68 results about "Neural network design" patented technology

The invention discloses a transformer electromagnetic thermal field real-time prediction method based on a physical constraint embedded neural network, and belongs to the technical field of transformer monitoring. The method aims at solving the problems that a traditional finite element method is poor in real-time performance, low in precision and weak in pure data driving model generalization. The method comprises the following steps: selecting a load rate, an environment temperature and a shell convective heat transfer coefficient as key parameters, generating a sample by optimal Latin hypercube sampling, and establishing a three-dimensional electromagnetic-thermal-fluid coupling finite element model to construct a training / testing database; constructing a deep full-connection neural network of which the input is five parameters easy to measure and the output is a winding temperature nephogram, and designing total loss function training containing data / physical loss; and deploying an on-line monitoring system after verification is qualified, and collecting parameters in real time to output a winding temperature cloud picture. The method has the characteristics of high precision, strong generalization and easy deployment, and provides support for intelligent operation and maintenance and digital twinning of the transformer.

An adaptive impedance-based multi-mobile-robot collaborative transportation control method. Each mobile robot estimates the actual pose and ideal pose of a reference point and the first and second derivatives of the ideal pose by means of two finite-time fully-distributed observers, respectively; and then, on the basis of the estimated poses of the reference point, the pose of an end-effector of a mechanical arm, and closed-chain constraints for collaborative transportation, an ideal trajectory of the end-effector of the mobile robot, and an estimated value of a pose deviation between the end-effector of the mobile robot and the reference point are obtained. An adaptive impedance system of each mobile robot is interconnected with a virtual energy tank, and the energy tank is used to guide the updating of impedance parameters, thereby ensuring the passivity of the entire collaborative adaptive impedance system. To process unknown system dynamics of mobile robots, an asymptotic tracking adaptive neural network controller is designed using a neural network, thereby asymptotically achieving an ideal adaptive impedance relationship. The operational accuracy of multi-robot collaborative transportation systems is improved while ensuring safe collaboration.

The invention belongs to the technical field of robot control, and discloses a visual servo intelligent robust control method for a mobile mechanical arm for a complex operation task, which comprises the following steps of: 1, acquiring a working scene image, and calculating a space coordinate of a target feature point in real time through a projection transformation model; step 2, dynamically inhibiting visual measurement noise by adopting adaptive Kalman filtering, generating a feedforward compensation signal through an integral sliding mode observer, and decoupling chassis slippage disturbance; step 3, inputting the pose signal into a radial RBF neural network, designing an RBF gain scheduler, and adjusting an output proportion-integral gain in real time to suppress external time-varying disturbance; 4, adopting a hybrid visual servo mode switching mechanism, and generating a mobile platform control instruction through an integral sliding mode surface in a position servo mode; and 5, designing a three-order composite controller to drive the mechanical arm so as to realize accurate grabbing. According to the method, the grabbing deviation caused by kinematics uncertainty of the mobile platform and dynamic environment disturbance is eliminated.

The invention discloses a channel state information (CSI) positioning method based on improved piecewise averaging method (IPAA) dimensionality reduction and multi-scale feature extraction under multiple channels. In the data preprocessing stage, statistical features are extracted from obtained CSI measurement values by using an IPAA dimension reduction method, and data dimensions are reduced. Designing a parallel processing structure by using three deep neural networks, namely a 1D convolutional neural network (1DCNN), a long-short-term memory network (LSTM) and a Transform, extracting multi-scale features of the CSI measured value again, fusing the multi-scale features through a feature fusion strategy, and finally, performing offline classification learning on the fused features through a classification network, so as to obtain the multi-scale features of the CSI measured value. Therefore, a high-precision indoor positioning task is realized. The method has remarkable advantages in the aspects of positioning precision and calculation efficiency, and can effectively deal with a complex indoor environment.

The invention discloses an optimal neural network differential classifier searching method based on a cryptographic structure. An existing method is improved from the three aspects of design of an ISRN neural network, data set format optimization based on a password structure, self-adaptive evolutionary optimizer construction and candidate difference set search. In the ISRN neural network design part, initial convolution layers with different convolution kernel sizes are combined with SE module adaptive screening, so that the utilization rate of the neural network to feature information is improved; in the data set format optimization part based on the password structure, more features in the differential propagation process can be provided for the data set input format of the password structure; in the self-adaptive evolutionary optimizer construction and candidate difference set search part, the information entropy evaluation is introduced, and the interlace operation and mutation probability are dynamically adjusted in different stages. The global search capability and convergence capability of the evolutionary optimizer are ensured, the search efficiency and convergence speed of the algorithm are improved, and falling into a locally optimal solution is avoided.

