119 results about "Reverse mapping" patented technology

The invention relates to the technical field of data security, in particular to a data security sharing method and system. The method comprises the following steps: acquiring original shared data, carrying out reverse mapping and block marking, generating block marking data, carrying out inter-block random arrangement on the block marking data, carrying out differential disturbance simulation based on a random sequence, forming disturbance block data, and encrypting the disturbance block data by applying a progressive encryption technology. The method comprises the following steps: carrying out fragmentation storage, generating storage shared data, carrying out access monitoring on the storage shared data, collecting shared access data, and constructing an evidence chain and generating an access traceability data chain according to the data, thereby realizing the security and traceability of data sharing. According to the invention, a safer and more credible data security sharing method is realized.

The invention discloses a lower limb motion data real-time analysis method and system applied to rehabilitation guidance, and the method comprises the steps: calling a personalized model according to user information, converting real-time motion data into historical motion feature vectors, inputting the historical motion feature vectors into a real-time prediction module, and generating multiple frames of future motion feature vectors through coding and decoding; calculating a deviation value between each future motion feature vector and a preset threshold range to obtain a motion deviation value, calculating a motion risk proportion of each future motion feature vector, performing weighted fusion to obtain a motion risk probability, and generating an equipment parameter adjustment value and pose adjustment information through reverse mapping; and finally, according to risk probability assessment training stage promotion, synchronously displaying correction information and adjusting equipment operation parameters. According to the method, the dynamic model is constructed through pre-training, real-time prediction of future motion features is realized in combination with the lightweight LSTM, and the predictability and personalized adaptation ability of rehabilitation guidance are improved.

Errors in network service provisioning are upstream causes to service disruptions interruptions, and failures experienced by a telecommunications network’s subscribers. These challenges are overcome by the disclosed systems for diagnosing and auditing service provisioning errors in telecommunications networks. By addressing discrepancies between Customer Facing Services (CFS) as selected by network subscribers (e.g., voice calling, messaging, roaming, unlimited or limited broadband access) and the current states of network elements in the telecommunications network, the system ensures accurate service delivery. The system performs real-time diagnosis by comparing expected network service configurations, derived from queried subscriber CFS, with the actual network states or configurations through a reverse mapping process. The system's auditing process reduces operational costs and obviates existing techniques of manual inspection.

The invention aims to provide a multi-condition diffusion model-based Shengtuo calligraphy rubbing fusion restoration method, which comprises the following steps of: firstly, acquiring a plurality of Shengtuo calligraphy rubbing images from different sources, preprocessing the images, and constructing a standardized training data set; designing a multi-image feature extraction and alignment module to obtain a fused structural feature tensor; constructing a core network structure of the multi-conditional diffusion model, and supporting guidance of a plurality of conditional vectors; jointly inputting the obtained fusion structure feature tensor and a plurality of condition guide items into a diffusion model to generate a fused and repaired potential image; and finally, restoring the potential image obtained by reconstruction into a final output image through a reverse mapping module. According to the method, the problems of blurring, pen lacking, breakage and non-uniform styles during fusion and restoration of the rubbing image in the prior art are solved.

The invention discloses an agent construction and development system based on a knowledge graph, and the system comprises a data collection and ontology construction module which is used for generating basic graph data; the polymorphic knowledge cell construction module is used for generating a machine-readable rule and an evidence template; the semantic mapping and deduction module is used for determining forward and reverse mapping modes and setting a consistency threshold value and a violation degree weight; the task analysis and planning module is used for outputting a feasible path and generating a plan sequence and tool arrangement; the execution and correction module is used for generating an evidence packet, constructing a forward chain, carrying out reverse reconstruction by utilizing reverse mapping, generating anti-fact candidates and writing back or backing back the anti-fact candidates; and the version and audit module is used for generating a version fingerprint for the polymorphic knowledge cellular graph which is written back and providing query and audit functions. According to the method, adaptive planning, reliable execution and traceable auditing of the intelligent agent in a dynamic task are realized by constructing polymorphic knowledge cells and cross-layer semantic mapping.

