31 results about "Minimization algorithm" patented technology

The invention provides a method for assisting a movable antenna to realize secure communication by STAR-RIS, and relates to the technical field of wireless communication. According to the method, the safety and the rate of the system are maximized by jointly optimizing the beam forming of the base station, the position of the movable antenna and the STAR-RIS coefficient matrix. The method specifically comprises the following steps: constructing a downlink MIMO system and establishing a joint optimization problem; decomposing the joint optimization problem into three sub-problems by adopting an alternate optimization framework; respectively utilizing a particle swarm optimization algorithm and a weighted minimum mean square error method to be combined with a Majoritization-Minimalization algorithm to solve each sub-problem, and respectively utilizing the particle swarm optimization algorithm and the weighted minimum mean square error method to be combined with the Majoritization-Minimalization algorithm to solve each sub-problem; and alternately executing related steps until system security and rate convergence to obtain a solution of the joint optimization problem. By the adoption of the method, the safety communication performance of the system is remarkably improved, full-space coverage and precise beam regulation are achieved, the algorithm convergence is high, the solving efficiency is high, and the method is suitable for a downlink system with multiple users and multiple eavesdropping nodes.

The invention relates to the technical field of artificial intelligence and energy system operation control, and discloses an artificial intelligence-based energy dynamic aid decision-making method and system, and the method comprises the steps: collecting multi-source heterogeneous data of an energy system, and constructing an observation vector; decoupling the observation vector into an environmental causal factor and an environmental noise factor of a potential space by using a variational structured causal characterization model; pruning a causal factor based on an accumulated anti-fact regret value gradient to obtain effective decision information, generating an anti-fact trajectory by applying intervention to a noise factor to construct an expanded belief state set, and updating a control strategy by using a continuous anti-fact regret minimization algorithm on this basis; and finally, performing physical constraint verification on the sampling action and executing the physical constraint verification. Through causal decoupling and an anti-factual reasoning mechanism, decision robustness, interpretability and physical safety of an energy system in a dynamic non-stationary environment are effectively improved.

The invention provides a communication perception fusion waveform design method based on space-time constructive interference, which belongs to the field of wireless communication and comprises the following steps: establishing a downlink multi-user communication perception fusion system model; establishing a communication model and a sensing model; establishing an optimization problem for minimizing the transmitting power, wherein the optimization problem has a space-time constructive interference constraint condition and a CRB constraint; after the CRB constraint of the optimization problem is relaxed into linear constraint by using a continuous convex approximation technology, a second optimization problem is obtained through iterative solution, the second optimization problem is simplified through a Lagrangian multiplier method and a Carush-Kuhn-Tuck condition, and the simplified second optimization problem is solved through iterative solution of a main optimal minimization algorithm; and obtaining a designed transmitted waveform vector. According to the invention, the symbol error rate of transmission and the transmitting power of the transmitting end can be effectively reduced under the condition that the communication and sensing performance requirements are met.

The present invention provides a method for selecting cache nodes in an Internet of Vehicles (IoV) under complex scenarios, and belongs to the technical field of IoV. It solves the technical problems of communication interruption and cache node load imbalance caused by obstacles blocking the link and multiple requests from a single vehicle. Its technical solution is: including the following steps: S1: constructing three models; S2: proposing a link quality assessment algorithm; S3: based on the file load upper limit of the cache vehicle, dynamically allocating file requests, and proposing a cache node load balancing algorithm; S4: preferentially selecting cache nodes and task nodes, and proposing a cache node minimization algorithm; S5: giving a cache node selection algorithm process under complex scenarios. The beneficial effects of the present invention are: the present invention considers the maximum number of requests that can be served by the cache vehicle as a load constraint, dynamically allocates file requests, and achieves load balancing.

The application discloses a 5G communication network data dynamic privacy publishing method and system based on reinforcement learning, and belongs to the technical field of information security. The steps of the application are as follows: 1, determining the interval of the data to be published histogram; 2, using a data stream sampling method based on a sliding window to sample the data in the sliding window at the current time to obtain a sampling set M; 3, obtaining the approximate statistical result and the noisy statistical result of all intervals in the sliding window at the current time; then, according to the obtained result, using an error minimization algorithm based on the comparison between the approximation and the noise to obtain a better result; 4, using a grouping noise adding method based on reinforcement learning to obtain a locally optimal grouping result for noise adding, and obtaining the finally publishable noisy histogram data. When the application is used for the differential privacy histogram publishing of the sliding window data stream, the application can quickly and accurately publish the data meeting the user demand, improves the data availability and reduces the privacy budget.

