3348 results about "DSPACE" patented technology

A computerized annotation method achieves real-time operation and better optimization properties while preserving the architectural advantages of the generative modeling approach. A novel clustering algorithm for objects is represented by discrete distributions, or bags of weighted vectors, thereby minimizing the total within cluster distance, a criterion used by the k-means algorithm. A new mixture modeling method, the hypothetical local mapping (HLM) method, is used to efficiently build a probability measure on the space of discrete distributions. Thus, in accord with the invention every image is characterized by a statistical distribution. The profiling model specifies a probability law for distributions directly.

Using a transformation based at least in part on a non-simple orthogonal or unitary matrix, data may be transmitted over a data bus in a manner that is resilient to one or more types of signal noise, that does not require a common reference at the transmission and acquisition points, and / or that has a pin-efficiency that is greater than 50% and may approach that of single-ended signaling. Such transformations may be implemented in hardware in an efficient manner. Hybrid transformers that apply such transformations to selected subsets of signals to be transmitted may be used to adapt to various signal set sizes and / or transmission environment properties including noise and physical space requirements of given transmission environments.

The invention provides a method for analyzing spatial variations in field and crop performance. The method allows analysis of the field and crop performance as a function of various treatment and / or environmental conditions. This method is an advance over the prior art in that it allows better defined, more accurate comparisons of adjacent test areas, eliminating the need for extrapolation of data and avoiding inaccuracies therefrom. Also, the more accurate comparisons allow better characterization of effects due to the environment, such as those caused by genotype by environment (GxE) interaction. The more accurately defined comparison units also lead to improved analysis when comparing to other data layers for non-spatial analysis. The invention also provides a system for carrying out this method. The system preferably includes a device for inputting spatially-referenced data, a memory, an output device, and a processor. The output data may be represented spatially or non-spatially. The information provided by the method and system of the invention will allow farmers to make more informed decisions regarding seed and farming inputs and practices, thereby maximizing agricultural productivity.

A closed-loop vision system is disclosed that utilizes a concept known as Dynamically Reconfigurable Vision (DRV), which is adaptive image sensing driven by a computer or human operator's response to changing scenery. The system reduces the amount of irrelevant video information sensed and thus achieves more effective bandwidth and computational resource utilization, as compared to traditional vision systems. One or more reconfigurable photodetector arrays sensitive to either visible, infrared or ultraviolet radiation are present in the DRV system. These photodetector arrays feature on-chip means for spatial and temporal data reduction implemented through multiple independently controllable, time-correlated, frequently overlapping windows on the photodetector array that may be programmed according to their size, location, resolution, integration time, and frame rate. All photodetector array windows are dynamically reconfigurable in real time on a frame-by-frame basis. Furthermore, a DRV system is constructed in a client-server architecture in which a vision processor client passes window request command messages to the reconfigurable photodetector array server, which in turn delivers the requested video back to the client processor. The ability to simultaneously reconfigure, integrate, process, and readout multiple photodetector array video windows is an important characteristic of the DRV system.

A methodology for process modeling and control and the software system implementation of this methodology, which includes a rigorous, nonlinear process simulation model, the generation of appropriate linear models derived from the rigorous model, and an adaptive, linear model predictive controller (MPC) that utilizes the derived linear models. A state space, multivariable, model predictive controller (MPC) is the preferred choice for the MPC since the nonlinear simulation model is analytically translated into a set of linear state equations and thus simplifies the translation of the linearized simulation equations to the modeling format required by the controller. Various other MPC modeling forms such as transfer functions, impulse response coefficients, and step response coefficients may also be used. The methodology is very general in that any model predictive controller using one of the above modeling forms can be used as the controller. The methodology also includes various modules that improve reliability and performance. For example, there is a data pretreatment module used to pre-process the plant measurements for gross error detection. A data reconciliation and parameter estimation module is then used to correct for instrumentation errors and to adjust model parameters based on current operating conditions. The full-order state space model can be reduced by the order reduction module to obtain fewer states for the controller model. Automated MPC tuning is also provided to improve control performance.