The invention discloses an image return method and device of an endoscope, a storage medium and a terminal. The method comprises the following steps: acquiring an initial image and a triaxial angular velocity signal of the endoscope; according to the triaxial angular velocity signal, attitude calculation is carried out by using a quaternion differential equation to obtain an attitude quaternion, and the attitude quaternion is updated in real time; performing interpolation on the attitude quaternions at adjacent moments within one frame of exposure time to obtain a row-level quaternion and generate a row-level rotation matrix; and based on the row-level rotation matrix, performing return and compensation on the initial image by using reverse mapping and bilinear interpolation to obtain a return image. The problems of universal joint deadlock and function failure caused by the universal joint deadlock in Euler angle attitude calculation are avoided, line-by-line compensation is carried out through an interpolation algorithm, the jelly effect is eliminated, the problems of large delay, insufficient real-time performance and difficulty in integration in a limited space in post-processing of a traditional software algorithm or an external micro-control unit are avoided, and the method is suitable for being applied to the field of Euler angle attitude calculation. And the imaging requirement of clinical operation of the medical endoscope can be better met.

The invention discloses a cross-view multi-target collaborative perception and global tracking method. The method comprises the following steps: receiving original video streams of different views in parallel; carrying out feature extraction and content perception feature recombination processing; establishing a homography transformation matrix representing the mapping relation between the visual angles; performing forward geometric mapping by using the homography transformation matrix, and performing reverse mapping, secondary confirmation and state updating; a joint cost model based on space-time-appearance fusion is constructed, the potential association degree of unassociated residual targets is quantified, and missing targets are recalled; performing duplicate removal and optimization on the targets in the association set; and training and optimizing the network. According to the embodiment of the invention, the method achieves the capturing of a pixel-level target, and improves the texture and edge expression capability of a tiny target in a deep feature map. Efficient fusion of multi-dimensional spatio-temporal information among visual angles is realized; higher robustness and tracking continuity are achieved, and the integrity and fineness of a final trajectory result are improved; and the recall is ensured, and the generation of redundant tracks is avoided at the same time.

The invention discloses a fuzzy video real-time jitter removing method and system, and the method comprises the steps: distributing feature points through multi-scale FAST detection, lightweight MLP detection and dynamic weight, combining a MobileNet-FPN network and the lightweight MLP dynamic weight to extract the robust features of the feature points, modeling the global matching relation of the feature points of two adjacent frames through Transform, and carrying out the real-time jitter removing of a fuzzy video. And a Bayesian adaptive RANSAC algorithm is adopted to realize high-efficiency mismatching pair elimination, and finally an affine transformation matrix is solved and reverse mapping is performed to compensate jitter. According to the method, while the real-time performance of the mobile terminal is ensured, the image stabilization precision in a violent shaking scene is remarkably improved.

The invention discloses a mirror image display image overturning processing method and system for AR equipment. The method comprises the steps that layered mirror image overturning is carried out; performing unified horizontal flipping processing on the input image of the physical camera and the virtual rendering layer through a GPU shader; independently rendering the UI channel; in a UI channel independent of the main image channel, drawing a UI layer according to the original space coordinates; calibrating interactive coordinates; performing reverse mapping on the position coordinate of the user operation event through a coordinate reverse mapping algorithm, so that the position coordinate is matched with the logic position of the UI control which is not overturned; optimizing dynamic performance; and monitoring a system load index in real time, and dynamically adjusting an image processing parameter according to a performance threshold so as to maintain a stable frame rate. The invention provides an efficient integrated solution, and the consistency of vision and interaction is realized through a cooperative processing mechanism of Shader image flipping and multi-channel rendering decoupling design. And the image direction logic consistency, the operation interface readability and the user experience real-time performance are effectively considered.

The invention discloses an unmanned aerial vehicle aerial photography target detection method and device and a storage medium, and relates to the technical field of computer vision. The method comprises the steps of constructing a detection model based on a YOLO framework, replacing a convolution module of a feature pyramid in a backbone network of the detection model with a block local-global attention module, designing a triple feature fusion module in a neck network, and adopting a dynamic detection head fusing a triple attention mechanism at a detection head; the block local-global attention module is used for decomposing an input feature map into position-weighted sub-map blocks and fusing the sub-map blocks, enhancing each sub-map block in a compact feature space in parallel by utilizing unified perception context attention, meanwhile, carrying out implicit modeling on a global context, and finally, recovering an original structure through reverse mapping; and training and verifying the detection model by using the unmanned aerial vehicle aerial image data set. According to the method, high-speed reasoning is kept, and meanwhile, the detection precision of tiny and dense targets is remarkably improved.