The invention discloses a piezoelectric micro-positioning platform preset performance control method based on an extended state observer, and the method mainly comprises the steps: firstly, building a piezoelectric micro-positioning platform system model considering input hysteresis and unknown disturbance; secondly, designing an extended state observer based on a radial basis function neural network to estimate an unmeasurable state of the system and lumped disturbance containing input hysteresis; then, a preset performance function is used, so that the overshoot performance of the system is effectively improved; thirdly, in combination with a preset performance function, a first-order sliding mode differentiator and an expansion state observer, providing a virtual control law and an adaptive control law; and finally, a preset performance controller is utilized, a Lyapunov stability theory is combined, proper parameters are selected to ensure that the closed-loop system is kept stable within preset time, and tracking errors are controlled within a set error range.

The invention relates to a cloud computing resource load prediction method based on TCN-PatchTST (Trusted Cryptographic Network-PatchTST). Comprising the following steps: collecting CPU load time sequence data to form a training sample set; carrying out preprocessing and feature construction on the training sample, and selecting a memory utilization rate load value which has obvious influence on a CPU load value as an input feature to predict the CPU load value; according to the method, a TCN-PatchTST neural network is constructed, a two-channel feature extraction architecture is designed, local time sequence features are extracted through causal convolution and expansion convolution of TCN, a long-term dependency relationship is captured by using partitioning processing and a channel independence mechanism of PatchTST, and a feature fusion layer is designed to integrate multi-scale information; the TCN-PatchTST neural network is trained; and inputting a test sample into the trained TCN-PatchTST neural network to obtain a CPU load value prediction result of the test sample. According to the method, load prediction is carried out through the historical load time sequence data, the prediction precision of the CPU load value of the cloud computing resources is effectively improved, and technical support is provided for intelligent scheduling of the cloud resources.

The invention discloses a five-phase permanent magnet synchronous motor harmonic suppression control method based on machine learning, and the method comprises the following steps: S1, constructing a mathematical model of a five-phase permanent magnet synchronous motor drive system in a rotating coordinate system, and designing a five-phase permanent magnet synchronous motor control system based on a deep neural network; and S2, deeply fusing the physical prior information of the motor into a training process to enhance the generalization ability of the model, and proposing a hybrid optimization strategy to improve the fitting precision of the predicted network voltage and the actual working voltage, suppress the fluctuation in the prediction process, and enable the predicted value to better meet the actual operation requirement. The invention overcomes the limitation of the existing dual-plane current control method, and makes full use of the technical advantages of the method.

The invention discloses a watermarking method and device suitable for any resolution image input, and the method comprises the steps: generating a fixed low-resolution watermark residual error through a watermark encoder according to an input carrier image and a watermark message, restoring the watermark residual error to the size of the carrier image, and superposing the watermark residual error with the carrier image to obtain a watermark-containing image; designing a watermark discriminator based on the residual neural network, wherein the watermark discriminator is used for distinguishing the difference between the watermark-containing image and the carrier image; a watermark noise layer is used for simulating various distortion operations suffered by an image under a digital channel, and disturbance is applied to the watermark-containing image processed by a watermark discriminator; and transmitting the watermark-containing image to which the disturbance is applied in a real channel, inputting the watermark-containing image into a robust watermark decoder, and decoding and extracting a watermark message. According to the method and the device, image input of any resolution can be processed, good invisibility and robustness can be kept, and copyright protection and leakage traceability of the image of any resolution can be realized.

The application discloses a design method and device of a metasurface optical neural network, and the design method comprises the following steps: obtaining an artificial neural network model and a metasurface optical neural network model, wherein the artificial neural network model is a trained model; selecting n intermediate layers in the artificial neural network model in the order from shallow to deep, one-to-one corresponding the n intermediate layers and n metasurface layers in sequence, and calculating an intermediate feature loss according to the similarity between the features output by each layer in the n intermediate layers and the features output by the corresponding metasurface layer; training the metasurface optical neural network model; and determining the process parameters of the metasurface according to the trained metasurface optical neural network model. The design method completely transplants the calculation capacity of the artificial neural network into the metasurface optical neural network, solves the problems of low design efficiency and low precision of the metasurface optical neural network, and is beneficial to the development of the metasurface optical neural network and the deployment in an environment with limited computing power and storage.