A method of vehicle control, in which a global target trajectory is tracked by successively calculating an optimum local trajectory in 4D output space to approach the global trajectory and observe vehicle performance limits and surrounding obstacle clearance. A receding horizon framework is proposed which successively updates the optimum local trajectory according to the current state of the vehicle. Processing overheads can be kept to a minimum by calculating performance limits offline, and optimisation is simplified by using a cost function approach.

This present application relates to the field of cloud storage security technology, and in particular, relates to a cloud storage data encryption method, apparatus and system. The method comprises: according to the amount of data X expected to be stored within the preset time, the proportion of local storage space R and the security level of data Z, calculating the size H of a random seed that should be generated; according to the amount Y of plaintext data every time, calculating the times u of random seed acquired; according to the times u, acquiring data from the generated random seed with the size of H for several times to generate a plaintext encryption bit identifier data string; by use of the data string, selecting more than one half of the plaintext data for encryption to form a ciphertext. This application also provides a cloud storage data encryption apparatus and system. This invention has reduced the amount of encrypted data to be stored without sacrifice in the degree of data security protection, thus greatly improves the cloud storage data encryption and decryption performance.

A random k-nearest neighbors (RKNN) approach may be used for regression / classification model wherein the input includes the k closest training examples in the feature space. The RKNN process may utilize video images as input in order to predict motor command for controlling navigation of a robot. In some implementations of robotic vision based navigation, the input space may be highly dimensional and highly redundant. When visual inputs are augmented with data of another modality that is characterized by fewer dimensions (e.g., audio), the visual data may overwhelm lower-dimension data. The RKNN process may partition available data into subsets comprising a given number of samples from the lower-dimension data. Outputs associated with individual subsets may be combined (e.g., averaged). Selection of number of neighbors, subset size and / or number of subsets may be used to trade-off between speed and accuracy of the prediction.

The invention provides a real-time target detection method based on a region convolutional neural network. The real-time target detection method mainly comprises an input image, a target detection system, alternative optimization learning and sharing, and classifier classification and detection. The real-time target detection method comprises the steps of: regarding an image of any size as input, inputting a plurality of regions of interest (RoIs) while inputting the image, proposing a detection region by means of a region proposal network (RPN), utilizing the proposed detection region by an R-CNN detector, sharing all spatial positions by means of complete connection layers, learning shared characteristics by adopting alternative training optimization, and carrying out classification detection by using the classifier. According to the real-time target detection method, the RPNs are used for generating region proposals, and the network parameters are reduced by using shared weights, thus the region proposing step costs almost nothing; and the region proposal network (RPN) and the region convolutional neural network (R-CNN) share two network between a convolutional layer, thereby the cost is significantly reduced, the detection speed is fast, and the efficiency is high.

A system and method are provided for enhanced navigation along execution time and code space in a debugger to assist a user in remediating errors, streamlining, or reverse engineering a computer program and the source code thereof. Snapshots of system states are recorded, a causality tree of commands is constructed through execution of the program to be debugged, and an intelligent display of system states captured during runtime and indexed or cross-referenced by time are displayed to the user in an intelligent manner to aid the user with certain debugging tasks. Additionally, further features in assisting the user to locate a root cause of an error or unexpected value and remediate that cause are also provided.

Data center interconnections, which encompass WSCs as well as traditional data centers, have become both a bottleneck and a cost / power issue for cloud computing providers, cloud service providers and the users of the cloud generally. Fiber optic technologies already play critical roles in data center operations and will increasingly in the future. The goal is to move data as fast as possible with the lowest latency with the lowest cost and the smallest space consumption on the server blade and throughout the network. Accordingly, it would be beneficial for new fiber optic interconnection architectures to address the traditional hierarchal time-division multiplexed (TDM) routing and interconnection and provide reduced latency, increased flexibility, lower cost, lower power consumption, and provide interconnections exploiting N×M×D Gbps photonic interconnects wherein N channels are provided each carrying M wavelength division signals at D Gbps.