A modelling method for making a virtual model of a user's head includes operations of: acquiring a plurality of depth images of a user's head in different head poses; processing the acquired depth images; localising on the two-dimensional images first marker points corresponding to characteristics of the head; estimating the head by estimating the yaw, pitch and roll angles of the head and the positioning of the head along three axes of a reference system of Cartesian axes for each of the two-dimensional images of the depth images, performing a transformation of the head pose; projecting the first marker points located in the respective cloud of points; associating iteratively in pairs the clouds of points until obtaining a cloud of synthesis points; aligning the cloud of synthesis points; and deforming the statistical head model by means of a minimization algorithm of a predefined cost function.

The application discloses an evolution modeling and prediction method and system for extreme risk coupling of an aviation hub network, and the method comprises the following steps: acquiring static section data of an aviation transportation system; constructing a space coupling dynamics model according to the static section data, wherein each airport node in the space coupling dynamics model corresponds to an attraction weight, and the attraction weight represents the co-mobility characteristics of each airport node under an extreme load state; using a maximization minimization algorithm to inverse the attraction weight; based on the inverted attraction weight, acquiring a space core parameter group through a Markov chain Monte Carlo algorithm, and obtaining regional congestion co-mobility early warning information based on the space core parameter group. The application can accurately simulate the dynamic growth logic of a multi-center airport group, and can more quantitatively identify the cascade load pressure distribution within the space dependence under the extreme scenario of the failure of a specific core hub.

This invention discloses a method for 3D modeling of small celestial bodies based on SFS (Self-Focused Spectroscopic Method), belonging to the field of intelligent 3D modeling and image reconstruction technology for small celestial bodies. To improve the accuracy of 3D modeling of small celestial bodies, this invention performs modeling under the condition that the camera's extrinsic parameters are determined. This includes setting the positions of the camera and the small celestial body in a simulation environment and setting the camera's motion trajectory to obtain multiple frames of small celestial body observation images; extracting the contours from the small celestial body observation images; transforming the 3D modeling problem of small celestial bodies into a pseudo-Boolean function minimization problem; applying local minimization and hierarchical minimization algorithms to solve for the minimum value of the pseudo-Boolean function to obtain a 3D voxel model of the small celestial body. For modeling under the condition that the camera's extrinsic parameters are uncertain, this includes converting the 3D voxel model of the small celestial body into a polygon model; selecting 3D model points; matching the 3D model points with 2D image points; and applying the LM algorithm to optimize the matching error between the 3D model points and 2D image points. This patented method is applied to the process of small celestial body detection.

The invention relates to the technical field of radar observation, and discloses a radar observation intelligent collaborative assimilation method, which comprises the steps of constructing a Python-Fortran hybrid programming platform, commanding a weather forecast mode simulation environment, generating multi-scale background field data, and enabling the programming platform and a weather forecast mode to share data mutually. A variational auto-encoder and a decoder are used to realize rapid mutual conversion between a physical space and a low-dimensional hidden space, the encoder realizes data dimension compression through a neural network, and the decoder realizes data physical reconstruction through the neural network; in the hidden space, radar observation is fused into a mode background by using a minimization algorithm to realize multivariable nonlinear assimilation. The physical environment state after assimilation is output to form a continuous time sequence which can be used for three-dimensional environment monitoring service, and an assimilation result at the latest moment is used as an initial state of a weather forecasting mode, so that the weather subjected to numerical simulation naturally expands to the future in a time domain, and forecasting service can be formed. According to the method, the efficiency and precision of three-dimensional environment monitoring and forecasting are improved.