The application relates to a personalized privacy risk dynamic measurement method based on a deep neural network, and specifically comprises the following steps: S1, collecting personalized privacy demands of a user, and establishing a personalized user privacy demand vector; S2, according to the personalized user privacy demand vector, extracting a feature matrix in a user picture by using a convolutional neural network, calculating a privacy feature vector from the feature matrix by using the convolutional neural network, multiplying the privacy feature vector with the personalized user privacy demand vector, obtaining a personalized privacy feature vector of the user in the picture, and identifying sensitive privacy information in the picture; and S3, predicting a leakage risk of the personalized privacy feature vector. The application is designed based on a deep neural network, solves the problem of privacy risk assessment from the perspective of user shared privacy image historical behavior characteristics, measures the risk size of user privacy information leakage, reduces the user privacy information leakage risk from the source, and enhances the user privacy consciousness.

The invention is suitable for the technical field of optical computing, and provides a knowledge distillation-based on-chip optical neural network design method, which comprises the following steps of: constructing an initial population of an optimization algorithm based on a preset process constraint condition; the plurality of individuals in the initial population are in one-to-one correspondence with the plurality of groups of optical neural network parameters; performing iterative optimization on the initial population by using an optimization algorithm to obtain an optimal optical neural network parameter meeting a preset process constraint condition; constructing an optical neural network based on the optimal optical neural network parameters, and taking the constructed optical neural network as an on-chip optical neural network; the calculation process of the fitness value of the individual in the optimization algorithm comprises the following steps: training the optical neural network corresponding to the individual by using the knowledge distillation framework, and determining the fitness value based on the training loss value and the physical parameter of the optical neural network. According to the invention, the area of the on-chip optical neural network is reduced, the complexity of the on-chip optical neural network is reduced, and the performance of the on-chip optical neural network in processing complex tasks is improved.

The application discloses a five-phase permanent magnet synchronous motor harmonic suppression control method based on machine learning, which comprises the following steps: S1, constructing a mathematical model of a five-phase permanent magnet synchronous motor driving system in a rotating coordinate system, and designing a five-phase permanent magnet synchronous motor control system based on a deep neural network; S2, deeply integrating motor physical prior information into a training process to enhance the generalization ability of the model, and proposing a hybrid optimization strategy to improve the fitting precision of a predicted network voltage and an actual working voltage, suppress fluctuations in the prediction process, and make the predicted value more suitable for actual operation requirements. The application overcomes the limitations of the existing double-plane current control method, and fully utilizes the technical advantages.

The invention relates to a method for extracting the power consumption of an individual device within a set of devices (E1, E2, E3,..., En), the method comprising transforming a stream of measurements of an overall power consumption into a time series, and dividing (S1) the time series into a set of sequences, and then, for each sequence, the steps of: - generating (S4) a positional encoding matrix (MEP); - extracting (S5) features from the sequence to form a feature matrix (MF); - projecting (S6) statistical metrics into a statistics vector (MS); - concatenating (S7) these various elements (MEP, MF, MS) to form an input matrix (ME) provided to a transformation neural network designed to infer (S8) a time sequence corresponding to the individual device.

The application discloses a kind of based on deep reinforcement learning's decoy penetration aircraft cooperative trajectory planning method, it is related to aircraft path planning technical field, including: the aircraft motion trajectory model of penetration aircraft and decoy aircraft is constructed;Penetration aircraft process Markov decision model is constructed, including state space model, action space model and reward function, and reward function is designed based on threat area;Using deep deterministic policy gradient (DDPG) algorithm, penetration framework based on actor-critic (Actor-Critic), by improving deep Q network, the Actor network and Critic network of penetration aircraft and decoy aircraft are designed using deep neural network, according to aircraft penetration process Markov decision model, heading angle change is obtained, the position of aircraft is updated after being substituted into trajectory model, and motion trajectory is obtained.The application has excellent exploration ability, and can solve the aircraft trajectory planning problem of higher real-time requirement in dynamic environment.