The invention provides a binocular camera-based panoramic image splicing method. According to the method, a binocular camera is arranged at a certain point of view in the space, the binocular camera completes photographing for once and obtains two fisheye images; a traditional algorithm is improved according to the defect of insufficient distortion correction capacity of a latitude-longitude correction method in a horizontal direction; corrected images are projected into the same coordinate system through using a spherical surface orthographic projection method, so that the fast correction of the fisheye images can be realized; feature points in an overlapping area of the two projected images are extracted based on an SIFT feature point detection method; the search strategy of a K-D tree is adopted to search Euclidean nearest neighbor distances of the feature points, so that feature point matching can be performed; an RANSAC (random sample consensus) algorithm is used to perform de-noising on the feature points and eliminate mismatching points, so that image splicing can be completed; and a linear fusion method is adopted to fuse spliced images, and therefore, color and scene change bluntness in an image transition area can be avoided.

This invention provides a system and method to propagate uncertainty information in a deep learning pipeline. It allows for the propagation of uncertainty information from one deep learning model to the next by fusing model uncertainty with the original imagery dataset. This approach results in a deep learning architecture where the output of the system contains not only the prediction, but also the model uncertainty information associated with that prediction. The embodiments herein improve upon existing deep learning-based models (CADe models) by providing the model with uncertainty / confidence information associated with (e.g. CADe) decisions. This uncertainty information can be employed in various ways, including (a) transmitting uncertainty from a first stage (or subsystem) of the machine learning system into a next (second) stage (or the next subsystem), and (b) providing uncertainty information to the end user in a manner that characterizes the uncertainty of the overall machine learning model.

In accordance with the objects of the present invention, a method of complex product design is provided. The method defines a product design space a base design analysis model is then automatically generated within the product design space. A topology optimization of the base design analysis model is performed including solving the topology optimization for multiple mass fractions to identify critical structure areas. The results of the topology optimizations are utilized to identify a plurality of load paths. An optimization of said plurality of load paths is performed to find the optimized load paths. Critical structure areas along the plurality of optimized load paths are then parameterized. An optimization is performed utilizing a multi-disciplinary optimization to optimize geometry, material and thickness along the optimized load paths. A manufacturing analysis of the shape optimized structure is performed and the shape optimized structure is modified to comply with manufacture requirements. An optimization is then performed using multi-disciplinary optimization comprising a gauge and grade optimization.

A flow distribution technique associates probabilistic data structure with each processor of a multi-processor system that is used to determine whether a received packet is included in a flow assigned to the associated processor. A set of tuples of a received packet is processed by multiple hash functions to generate a set of values utilized as indices to search for hits of the probabilistic data structures. Only those flows having hashed indices that hit more than one probabilistic data structure are saved in a collision flow table. The probabilistic data structure require much less memory space than a flow table and can be designed so that the frequency of collisions is very low thereby requiring a small amount of memory for the collision flow table.

An automated tuning method of a large-scale multivariable model predictive controller for multiple array papermaking machine cross-directional (CD) processes can significantly improve the performance of the controller over traditional controllers. Paper machine CD processes are large-scale spatially-distributed dynamical systems. Due to these systems' (almost) spatially invariant nature, the closed-loop transfer functions are approximated by transfer matrices with rectangular circulant matrix blocks, whose input and output singular vectors are the Fourier components of dimension equivalent to either number of actuators or number of measurements. This approximation enables the model predictive controller for these systems to be tuned by a numerical search over optimization weights in order to shape the closed-loop transfer functions in the two-dimensional frequency domain for performance and robustness. A novel scaling method is used for scaling the inputs and outputs of the multivariable system in the spatial frequency domain.