The invention provides a self-anti-S box construction method and system based on m-sequence generation and logic optimization, and belongs to the technical field of cryptography and hardware security. The method comprises the following steps: firstly, constructing an n-level linear feedback shift register according to a set S box number n and a primitive polynomial, generating an m sequence in a non-all-zero initial state, and adding an all-zero state at the tail; a self-reverse S box lookup table is constructed through a symmetric mapping strategy, so that the ith element in the sequence and the element at the symmetric position form mapping, and the same S box is used for encryption and decryption. Thirdly, the lookup table serves as a Boolean function truth table, is simplified into a simplest expression through a logic minimization algorithm and is decomposed into a layered logic circuit; during actual processing, N data blocks are converted into a bit plane format from block type storage, then parallel operation is carried out through a hierarchical circuit, and finally a result is transposed back to block type storage. According to the method, the same S box hardware module can be reused in the encryption and decryption process, and hardware resources are remarkably saved.

The invention belongs to the technical field of circuit fault diagnosis, and particularly relates to a solving method and device based on multiple observation aggregation fault diagnosis and a medium. In order to solve the problems of one-sided single observation information, redundant diagnosis results and the like, multiple times of fault observation are performed on a circuit system, and multiple independent diagnosis results are generated and integrated into a diagnosis set; establishing a mapping relation between a diagnosis set and a unique keyword by using a map function of the STL, and converting the component combination into an integer identifier to simplify operation; calling a BWSS algorithm to perform collision calculation on the mapped keywords to obtain a candidate solution set covering all observation scenes; after the keyword solution set is converted to an original diagnosis form, deleting a superset according to a redundancy rule to form a simplified intermediate diagnosis solution set; and finally, through a branch and bound minimization algorithm, according to potential classification and ascending order arrangement of the diagnosis sets, high-order supersets are deleted layer by layer by adopting a heuristic strategy, and finally, a minimum diagnosis solution set of all faults is output.

This invention discloses a semi-blind OFDM receiving method that combines channel estimation, phase noise compensation, and signal detection. The method first constructs a constrained objective function for joint channel estimation, phase noise compensation, and signal detection based on phase noise modulo-1 constraints and the mapping constraint of the signal to be detected. Then, given the phase noise and the signal to be detected, the optimal channel estimate is obtained, and this estimate is brought back to the original optimization objective to obtain an updated objective function. Subsequently, by introducing a penalty term, a penalty factor, and auxiliary variables, an augmented Lagrangian function of the optimization objective is constructed, and the corresponding optimization problem is solved using the cross-direction multiplier method (ADMM) and the minimization-maximization (MM) algorithm. When the iteration termination condition is met, the joint estimate of the channel, phase noise, and transmitted signal is obtained. This invention ensures optimal performance from the optimization objective and can effectively save pilot overhead, thereby improving spectral efficiency.

The invention belongs to the field of laser welding numerical simulation, and discloses a laser welding numerical model high-precision efficient calculation method and system based on a physical information neural network. Generating model basic training data based on the laser welding multiphase flow numerical model; a 3D ConvLSTM-based laser welding physical field prediction model is constructed, energy input information is introduced into a hidden layer, and the randomness of the laser welding process is restored; designing a loss function containing various error sources; adjusting the network weight by using a gradient minimization algorithm, and minimizing a loss function; and inputting the three-dimensional physical field data of T moments to predict the change of subsequent T moments, extracting the data of (T-T) moments after the input data and the output data to reform a model for input, and predicting the physical field change of the subsequent moments. The invention provides a high-precision and high-efficiency novel solving method for laser welding numerical model calculation.

The application discloses an industrial surface defect detection method and system based on sampling disturbance optimization and history reuse, and relates to the technical field of industrial visual defect detection. In view of the problems that the background of an industrial defect image is complex, the loss topography is highly non-convex, which leads to model overfitting, and the traditional sharpness perception minimization algorithm relies on double back propagation, which leads to excessive algorithm cost, a lightweight optimization method is provided. The method introduces a sampling sharpness estimation mechanism in the network parameter update stage, generates a disturbance tensor by sampling within the preset norm boundary to explore the flat minimum value, and introduces a history reuse mechanism to dynamically evaluate and reuse the historical state, so as to bypass the back propagation calculation of the current calculation graph in part of the training period. With less memory cost, the application greatly reduces the disturbance calculation cost, effectively crosses the non-convex loss valley, significantly improves the generalization accuracy of industrial surface defect detection, and has a wide application prospect.