The application discloses a kind of unmanned aerial vehicle trajectory tracking methods based on reinforcement learning and event triggering, belongs to unmanned aerial vehicle automatic control technical field.Establish the six degrees of freedom model of unmanned aerial vehicle, given desired tracking trajectory, according to the desired position, attitude angle, speed and angular velocity obtained according to given trajectory;Trajectory tracking error calculation: the error between actual trajectory and desired trajectory is calculated, filter tracking error is defined, and it is derived;Event-triggering mechanism design is carried out to input, state quantity is designed, and reasonable event-triggering parameter is designed;Evaluator neural network design, design the tracking performance of evaluation system, for tracking performance improvement;Actuator neural network design, approximation uncertain term in model;Model nonlinear control law calculation is obtained in combination with neural network.The method of the application is based on reinforcement learning and event triggering method, realizes optimal control, solves the problem of long-time hovering actuator loss and communication burden of unmanned aerial vehicle under parameter uncertainty and external disturbance.

This invention provides a method and related apparatus for designing a neural network. The method includes: obtaining necessary bit data from a bitstream file of an initial neural network to obtain a set of necessary bit data; the necessary bit data is configuration memory data that affects the performance of the initial neural network; flipping each necessary bit data in the set of necessary bit data to obtain updated necessary bit data, and obtaining an updated bitstream file based on the updated necessary bit data; and obtaining a final neural network based on the updated bitstream file. The neural network designed by this method has high accuracy.

The invention discloses a circRNA (Ribonucleic Acid) and miRNA (Micro Ribonucleic Acid) interaction prediction system and a circRNA and miRNA interaction prediction method of a graph Fourier pulse neural network. The method comprises the following steps: on the basis of high-throughput sequencing omics data of complex diseases, constructing a heterogeneous biological information network containing drugs, diseases, proteins, circRNA, miRNA and lncRNA; converting the topological features of the entities into a unified feature space by using a graph convolutional network; designing a pulse graph neural network in combination with Fourier coding and a pulse neural network, and extracting a topological structure and high-order semantic features in the network; fusing sequences, topologies and semantic features of circRNA and miRNA through a gate multilayer perceptron to obtain embedding features of circRNA and miRNA; and finally, the interaction of circRNA and miRNA is predicted by adopting a Bayesian classifier. According to the method, heterogeneous biological information is modeled from the perspective of network science, Fourier coding, spiking neurons and graph embedding learning are utilized, the action mechanism of circRNA and miRNA in complex diseases can be disclosed, and the method has good practicability and application prospects in the fields of artificial intelligence, life science, clinical medicine and the like.

A method for extracting the power consumption of an individual piece of equipment within a set of equipment (E1, E2, E3, …, En), comprising transforming a stream of measurements of overall power consumption into a time series, and dividing this series (S1) into a set of sequences. Then, for each sequence, the following steps are performed: generation (S4) of a positional encoding matrix (PEM); extraction (S5) of features from the sequence to form a feature matrix (FM); projection (S6) of statistical metrics into a statistics vector (SM); and concatenation (S7) of these different elements (PEM, FM, SM) to form an input matrix (IM) provided to a transformation neural network designed to infer (S8) a time sequence corresponding to the individual piece of equipment. Figure for the abstract: Fig. 1

The present application relates to the technical field of robot control, in particular to a self-adaptive formation control and obstacle avoidance method for distributed multi-robot cluster, which obtains state information of neighbor robots, target points and environmental obstacles through wireless communication and sensors, constructs an error function describing distance and speed deviation based on the information, designs a formation control and obstacle avoidance strategy with adaptive ability, uses RBF neural network to design an adaptive controller, outputs local obstacle avoidance and speed deviation items according to the changes of neighbor robot and obstacle information, and adopts three-dimensional space description, three-layer structure, three neurons in the input layer, six neurons in the hidden layer and one neuron in the output layer for the RBF neural network, and then generates a complete control strategy by weighting and summing the RBF neural network output and the neighbor robot state, so that the robot cluster can effectively avoid obstacles while maintaining formation, and the adaptive ability of the system to environmental changes is significantly improved.