The invention discloses a repeated data deletion method and system based on ZNS SSD, and belongs to the field of repeated data deletion.The method comprises the steps that a fingerprint index table and a fingerprint updating cache are maintained, three-time opportunity repeated data deletion is achieved, and repeated data search is conducted in the fingerprint index table, the fingerprint updating cache and an address mapping table respectively; an address mapping table, an address reverse mapping table and a reference counting table are maintained, physical space garbage collection / defragmentation and write request merging are achieved, the address mapping table, the address reverse mapping table and the reference counting table conduct garbage collection with partition as granularity when free space is insufficient, and a fingerprint index table is not updated after garbage collection data migration; and instead, the mapping of fingerprints to migration-in physical pages is stored in a fingerprint updating cache, the physical pages are sequentially distributed in partitions by the fingerprint updating cache, normal data write requests and garbage collection data migration write requests are respectively cached, and the write requests which are continuous in sequence are merged and then issued. According to the method, the high-performance and low-cost efficient storage system can be constructed on the basis of the ZNS SSD at relatively low overhead.

The embodiment of the invention provides an image distortion correction method and device, computer equipment and a storage medium. The method comprises the following steps: acquiring a plurality of first pixel coordinates in an initial correction image, and determining a reverse mapping coordinate corresponding to each first pixel coordinate to obtain an initial reverse mapping table; partitioning the initial reverse mapping table to obtain a plurality of blocks, and sorting the plurality of blocks according to a preset rule to obtain a target reverse mapping table; analyzing the target reverse mapping table, and determining a distortion correction sequence of the initial correction image; and performing distortion correction on the corresponding original image area of each block based on the distortion correction sequence to obtain a target correction image. According to the embodiment of the invention, the reverse mapping table is obtained, and the reverse mapping table is partitioned and sorted, so that the distortion correction sequence of the initial correction image can be determined, the image is processed block by block according to the distortion correction sequence, the whole image does not need to be processed at the same time, and the memory occupation and the calculation cost are reduced.

The invention discloses a distribution network complex fault waveform-oriented feeder automation test method and system, and relates to the technical field of distribution automation, and the method comprises the steps: building a corresponding distribution network topology model according to the grounding mode of a target region distribution network; configuring the distribution network topology model in a preset simulation environment to simulate different fault scenes so as to establish a complex fault waveform library; sequentially carrying out expansion and feature extraction on each waveform in the complex fault waveform library; carrying out sequence arrangement on the extracted fault feature data and carrying out reverse mapping to obtain simulation adjustable parameters, and updating a simulation environment by using the simulation adjustable parameters; and connecting the updated simulation environment to a feeder automation test system, and carrying out an iterative test on the fault of the power distribution network in the target area, so that the response accuracy and reliability of the feeder automation test can be realized.

The invention discloses a memory controller, a memory address processing method thereof and a computing system.The method comprises the steps that a reverse mapping table between a system physical address area and an available DRAM physical address area is obtained through bad area data based on a DRAM, and during running, the memory controller sends a system physical address sent by a processor to the DRAM physical address area according to the reverse mapping table; the method comprises the following steps: dynamically remapping to a DRAM physical address of a non-failure unit, and generating a corresponding control signal for accessing the DRAM, thereby shielding a DRAM physical region containing a failure storage unit, providing a continuous and complete system physical address space for an upper layer, improving the effective utilization rate of the DRAM and the operation stability of a computing system, and improving the reliability of the computing system on the premise of not changing the memory access mode of a processor. And the risk that the whole memory is unavailable due to local storage failure is reduced.

The invention provides a hydraulic engineering operation and maintenance management system based on digital twinning, and relates to the technical field of operation and maintenance management, and the system comprises a hydrological sensitive point screening module which carries out the finite element stress hot spot and hydrological sensitive point screening of a target hydraulic engineering; the sensor layout module is used for carrying out multi-modal sensor layout; the twinning model construction module pushes real-time data to a digital twinning synchronization engine; a comparison loss value set determination module performs comparison loss analysis on the real-time digital twinborn model and a historical digital twinborn model synchronously constructed last time; the reverse mapping module performs reverse mapping on the multi-mode sensor network; the asynchronous sampling module carries out asynchronous sampling on the multi-mode sensor network and carries out operation and maintenance management scheme identification. The technical problem that the sampling frequency cannot be adaptively adjusted according to the real-time change in the prior art can be solved, the technical target of intelligent sampling of the multi-modal sensor is achieved, and the technical effect of improving the key data acquisition precision is achieved.