This invention discloses a semi-blind direct localization method for thermal noise communication systems, comprising: based on the frequency domain received signal model of the thermal noise communication system, utilizing the statistical characteristics of the received signal to construct a maximum likelihood estimation problem concerning the mobile station position and channel fading coefficient, obtaining an objective function containing logarithmic determinant terms and trace function terms of a matrix; the objective function simultaneously includes a contribution from pilot power and an expected contribution from the power of the received data signal; using an optimization minimization algorithm framework, alternately solving for the mobile station position and channel fading coefficient until a preset convergence condition is met, obtaining the localization result; wherein, by constructing a substitution function containing the statistical characteristics of blind pilots and data, a closed-form solution for the channel parameters is derived in each iteration, and the mobile station position is updated through a two-dimensional grid search. This invention mines the statistical information of the data portion, effectively improving the utilization rate of the received signal and increasing the localization accuracy under low signal-to-noise ratio conditions.

The invention provides an automatic driving key scene generation method and system based on a nested game algorithm, relates to the technical field of automatic driving, and aims to solve the problems that a single game model is adopted in the prior art, strategy updating is not clear, calculation complexity is high, the convergence speed is low, and interpretability is poor. The method comprises the following steps: constructing an automatic driving antagonism test environment, respectively constructing game models of two stages for vehicle driving, determining a game player set, an action space, a revenue function and a constraint condition, and performing optimization test on the automatic driving antagonism test environment by using a nested anti-fact regret minimization algorithm. And obtaining an optimized adversarial test environment, performing a simulation test on the optimized adversarial test environment in a simulation environment, and generating a key safety scene for testing the safety of automatic driving. The problems in the prior art are solved, the representativeness and diversity of the generated scene are improved, and the efficiency and accuracy of automatic driving safety verification are improved.

The invention discloses a high mobility scene channel estimation and prediction method based on a channel map, and relates to the field of wireless communication system research, and the method comprises the steps: building channel key parameter prior distribution based on a power spectrum density function provided by the channel map; constructing and converting a pilot frequency training model considering inter-carrier interference into a parameterized linear estimation problem; and iteratively solving the problem by using a maximization-minimization algorithm, and obtaining a current channel estimation value and a subsequent channel prediction value based on a channel multipath structure. Compared with an existing method, the method has the advantages that the channel estimation precision can be improved by 10 times, high-precision time domain channel prediction can be realized, and a high-precision channel estimation and prediction solution is provided for high-mobility scenes such as high-speed railways and unmanned aerial vehicles. Instantaneous channel state information estimation of a double-selective channel can be realized, and time domain channel prediction in a quasi-static time interval can be realized.

The invention discloses a wireless spectrum map data completion method, which comprises the following steps of: modeling spectrum data containing missing data and mixed noise into a three-dimensional block term tensor decomposition model according to three dimensions of two-dimensional space and frequency and a spectrum situation superposition model; in order to limit the influence of abnormal data such as impulse noise, a non-convex optimization function based on a self-adaptive threshold value is provided, so that a spectrum data restoration model based on low-rank tensor completion is constructed; converting the non-convex optimization problem into an optimization problem of a convex function by adopting a semi-quadratic theory; and according to a block continuous upper bound minimization algorithm framework, a multivariable optimization problem is converted into an optimization problem of each single variable, and a projection gradient method is utilized to iteratively optimize each factor until a convergence condition is met. According to the method, an accurate spectrum map can be obtained in a spatial sparse sampling and multi-noise scene, and the urgent demand for high-precision spectrum situation sensing in a complex noise environment is met.

The application discloses a personalized advertisement content accurate delivery method and system based on user portrait, and belongs to the technical field of intelligent promotion, which combines gender / age and other user portrait features and real-time elevator scene data, dynamically calculates advertisement playing benefit score R, realizes accurate matching of advertisement content and current audience, and improves the conversion efficiency of single advertisement. In addition, the application adopts a double evaluation mechanism of benefit score sorting and advertisement importance, dynamically generates an optimal playing scheme through an H1 value minimization algorithm, guarantees the exposure rights and interests of high-value advertisements, and can optimize the playing strategy in real time according to historical effects. The application introduces a completion proportion standard deviation monitoring, dynamically adjusts the playing order, balances the progress differences of various advertisements, and effectively avoids the long-term accumulation of some advertisements due to low benefit score.