The invention relates to a device for the automatic recognition of at least one visual signature, the device comprising a detector designed to receive an input image and optionally return at least one region of interest, the detector comprising a convolutional neural network designed to receive the input image as input and to return, as a region of interest, a portion of the input image, the convolutional neural network having been trained to detect the potential presence of a visual signature in the image received as input, the detector further comprising an estimator designed to receive an input region of interest and at least one image comprising a visual signature to be recognized, and to return a value representative of the presence, in the input region of interest, of the visual signature to be recognized. The estimator comprises a triplet neural network designed to receive as input an image and to return as output an associated vector, the triplet neural network comprising three sub-networks of the same architecture and having identical connection weights between layers within each respective sub-network, which sub-networks comprise a transformer designed to receive said image and to output the associated vector. The estimator is designed to determine a representation vector from the vector associated with the input region of interest and to calculate a value representative of the similarity between this representation vector and each representation vector determined from the vector associated with an image comprising a visual signature to be recognized, and to return the identifier of the image comprising a visual signature to be recognized for which the similarity value is the highest and exceeds a chosen threshold. The device is designed, on the basis of one or more input images, to call the detector to optionally obtain at least one region of interest for each input image, and to call the estimator with each region of interest returned by the detector and with one or more images comprising a visual signature to be recognized.

The invention discloses a celestial body physical parameter prediction method and system, and relates to the technical field of astronomical data processing and analysis, and the method comprises the steps: obtaining a photometric image and observation spectrum data of a to-be-measured celestial body; inputting the photometric image and the spectral data into a lightweight multi-modal parameter prediction model, performing coarse-grained feature extraction on the photometric image through two layers of large-kernel inverted bottleneck 2D convolution blocks, and performing coarse-grained feature extraction on the spectral data through a 1D convolution operator; aligning the granularity of the extracted coarse-grained features through a dimension transformation module, and performing multi-modal feature fusion on the two groups of aligned coarse-grained features to obtain fine-grained depth semantic features; the fine-grained depth semantic features pass through a parameter regression module, and a parameter prediction result is output; according to the method, the calculation consumption in the reasoning process is reduced by adopting a lightweight neural network design, and the accuracy of a prediction result is improved through feature information complementation among different modalities.

This invention discloses a pre-time formation switching control method for heterogeneous UUVs in confined environments based on mobility perception. The method includes: establishing a heterogeneous UUV formation switching system model; modeling the confined area as a time-varying curve virtual pipeline and defining a mobility index as the trigger for formation switching; constructing a data-driven neural network to approximate uncertainties and disturbances; designing a pre-defined time restraint controller for the actual root leader; generating a reference trajectory from the virtual root leader; and determining the leader formation's position and velocity based on the actual root leader information; designing a pre-defined time affine switching observer to observe the desired state of followers; and designing a full-state pre-time controller by combining a unified obstacle function and a neural network, ultimately achieving formation switching control. This invention integrates core technologies such as mobility perception, pre-time control, a unified obstacle function, and data-driven neural networks to construct a complete heterogeneous UUV formation switching control scheme, achieving comprehensive improvements in environmental adaptability, control accuracy, convergence performance, operational safety, and formation cooperative robustness.

The invention provides a finite time cross-domain aircraft cluster formation control method based on a neural network, and relates to the field of aircraft control, and the method specifically comprises the steps: building a unified nonlinear non-strict feedback system for the dynamic characteristics of cross-domain aircrafts in different domains; collecting real-time flight states and position information of a plurality of followers; designing an improved Tan type nonlinear mapping function, and identifying an initial output constraint type of the system to obtain an unconstrained new variable; constructing a dual neural network combining a shared network and an exclusive network, designing a backstepping method, combining a switching function and a non-singular fast finite time control strategy, and controlling the output of a dual neural network structure; and a switching threshold event triggering mechanism is designed, so that the unmanned aerial vehicle group has different responses in different environments. According to the technical scheme, the problem that mode switching, dynamic coupling and cooperation consistency in cross-domain cooperation are not considered in a formation control method in the prior art is solved.

The invention discloses a zero neural network design mode with preset time convergence and robustness, and belongs to the technical field of neurodynamics and automatic control. The method is suitable for solving the time-varying linear matrix equation TV-LME and tracking and controlling the trajectory of the mechanical arm. According to the method, a composite structure predetermined time convergence robust activation function PTCR-AF is developed based on a TV-LME time-varying error model, and convergence of different amplitude errors is adaptively accelerated through parameter adjustment. A predetermined time convergence robust zero neural network PTCR-ZNN is constructed by using the PTCR-AF and an error dynamic evolution mechanism, and global convergence can be realized in an explicit time upper boundary under undisturbed and bounded disturbance conditions. The time and space complexity of the method only depends on the TV-LME dimension, and the expandability of medium and large scale problems is achieved. Simulation shows that the PTCR-ZNN can converge according to preset time under different orders of TV-LME and keep high precision under strong disturbance; the mechanical arm controller based on the model achieves high-precision trajectory tracking and stable disturbance suppression, and the engineering application value is remarkable.

