherein the dynamic human body reconstruction network comprises an action branch for acquiring an action field and a reconstruction branch for acquiring a neural radiation field; S2, initializing the action branch and the reconstruction branch in a specified resting posture by using a human body parameterized model; S3, using real images in a given posture and given camera internal and external parameters as a supervision signal, and simultaneously training the action branch and the reconstruction branch by taking the minimum mean square error between the real images and synthetic images as a loss function; and S4, inputting a new human body posture and a camera pose to the trained dynamic human body reconstruction network, and rendering a human body image in any view point in the new posture.

The invention is suitable for the technical field of geographic element data analysis, and provides a geographic spatiotemporal weighted vector autoregression-based multi-element collaborative change modeling method, which comprises the following steps of: constructing a GTWR-VAR model, and combining a multivariable feedback mechanism of vector autoregression with locality of geographic spatiotemporal weighted regression to obtain a GTWR-VAR model; the method comprises the following steps of: quantifying spatial-temporal heterogeneity of mutual influence among geographic elements, introducing a Bayesian optimization mechanism, adaptively determining an optimal lag order, a spatial bandwidth and a time bandwidth by taking verification set mean square error minimization as a target, and finally adopting Monte Carlo block permutation test to determine the optimal lag order, the spatial bandwidth and the time bandwidth on the premise of retaining inherent time autocorrelation of data. And evaluating and eliminating non-significant items in the local coefficient to ensure the statistical effectiveness of the result. According to the method, the continuous change rule of the influence mechanism between the geographic elements in time and space is disclosed, and a more accurate and reliable basis is provided for regional response research under fine management, prediction and climate change.

The invention discloses an interference equalization method, device and equipment for multi-carrier modulation in a high-dynamic scene and a storage medium, and relates to the technical field of mobile communication, and the method comprises the steps: obtaining a time domain channel matrix of a time-frequency double-color dispersion channel, and determining an inversion core matrix in a minimum mean square error equalization matrix based on the time domain channel matrix; based on the sparsity structure of the inversion core matrix, decomposing the inversion core matrix to obtain a decomposition matrix; based on the time domain channel matrix, the inversion core matrix and the decomposition matrix, determining a multilayer linear equation set with minimum mean square error equalization; and solving the multi-layer linear equation set to obtain an equalization result. By means of the mode, the calculation complexity of the equalization process in the process of resisting inter-symbol interference and inter-carrier interference existing in multi-carrier modulation in a high dynamic scene is remarkably reduced, excellent performance is still kept under the extreme condition that no cyclic prefix exists or the length of the cyclic prefix is zero, and high scene adaptability and system practicability are embodied.

The invention relates to the field of sound signal processing, and discloses a relay gateway echo cancellation method and device and a medium, and the method comprises the steps: obtaining a near-end voice instruction digital signal X (n) and a far-end reference signal Y (n) through a TDM bus interface, and n is a time domain sampling point sequence number; inputting the Y (n) into an adaptive filter bank, calculating a filter coefficient W (n) through a minimum mean square error criterion, and generating an analog echo signal; performing time domain subtraction operation on the X (n) to obtain a primary residual signal; performing non-linear distortion detection, and activating a secondary cancellation module when a non-linear component is detected; performing sub-band decomposition on the X (n) in the secondary elimination module, and generating a compensation signal C (n) by adopting a frequency domain weighting method; carrying out superposition processing on the residual signal and C (n), and outputting a final residual signal; the method has the beneficial effect that linear and nonlinear echoes are remarkably reduced.

The invention relates to the technical field of 5G communication channel estimation, and discloses a 5GMIMO channel estimation optimization method based on deep learning. According to the method, a deep learning model architecture is constructed, and model parameters are dynamically initialized according to channel coherence time so as to adapt to 5GMIMO channel multipath propagation characteristics. Received signal streams from a plurality of antenna elements containing different orthogonal frequency division multiplexing sub-carrier frequency sequences are processed, channel impulse response estimates are generated based on frequency domain correlation and input into a model for nonlinear transformation. And performing model parameter collaborative optimization, updating the network weight by calculating the gradient variation, constructing a target function according to channel delay extension and model depth association, and adjusting the estimation output towards the direction of minimizing the mean square error. And triggering model structure adaptation based on a norm of gradient variation, acquiring model copy parameters matched with the current channel state from an adjacent cell base station, normalizing the model copy parameters, and adjusting the number of attention heads or the number of residual blocks.

The invention discloses a message passing detection algorithm for a bidirectional transmission symbol probability of an orthogonal time frequency air conditioning signal, and provides a calculation formula of a bidirectional transmission probability message, a convergence judgment criterion of an iterative detection process and two methods for reducing calculation complexity. The error code performance of the detection algorithm is superior to that of a linear minimum mean square error detection algorithm and a traditional message passing detection algorithm, and the convergence speed is higher than that of the traditional message passing detection algorithm. After the two methods for reducing the calculation complexity are applied, the bit error rate performance is only slightly reduced, but the calculation complexity is obviously reduced.