The invention relates to the technical field of AR interaction control, in particular to an AR glasses cursor control method and system based on head posture induction, and the method comprises the steps of data acquisition, mapping establishment, displacement accumulation and cursor control. Through the AR glasses equipment body, the natural rotation posture of the head of the user is collected and used as an input instruction, and a reverse mapping normal form is constructed, so that the physical direction of cursor movement is opposite to the physical direction of user head rotation, and a non-intuitive chasing relation between the cursor and a target is avoided; intuitive mapping between natural rotation of the head of the user and cursor control of the AR glasses is achieved, accurate control over the cursor can be achieved without additional auxiliary actions or a controller, and convenience and privacy are high.

The invention discloses a multi-modal model training method, a model, an image recognition method and a related product, and relates to the technical field of machine learning, and the multi-modal model training method comprises the steps: generating a multi-modal prompt feature based on multi-modal sample data, carrying out the cross-modal mapping of the multi-modal prompt feature, and obtaining target modal prediction information; determining modal alignment loss through the target modal prediction information and the corresponding sample data; content alignment loss is determined through reverse mapping modal information and corresponding sample data, the reverse mapping modal information is generated by target modal prediction information through reverse cross-modal mapping, and the reverse cross-modal mapping is an inverse process of cross-modal mapping; and optimizing learnable parameters involved in cross-modal mapping, reverse cross-modal mapping and prompt information through modal alignment loss and content alignment loss so as to complete training. According to the embodiment of the invention, the performance of the multi-modal model and the use experience of a user can be improved.

The invention provides an improved ship target distance measurement method and system, and the method comprises the steps: calculating the radial and tangential distortion parameters of a camera and the internal reference of the camera through a camera calibration technology, building a mapping relation through a distortion automatic correction algorithm, carrying out the reverse mapping calculation, carrying out the restoration and filling of vulnerability pixel points, and generating a distortionless correction image. A deep learning model is used to identify a sea level line in a camera video picture, a distortion automatic correction algorithm is used to correct the sea level line into a horizontal straight line, and according to a camera imaging principle, a proportional relation of a tangent function related to a camera lens direction and a sea level angle is constructed. Automatically calculating the installation angle of the camera by using an arc tangent function; and according to the automatically calculated camera installation angle and the pixel point coordinates of the distortionless correction image, the distance between the position of the opposite ship on the corrected sight line video picture and the world coordinate system of the camera is obtained, the ship target distance measurement improvement is realized, and the method has usability and high precision.

The invention provides a parallel conversion and loading method for a multi-channel DDR (Double Data Rate) address interleaving mapping mirror image, which comprises the following steps of: inserting a MemCore bit and a Rank bit into a physical address, and carrying out reverse mapping on the DDR physical address to generate an HIF (High Interleaved Format) address; a Channel bit and a DFI bit are inserted into the HIF address, and an AXI address is generated; converting a system mirror image which needs to be loaded into the multi-channel DDR particle model of the chip simulation platform into a RAW format interleaving mapping mirror image; and based on the AXI address, the chip simulation platform loads the converted interleaving mapping mirror image into the multi-channel DDR particle model according to file blocks through a DDR rear gate. The system bandwidth performance verification efficiency of an SoC prototype based on a chip simulation platform in a DDR multichannel interleaving scene is improved, and bandwidth performance test data in a complex scene is provided for iterative tuning of DDR address interleaving mapping parameters in real time.

The invention discloses a service system migration method and device based on a gateway and a storage medium, and relates to the technical field of data management, and the service system migration method based on the gateway comprises the steps: receiving and analyzing a service request message of a source system, and extracting a feature tuple of the service request message; based on a migration strategy of a migration batch configuration rule, matching the feature tuple, and determining a corresponding target system; mapping the service request message into a corresponding target request message according to an interface mapping configuration file of the target system, and sending the target request message to the target system; and based on a target response message of the target system and the interface mapping configuration file, reversely mapping the target response message into a final response message of the source system. The technical effect of ensuring continuous execution of the service request in the migration process is achieved.