The present application relates to a kind of machine learning-based groundwater funnel area monitoring data processing method, belong to artificial intelligence technical field.It includes the following steps: groundwater level and space-time auxiliary feature data are collected to construct historical data set, then through the strategy of fusion time-space correlation and periodic priori, data interpolation denoising is completed, and regular sequence is obtained;Subsequently, multi-scale periodic-trend decomposition and time-frequency analysis are realized by adaptive sliding window, and multi-dimensional feature matrix is formed by splicing;Then, the spatiotemporal graph neural network of multi-scale spatiotemporal graph convolutional encoder and physically guided attention decoder is constructed, and the joint loss function of mean square error and physical regularization term fusion is designed;After obtaining the optimal model by Adam optimizer training, it is deployed to monitoring system to realize water level real-time prediction, and the development risk of funnel area is analyzed in combination with historical data and warning threshold, and early warning is issued.The present application can improve prediction accuracy and robustness, and provide accurate decision-making basis for groundwater funnel area management.

The invention belongs to the technical field of graph neural networks, and particularly relates to a graph neural network-based sampling strategy dynamic optimization method, which comprises the following specific steps of S1, constructing an auxiliary graph neural network: designing a graph neural network as a sampling strategy learning device to construct the auxiliary graph neural network; the graph neural network comprises a graph convolutional network and a graph attention network; s2, calculating node sampling probability: inputting the original graph data into the auxiliary graph neural network, and obtaining the sampling probability of each node through forward propagation calculation; the calculation of the sampling probability is based on the comprehensive understanding of the auxiliary graph neural network on the position of the node in the graph structure, the connection relation between the node and other nodes and the characteristics of the node. According to the method, the sampling strategy is dynamically optimized, so that the training efficiency of the graph neural network on large-scale graph data can be remarkably improved, and meanwhile, the prediction performance of the model on various tasks is kept and even improved.

The invention discloses a nonlinear unmanned aerial vehicle formation control method based on preset time reinforcement learning. The nonlinear unmanned aerial vehicle formation control method comprises the steps of 1, constructing a nonlinear dynamic model of following unmanned aerial vehicles; 2, designing a position tracking control strategy based on the quadratic performance index function, a Hamilton-Jacobian-Bellman equation, a preset time reinforcement learning method and an evaluation neural network; 3, designing an attitude tracking control strategy based on the quadratic performance index function, a Hamilton-Jacobian-Bellman equation, a preset time reinforcement learning method and an identification-evaluation neural network; and 4, based on the target reference signal sent by the pilot unmanned aerial vehicle, the nonlinear dynamic model, the position tracking control strategy and the attitude tracking control strategy, the following unmanned aerial vehicle is controlled to realize formation flight, the response speed and the formation precision of the unmanned aerial vehicle formation can be improved, and the control energy consumption is reduced.

Disclosed are systems and methods for accelerated reconstruction of dynamic, undersampled MRI data using a temporally-aware neural network system. An exemplary method includes receiving, for each of a plurality of time frames, undersampled k-space data; for one of the plurality of time frames, generating a reconstructed image by processing the undersampled k-space data for the time frame together with prior information derived from at least one previous time frame of the plurality of time frames, wherein the prior information comprises at least one of: (i) a k-space memory representing previously estimated or acquired k-space data, or (ii) an image-domain memory representing previously reconstructed images.

Techniques for a BLE-CS processing architecture using a parametric, data-driven neural network design for phase-based distance (PBR) measurement are described. The neural network can integrate feature transformation and distance estimation to simultaneously generate a clean spectrum and distance estimate. The neural network can receive PBR measurement data from constant-tone signals across a frequency range exchanged between two devices. From the PBR measurement data, the neural network can extract features representative of non-integer frequencies across this frequency range. These non-integer frequencies can be sampled at non-fixed positions and in ascending order. Based on the extracted features, the neural network can estimate the distance between the two devices.In a single embodiment, the neural network can combine scene identification, denoising, feature transformation, and distance estimation steps into a single model. The neural network can adapt to different indoor or outdoor scenes without requiring an explicit scene identification step.