The application discloses an improved traveling wave distance measurement method for fault location analysis of a power distribution network, and relates to the technical field of power distribution networks. The method solves the problem of positioning deviation caused by a traditional single wave speed model, and especially the problem of false fault points caused by refraction and reflection at the connection between a cable and an overhead line. The method performs waveform translation and coincidence verification, calculates the amplitude difference average under different translation amounts, automatically identifies the optimal time alignment point, and locks the associated time difference t1. This process can effectively correct the time error caused by signal distortion, non-synchronous sampling and the like, and can improve the time difference accuracy from the microsecond level to the nanosecond level, thereby laying a foundation for subsequent accurate positioning. In the locked range interval, a plurality of associated points are randomly selected, a corrected waveform is constructed based on the attenuation ratio, and repeated verification is performed. The optimal solution is selected through a characteristic mean minimization algorithm. The mechanism can effectively suppress single measurement point error, and can improve the robustness of the positioning result through statistical average effect.

The application provides a method for realizing safe communication by STAR-RIS assisted movable antenna, and relates to the technical field of wireless communication. The method maximizes the safety and rate of the system by jointly optimizing the base station beam forming, the movable antenna position and the STAR-RIS coefficient matrix. Specifically, the method comprises the following steps: constructing a downlink MIMO system and establishing a joint optimization problem; using an alternating optimization framework to decompose the joint optimization problem into three sub-problems; using a particle swarm optimization algorithm, a weighted least mean square error method combined with a Majorization-Minimization algorithm to solve each sub-problem; and alternately performing relevant steps until the system safety and rate converge, to obtain the solution of the joint optimization problem. The application uses the above method, significantly improves the system safety communication performance, realizes full space coverage and accurate beam control, and has strong algorithm convergence and high solution efficiency, and is suitable for a downlink system with multiple users and multiple eavesdropping nodes.

The invention belongs to the field of image processing circuits, and particularly provides a hyperspectral image target detection circuit based on a constraint energy minimization algorithm innovatively from the perspective of an analog circuit, so that one-step hyperspectral image pixel point target detection of a hardware circuit is realized, and the defects in the prior art are overcome. The built integral circuit model can be divided into two parts, one part is an optimal filter coefficient calculation module realized through an integral negative feedback circuit, and the other part is an FIR linear filter module realized through a multiply-accumulate circuit. According to the designed circuit model, the actual calculation effect of the pixels of the hyperspectral image data set can be further studied if the spectral information of the target is known in the target detection process. The circuit can be applied to scenes such as remote sensing monitoring and environment detection, hardware implementation of the circuit has important engineering application value and development potential, and core intellectual property right can be formed in the field.

A patch-based deblurring method to deblur images blurred in a non-uniform manner (e.g., motion blur or camera shake) uses a modified total-variation (TV) minimization algorithm to make a network specifically designed to handle non-uniform blur.

The invention discloses a multi-component collaborative vibration monitoring and fault positioning system for a transmission chain of a wind turbine generator, and relates to the technical field of state monitoring and fault diagnosis of wind power generation equipment. An improved order analysis method based on second-order fitting and absolute value nearest neighbor matching is adopted to obtain a stable order spectrum resisting rotation speed fluctuation; after an observation feature matrix is constructed, an SCAD penalty function is introduced to establish a sparse non-negative matrix factorization model, efficient and stable solution is carried out by adopting a near-end alternating linearization minimization algorithm, and accurate decoupling of fault features of multiple components such as a main bearing, a gearbox and a generator is realized; and finally, accurate fault positioning and identification are realized through cosine similarity matching. According to the method, the problems of coupling, aliasing and feature extraction of fault signals of multiple parts under the variable rotating speed are effectively solved, and the diagnosis accuracy, the coverage range and the positioning precision are remarkably improved.

The invention relates to the technical field of hydraulic machinery operation prediction and control, in particular to an Informer-based pump turbine pressure pulsation prediction method, which comprises the following steps of: obtaining a real-time pressure pulsation signal acquired by a pressure sensor arranged on a pump turbine, and sequentially carrying out median filtering method denoising and normalization preprocessing on the real-time pressure pulsation signal; a whale optimization algorithm WOA is adopted to optimize VMD feature parameters based on a signal reconstruction error minimization algorithm; a whale optimization algorithm WOA is adopted to optimize Informer model parameters based on a signal reconstruction error minimization algorithm, target parameter combination is adopted to predict each IMF component, and all predicted IMF components are reconstructed; and calculating a prediction result index, and performing an Informer model adaptive updating mechanism or re-optimization. According to the method, the problems that the performance of a prediction model is unstable and the prediction model is caught in local optimum in the prior art are solved.