The invention belongs to the technical field of measurement, and discloses a single-frame stripe depth reconstruction method based on indirect mapping and deep learning. The method comprises the following steps: firstly, constructing a data set for training a double convolutional neural network, and training a network model by using the data set; secondly, predicting high-frequency wrapped phase and coarse depth information by adopting the trained network model; thirdly, performing reverse mapping on the coarse depth information through a phase-depth model to obtain an uncorrected absolute phase, and performing reforming correction by combining a high-frequency wrapped phase to obtain a corrected absolute phase; and finally, the corrected absolute phase is subjected to forward mapping of a phase-depth model, final depth information is obtained, and high-precision three-dimensional reconstruction is realized. The method has the advantages that three-dimensional reconstruction of the measured scene can be completed only through a single-frame fringe pattern, the measurement efficiency is remarkably improved, the measurement precision is guaranteed, and meanwhile some important intermediate information can be obtained.

The invention belongs to the field of computer system compatibility and peripheral interfaces, and relates to a domestic desktop terminal peripheral compatibility method and system based on protocol reverse mapping and a computer readable storage medium. According to the method, after peripheral access, three-layer analysis of physical layer signal reversal, data link layer structure reversal and application layer semantic reversal is carried out in sequence, and an equipment interface type, a frame structure and instruction semantics are dynamically identified. And based on the reverse result, dividing the equipment into a standard protocol or a private protocol through a protocol type decision, respectively routing the protocol to a drive straight-through module or an instruction translation executor for processing, and finally providing compatible services for a domestic service system through a unified virtual equipment interface. According to the method, the adaptation problem of a document-free private protocol is solved, plug and play of a Wintel peripheral by a non-x86 architecture domestic terminal is achieved, the protocol compatibility coverage rate is remarkably increased, meanwhile, the CPU occupancy rate is reduced, and the harsh requirement for millisecond response of specific industries such as electric power is met.

A computing device comprises a processor, a table walker, and a memory storing a segmented reverse map table in multiple non-contiguous portions of the memory. The table walker is configured to translate a virtual memory address specified by a memory access request to a physical memory address associated with the virtual memory address; and provide a requester associated with the memory access request with access to the associated physical memory address in response to an indication at the reverse map table that the requester is authorized to access the associated physical memory address.

The present invention discloses a real-time analysis method and system for lower limb motion data applied to rehabilitation guidance. The method calls a personalized model based on user information, converts real-time motion data into historical motion feature vectors, inputs them into a real-time prediction module, and generates multiple frames of future motion feature vectors through encoding and decoding. The deviation value of each future motion feature vector from a preset threshold range is calculated to obtain a motion deviation value, and the motion risk ratio of each future motion feature vector is calculated. The motion risk probability is obtained through weighted fusion, and the device parameter adjustment value and posture adjustment information are generated through reverse mapping. Finally, the training stage is advanced based on the risk probability assessment, and correction information is synchronously displayed and the device operating parameters are adjusted. This method constructs a dynamic model through pre-training and combines it with a lightweight LSTM to achieve real-time prediction of future motion features, thereby improving the predictability and personalized adaptation capabilities of rehabilitation guidance.

The application discloses an image righting method and device of an endoscope, a storage medium and a terminal, and comprises the following steps: acquiring an initial image and a three-axis angular velocity signal of the endoscope; obtaining a posture quaternion by using a quaternion differential equation according to the three-axis angular velocity signal for attitude solution; updating the posture quaternion in real time; interpolating the posture quaternion of adjacent time points to obtain a line-level quaternion and generate a line-level rotation matrix within one frame exposure time; and righting and compensating the initial image to obtain a righted image by using reverse mapping and bilinear interpolation based on the line-level rotation matrix. The application avoids the problems of gimbal lock and functional failure caused by the gimbal lock in Euler angle attitude solution, eliminates the jelly effect by using an interpolation algorithm for line-by-line compensation, avoids the problems of large delay, insufficient real-time performance and difficulty in integration in a limited space caused by traditional software algorithms or external micro control unit post-processing, and better meets the imaging requirements of medical endoscope clinical operation.