The invention discloses an MIMO (Multiple Input Multiple Output) resource allocation method considering a channel estimation error and a related device, and the method comprises the steps: constructing a multi-cell MIMO system, and carrying out the estimation of a channel through employing an uplink orthogonal pilot frequency, and obtaining a least square and minimum mean square error estimation channel; estimating a channel according to the least square and the minimum mean square error, and determining an expected closed expression of a channel estimation error matrix; based on the channel hardening effect, with the average minimization of all users as the optimization target, a pilot frequency power distribution optimization problem is constructed, the pilot frequency power distribution optimization problem is solved, and the optimal pilot frequency power distribution scheme is obtained.By means of the method and the related device, pilot frequency pollution can be effectively relieved, the channel estimation precision is improved, and the channel estimation efficiency is improved. And meanwhile, balanced optimization of the overall performance of the system is realized.

The invention discloses a frequency domain super-resolution reconstruction OFDM (Orthogonal Frequency Division Multiplexing) waveform distortion suppression method, and belongs to the field of wireless communication. The method solves the problems that an existing method is high in complexity, sensitive to channel conditions and poor in error rate performance. The method comprises the following steps: constructing an interpolation matrix at a receiving end, and carrying out weighted interpolation on a frequency domain receiving signal by utilizing adjacent spectral line weighting to generate an interpolated frequency domain signal; the method comprises the following steps of: deducing to obtain a time domain equivalent relation of interpolation processing of a transmitting end and a receiving end by combining the logic of performing frequency domain zero insertion on an original frequency domain signal by a transmitting end, further jointly estimating an interpolated channel, and designing an equalizer according to a minimum mean square error criterion to perform equalization processing on the interpolated frequency domain signal to complete data recovery. According to the method, the problem of inter-symbol interference caused by insufficient CP of a common OFDM system with insufficient cyclic prefixes is solved, inter-symbol interference and inter-subcarrier interference caused by insufficient cyclic prefixes can be effectively suppressed, and accurate demodulation data can be obtained. The method can be applied to the field of wireless communication.

The invention provides a power distribution transmission method and device under an OCDM frequency selection channel, and the method comprises the steps: converting a time domain signal into a frequency domain through a discrete Fourier transform matrix at a transmitting end, generating a power distribution matrix through combining channel state information and adopting a water injection algorithm, and carrying out the dynamic power precoding of a frequency domain subcarrier; generating a time domain OCDM signal through an inverse discrete Fourier transform matrix and an inverse discrete Fresnel transform matrix, adding a cyclic prefix, and transmitting the time domain OCDM signal; after a receiving end is subjected to prefix removal, a discrete Fourier transform matrix is converted into a frequency domain, a frequency domain equalizer is designed on the basis of a minimum mean square error criterion by utilizing channel diagonalization characteristics, signals are equalized in combination with transmitting power information, and finally time domain information is recovered through an inverse discrete Fourier transform matrix; therefore, the problem of traditional power distribution failure caused by OCDM channel energy dispersion is solved, and through receiving and transmitting joint optimization, diversity gain is fully mined, frequency selective channel fading is adaptively matched, the anti-interference and anti-multipath capabilities of the system are improved, and the spectrum efficiency and the transmission performance are maximized.

This invention relates to a symbol detection method for an orthogonal time-frequency spatial modulation system combining LMMSE and MRC, belonging to the field of mobile communication. Based on the orthogonal time-frequency spatial OTFS modulation system, linear minimum mean square error (LMMSE) detection is first performed in blocks in the time domain. The detection results are then combined and converted to the time-delay-Doppler domain for orthogonal amplitude demodulation. The demodulated results are then converted to the time-delay-time domain and used as the initial values ​​for the maximum ratio combining (MRC) detection algorithm in the time-delay-time domain, thereby improving the bit error rate performance of the MRC detection method. Simulation experiments verify the effectiveness of this invention. Compared with other initialization detection methods, this method has a lower bit error rate at high communication rates and also exhibits superior performance in high-speed mobile scenarios.