The invention discloses a fluctuation target detection method applied to a communication and induction integrated system, and the method comprises the steps: constructing an active IRS-assisted ISAC system, and considering that an active RIS can provide a view angle different from a line-of-sight link (LoS) to observe a target. In order to improve the comprehensive performance and environmental adaptability of the system, the influence of clutters on the sensing function is actively considered, and the clutters are incorporated into a system modeling and optimization framework. Constructing a joint optimization problem model; optimizing a transmitting beam forming matrix of the base station by adopting first-order Taylor expansion and positive semidefinite relaxation; and a reflected beam forming vector of the active RIS is optimized by adopting second-order Taylor expansion and a maximum-minimum (MM) algorithm. A numerical result proves that the fluctuation target detection performance in the ISAC system can be remarkably improved.

Method and device for generating the trajectory of a mobile device between a starting point and a destination point. This trajectory generation method includes assigning (52), to each point of a three-dimensional mesh, a plurality of corridors, each corridor corresponding to a solid angle of arrival at that point, and having an associated corridor index, forming a set of regular nodes, each regular node being associated with a point of the mesh and a corridor index; for the destination point, the calculation (54) of a destination corridor index as a function of a heading and an arrival slope;the determination (60), by a cost minimization algorithm, of a succession of nodes from the set of regular nodes, between a starting link node associated with the starting point and a destination link node associated with the destination point and the destination corridor, and the formation (62) of a trajectory from the points of said succession of nodes. Figure for the abstract: Figure 3;

The invention discloses an offline data caching and artificial intelligence preloading method and system for a data platform, and relates to the technical field of artificial intelligence. The method comprises the following steps: acquiring running state parameters of each cache node, and generating a multi-dimensional performance data set; generating a contribution degree distribution vector; obtaining a load peak point and a historical mean value of the service, and generating a weight distribution scheme; constructing a cost function model, and generating a resource allocation strategy by adopting a cost minimization algorithm; calculating a data migration path among the cache nodes by adopting a genetic algorithm, and generating a cache resource allocation scheme; executing batch data migration operation to obtain an initial system operation parameter set; and generating a multi-dimensional performance data set, and obtaining a final system operation parameter set. According to the method, efficient utilization of resources and stable operation of the system can be achieved, the cache hit rate is increased, access delay is reduced, storage and bandwidth use are optimized, and the overall performance and reliability of a data platform are improved.

This invention discloses a hybrid beamforming method for integrated MIMO communication and radar based on overlapping subarrays. Under the conditions of hybrid analog-digital arrays, including overlapping subarray structure, transmit power, and constant mode constraints, a joint problem of maximizing communication spectral efficiency and minimizing radar beam pattern matching error is established. This joint problem is relaxed into a weighted summation minimization problem, where the designed hybrid beamformer minimizes the Euclidean distance to the optimal communication / radar beamformer. This invention decomposes the hybrid beamforming design problem into three subproblems and develops an efficient alternating minimization algorithm. The proposed hybrid beamforming design based on overlapping subarray structure offers flexibility and effectiveness in terms of spectral efficiency and radar beam pattern performance.

The application discloses a workpiece point cloud fine registration method based on a pseudo-eliminating weighted variance minimization algorithm. Firstly, a CAD model is discretized into a point cloud model, and normal vector calculation and orientation are performed, and a three-dimensional scanner is used to acquire a measurement point cloud; then, a Kdtree algorithm is used to bidirectionally search for a nearest neighbor point pair set, based on the position relationship of the point pair, the measurement points are divided into positive measurement points and negative measurement points; then, a binary symbol is used to unify point-to-point distance and point-to-plane distance to establish an adaptive coordination distance; a weight function is constructed to distinguish normal point clouds, controllable abnormal point clouds and uncontrollable abnormal point clouds; based on the adaptive coordination distance and the weight function, a target function of the pseudo-eliminating weighted variance minimization algorithm is established, and a rigid body conversion matrix of CAD point cloud and measurement point cloud fine registration is solved, and fine registration of the workpiece is completed. The application fully considers matching distortion caused by various abnormal point clouds, effectively suppresses matching inclination, and improves registration accuracy and registration robustness.