The application belongs to the field of computer system compatibility and peripheral interface, and relates to a domestic desktop terminal peripheral compatibility method and system based on protocol reverse mapping and a computer readable storage medium. After the peripheral is accessed, the method sequentially performs three-layer analysis of physical layer signal reverse, data link layer structure reverse and application layer semantic reverse, dynamically identifies device interface type, frame structure and instruction semantics. Based on the reverse result, the device is divided into standard protocol or private protocol through protocol type decision, and is routed to a drive pass-through module or an instruction translation executor for processing, respectively, to finally provide compatibility services for domestic business systems through a unified virtual device interface. The application solves the adaptation problem of undocumented private protocols, realizes plug-and-play of Wintel peripherals for non-x86 architecture domestic terminals, significantly improves protocol compatibility coverage, reduces CPU occupancy, and meets the stringent requirements of millisecond-level response in the power and other specific industries.

The present invention proposes a three-dimensional model optimization method for rapid local Delaunay triangulation, including: Step S1, import a model file and reconstruct the input data file based on the triangular patch data structure of half edges; Step S2, extract all planar regions in the model and filter out non-compliant planar regions as optimization candidates; Step S3, perform a three-dimensional rotation on the three-dimensional plane to be optimized so that it is mapped into a two-dimensional space; Step S4, optimize the planar in the two-dimensional space using the Delaunay algorithm; Step S5, reverse-map the optimized two-dimensional plane into the original three-dimensional plane; Step S6, delete the original plane in the model and insert the optimized three-dimensional plane into the three-dimensional model, where the insertion process ensures the topological closure of the entire three-dimensional model; Step S7, output the result of the optimized three-dimensional model.

The embodiment of the invention provides an image processing method and device, equipment, a storage medium and a program product. The method comprises the following steps: acquiring a reverse mapping tensor; in the on-chip storage space of the calculation unit, a buffer area used for image distortion transformation is distributed; writing forward mapping information of a group of pixels corresponding to a first sliding window into the buffer area based on the first sliding window on the reverse mapping tensor; moving one first unit block in response to the first sliding window, writing forward mapping information in a first area in the buffer area into an off-chip storage space of the computing unit, and moving forward mapping information in a second area in the buffer area to the first area; and determining a forward mapping tensor for the reverse mapping tensor based on the forward mapping information written in the off-chip storage space to perform image warping transformation on the input image. In this way, the image distortion transformation efficiency can be improved.

The invention belongs to the field of integrated circuits, and particularly relates to an efficient and accurate mapping relation reverse construction method for an FPGA configuration file. The efficient and accurate mapping relation reverse construction method for the FPGA configuration file comprises the steps that S1, device structure information of a target FPGA is obtained through an ISE tool, and a two-dimensional plane matrix mapping library of the target FPGA is constructed; the two-dimensional plane matrix mapping library of the target FPGA comprises all on-chip resources, basic information of the on-chip resources and a mapping relation between the on-chip resources and configuration file coding information; s2, constructing a mapping relationship between smaller resources site in all on-chip resources and circuit functions and configuration parameters supported by the resource type; s3, aiming at a smaller resource site in each FPGA on-chip resource, generating configuration files of instantiated functional circuits with different configuration parameters in batches; s4, according to the coding rule of the configuration file, constructing a mapping relation between the address of the configuration file and the resource where the functional circuit is located; and S5, comparing the difference of the configuration files corresponding to different parameters of the same circuit function realized by the same resource in parallel, and forming a mapping library file between the circuit function and the configuration parameters thereof and the configuration file addresses. The problems that in the prior art, reverse engineering of the FPGA configuration bit mapping relation is low in accuracy, low in speed and small in application range are solved.

This disclosure provides a fall recovery method and apparatus for a tethered differential omnidirectional wheeled humanoid robot, relating to the field of control technology. The method includes: inputting the current angle, angular velocity, and current of the motors into a forward mapping network to obtain joint state estimates; inputting the joint state estimates, the current posture and angular velocity of the base, and the robot's current contact state into a reinforcement learning policy network for policy reasoning to obtain joint space action commands; inputting the joint space action commands into a reverse mapping network to obtain motor space commands; and controlling the motors to execute fall recovery actions based on the motor space commands. This disclosure absorbs the non-ideal characteristics of the tethered differential control interface layer through the forward and reverse mapping networks, enabling the policy to learn and think in joint space. After deployment, the learned mappings correctly drive the motors, improving the accuracy and smoothness of the robot's recovery actions.