The invention provides a channel estimation method for a minimum mean square error, which comprises the following steps of: performing Fourier transform on a received baseband signal to obtain OFDM (Orthogonal Frequency Division Multiplexing) frequency domain data, and obtaining frequency domain data of all symbols on a receiving side through solution resource mapping; the method comprises the following steps: processing by utilizing frequency domain data of a pilot symbol at a receiving side and a priori known pilot symbol sequence to obtain channel responses of SNR (Signal to Noise Ratio) estimation and LS (Least Square) channel estimation; processing according to the channel response of the LS channel estimation to obtain root-mean-square delay expansion; calculating a data-pilot frequency cross-correlation matrix and a pilot frequency self-correlation matrix according to root mean square time delay expansion, and calculating to obtain an MMSE filtering channel estimation coefficient according to the data-pilot frequency cross-correlation matrix and the pilot frequency self-correlation matrix; mMSE filtering is carried out according to the MMSE filtering channel estimation coefficient, and finally a final channel estimation value is obtained. According to the method, channel response is accurately estimated while operation is simplified, so that reliable communication and low operation delay are ensured.

This invention discloses a group-connected BD-RIS-assisted uplink communication system and a method for maximizing transmission rate, relating to the field of wireless communication technology. The method first establishes a channel model between the user, base station, and BD-RIS, derives the decoded signal expression for the user at the base station, and calculates the user's signal-to-interference-plus-noise ratio (SNR) and achievable rate. Then, it constructs an optimization problem with the goal of maximizing the system's sum and rate, with the optimization variables being the equalizer, user transmit power, and BD-RIS phase shift matrix. Finally, it employs an alternating optimization strategy, sequentially optimizing the equalizer based on the minimum mean square error criterion, transforming the non-convex problem into a convex problem to solve for the transmit power, and optimizing the phase shift matrix using the BFGS quasi-Newton method, iterating until the system performance converges. This invention overcomes the design limitations of traditional reconfigurable smart reflectors, fully leverages the control advantages of BD-RIS, effectively improves the sum and rate of the uplink communication system, and provides a new solution for future wireless network performance optimization.

This invention discloses an adaptive detection and separation method for human breathing and heartbeat signals in fire rescue. The invention includes software, hardware, and algorithms. It can automatically select the equipment's operating frequency band according to the needs of fire rescue, performing breathing detection after penetrating obstacles and when there are no obstacles or only a few obstacles. The detection function is implemented by controlling a USRP B210-based software-defined radar hardware through GNU Radio Companion software. Data preprocessing of the transmitted and received baseband signals yields a composite signal of superimposed human breathing and heartbeat signals. To address the difficulty in separating human breathing and heartbeat signals in software-defined radar detection, adaptive filtering and moving target display are applied to the composite signal to separate the breathing and heartbeat signals. The method includes: a human breathing and heartbeat signal detection method based on software-defined radar, and an adaptive human breathing and heartbeat signal separation method based on minimum mean square error moving target display.

The application discloses a diatom image classification system and method driven by data and knowledge, which comprises a generative adversarial network module, a knowledge learning module and a deep residual network module; the generative adversarial network module designs a generative adversarial network model suitable for diatom image generation to expand a diatom image dataset; the knowledge learning module embeds diatom knowledge into a knowledge vector space; and the deep residual network module embeds diatom images into a feature vector space; in order to overcome the semantic gap between the knowledge domain and the feature domain, a knowledge-assisted deep residual network classification is utilized, a multilayer perceptron is adopted to project the knowledge domain into the feature domain, and a least mean square error loss function is used to minimize the distance between the knowledge embedding and the feature embedding in the feature domain; compared with a traditional classification model without the generative adversarial network and the knowledge attribute, the method of the application improves the precision by 8.6%, the recall by 9.7% and the loss by 31.6%.

This invention discloses a self-calibrated image denoising method utilizing convexity, belonging to the field of image processing technology. It is used for image denoising under conditions of unknown noise levels. The method includes generating a pair of noisy images and images with secondary noise, constructing a conditional denoising network; deriving the denoising estimator and calculating the variance of the estimation error based on the linear minimum mean square error estimation theory, and introducing a noise ratio parameter for normalization analysis; utilizing the convexity of the error variance function, determining the optimal synthetic noise variance through differentiation or a ternary search algorithm; calibrating the optimal synthetic noise variance using an empirical formula; training and executing the denoising network using the calibrated optimal variance, and outputting the final denoised image. This invention, by constructing a robust training process under conditions of unknown noise levels, can effectively learn the statistical characteristics of noise, overcoming the limitation of traditional methods that require prior knowledge of the noise level.

Disclosed herein is an apparatus and method for channel estimation using Demodulation Reference Signal (DMRS). The apparatus receives a DMRS, computes a Least Squares (LS) channel estimation value corresponding to the received DMRS, generates a time-domain channel response by performing an Inverse Discrete Fourier Transform (IDFT) on the LS channel estimation value, estimates noise components by setting bilateral noise windows within the time-domain channel response, removes noise components from the time-domain channel response by performing element-wise Minimum Mean Square Error (eMMSE) correction using the estimated noise components, and recovers a frequency-domain channel response by performing an N-point DFT on the corrected time-domain channel response.

The invention provides an RIS-assisted cellular-free large-scale MIMO network security energy efficiency optimization method. The method comprises the following steps: establishing an RIS-assisted cellular-free large-scale MIMO network model under a Rician fading channel; acquiring estimated channel state information of aggregation channels between the base station and the user equipment and between the base station and the eavesdropper by adopting a linear minimum mean square error estimation method; performing downlink channel data transmission based on the estimated channel state information; deriving a lower bound of a user reachable rate and an upper bound of an eavesdropper leakage rate according to a downlink channel data transmission result; deducing a safety reachable rate expression and a safety energy efficiency expression according to the lower bound of the user reachable rate and the upper bound of the eavesdropper leakage rate; based on the safety energy efficiency expression, a non-convex optimization problem containing RIS phase shift and power distribution is established, an iterative optimization algorithm is adopted for solving, and the iterative optimization algorithm decouples the original problem into an RIS phase shift optimization sub-problem and a power distribution optimization sub-problem for alternate solving until convergence is achieved to a local optimal solution meeting all constraints.

The application discloses a one-code multi-stream transmission method based on random multiplexing and joint detection decoding, wherein original signals are subjected to random multiplexing at a transmitting end, so that signal energy is uniformly distributed on all fading branches; signals transmitted by the transmitting end are received at a receiving end, and an iterative process is performed by using a cascaded CD-OAMP detection decoding receiver or a joint CD-OAMP detection decoding receiver; a linear estimator of the CD-OAMP performs linear minimum mean square error estimation, a nonlinear estimator performs symbol-by-symbol demodulation, and iteration is performed until convergence, so that a decoding result removing inter-symbol interference is obtained. The application makes equivalent channels approximate AWGN channels by random multiplexing, so that a single AWGN code is still robust. Inter-symbol interference introduced by random multiplexing is eliminated by a CD-OAMP detector with low complexity and Bayesian optimality. Decoding information is fully utilized by joint decoding detection, so that signal estimation is more accurate.

The application discloses a method for analyzing bit error rate performance in a space modulation assisted affine frequency division multiplexing system; in the space modulation mode, information bits are mapped into constellation symbols and indexes of activated antennas, the system experiences a time-frequency double selective channel after affine frequency division multiplexing modulation, linear minimum mean square error equalization is performed on the affine Fourier transform domain received signal after affine frequency division multiplexing demodulation at the receiving end, then the equalized signal is grouped, maximum likelihood detection is performed on each carrier symbol to obtain estimated symbols and activated antenna indexes, and the estimated symbols and activated antenna indexes are demapped into bits; theoretical approximate closed expressions of error probabilities of transmitting symbol detection and activated antenna index detection are derived respectively, and the average bit error rate of the system is obtained; through simulation verification, the performance analysis method provided by the application can effectively measure the average bit error rate performance of the system.

The application discloses a digital air computing method based on binary complement coding, comprising the following steps: linear scaling and down rounding quantization are performed on a continuous analog signal to be calculated at a distributed device end, and the quantized value is converted into a binary complement sequence with a fixed length; a decoder is designed at a base station end based on a superposition result of each binary bit; each bit of the binary complement sequence is subjected to binary phase shift keying (BPSK) modulation, a frequency band is divided into a plurality of subcarriers by using orthogonal frequency division multiplexing (OFDM) technology, and a plurality of devices superimpose transmission on the same subcarrier for the same binary bit; equal power allocation is performed according to the weight priority of the binary bit, a transmission signal is preprocessed based on channel state information by using a truncated channel inversion strategy; a linear minimum mean square error (LMMSE) detector is applied to each subcarrier received signal at the base station end to estimate a superposition result; a multi-subcarrier mean square error minimization problem is constructed, and a greedy algorithm is used to optimize an active device set and a transmission power factor; and the application significantly improves the real-time performance and reliability of multi-device cooperative computing in actual application.

This invention discloses a gesture recognition method based on an improved tabu search-optimized BP neural network, comprising: S1 initializing neural network parameters, determining network input and output parameters, and tabu search-related parameters; S2 generating neighborhood boundary solutions according to randomly generated operators; S3 calculating the point with the minimum mean square error within the neighborhood; S4 moving the operator to the optimal point and regenerating the neighborhood until the mean square error of the operator is the minimum value within the operator's neighborhood; S5 storing the operator in a tabu table, randomly jumping to generate initial solutions, and continuing to search for local extrema until the tabu table is full or the search ends after a set number of iterations; otherwise, returning to step 2; S6 cyclically passing the local extrema points in the tabu table to the BP neural network training model; and S7 calculating the mean square error of each extrema point to obtain the optimal gesture recognition network model. This invention can overcome the local extrema problem, accelerate the convergence speed of the algorithm, and significantly improve the accuracy of gesture recognition.