The invention discloses data exchange and sharing technology among heterogeneous systems. Data and services in different systems can be accessed in a uniform mode. The data exchange and sharing technology can be implemented by the following steps of: 1, establishing an import data source; 2, establishing a data format template; 3, defining a transformation rule; 4, archiving the data format template; 5, importing the data; 6, exporting the data; 7, transforming formats; 8, sorting the data out; 9, recovering the data; 10, encrypting / decrypting and compressing / uncompressing the data; and 11, managing messages. The invention provides a data interface mechanism, enterprise-level data interaction in the systems can be realized by the mechanism, data exchange and sharing of each management system can be realized through an enterprise data platform, and a large number of memory spaces are not required when the data are processed, so that investment in memory equipment can be reduced; and data processing and transmission are performed on line without manual intervention, so that maintenance cost can be reduced.

An internet marketing analytics system to quantify visitor website activity, the system including a database having a plurality of parameters having any portion of the following; site visits, total revenue, bounce rate, conversion rate, orders, average order value, value per visit, pages per visit, new visits, units, cart additions, cart removals, internal search, click through rate, revenue per visit, open rate, email list, impressions, visit duration, percent new visitors, percent return visitors, post volume, net promoter score, social referring traffic, total cost, search term, exit rate, page views, and product type, plus non website related visitor internet activity. The system producing a multi-variate visual spatial display of the database including at least three parameters being simultaneously displayed and modifying the display to selectively emphasize a parameter to be displayed as an X, Y, or Z axis for clarity, for modifying a component of the database to optimize website efficiency.

The invention discloses a method for data fragmentation of a distributed storage system. The method comprises the steps that a physical node state is monitored through a cluster controller and the correctness of data transmission is guaranteed; physical spaces of storage nodes are managed in a unified mode through the consistent Hash algorithm, so that a logic storage space is formed; the reliability of the system is guaranteed in multi-transcript fragment mode. According to the method, the storage capacity is lager and flexible and the storage capacity virtualized by a host computer depends on resources of a native machine. Due to the limitation on hardware design of the host computer, the storage capacity cannot be infinitely increased; the storage capacity cannot be limited to the storage capacity of the host computer after the distributed system is adopted, storage resources can be increased or reduced at will according to the requirement of a client, and the storage capacity and the flexibility of the system are improved to the greatest extent.

The invention discloses a radar signal parallel processing method and system based on a heterogeneous multinucleated system. The method comprises the following steps: S1, initializing equipment-end parameters, opening up a task execution thread, and dividing thread grids of data processing and the dimensions of a thread block; S2, opening up a video memory space and a memory space which are needed for the signal processing; S3, acquiring data transmission and calculating time needed by signal processing of a single time, and scheduling a genetic task scheduling algorithm to obtain a task scheduling mode; and S4, acquiring laser sampling data for storing in the memory space by means of segments according to a time sequence, sending the data to a CPU and a GPU according to the task scheduling mode, mapping sampling points to each thread and performing concurrent execution according to the thread grids and the dimensions of the thread block, and scheduling a filter coefficient to perform orthogonal phase demodulation, pulse compression, moving object display, moving object detection, pulse accumulation and constant false alarm detection on the sampling data. By applying the method and system provided by the invention, the speed of a general processor in executing the signal processing is improved, and the requirement for real-time performance of radar signal processing is met.

The present invention, known as runspace, relates to the field of computing system management, data processing and data communications, and specifically to synergistic methods and systems which provide resource-efficient computation, especially for decomposable many-component tasks executable on multiple processing elements, by using a metric space representation of code and data locality to direct allocation and migration of code and data, by performing analysis to mark code areas that provide opportunities for runtime improvement, and by providing a low-power, local, secure memory management system suitable for distributed invocation of compact sections of code accessing local memory. Runspace provides mechanisms supporting hierarchical allocation, optimization, monitoring and control, and supporting resilient, energy efficient large-scale computing.

The invention discloses an estimation method of a co-prime array DOA (Direction Of Arrival) angle based on sparse reconstruction, and mainly solves the problems that a prior art is higher in operand, less in identification information source amount and large in passive location evaluated error, and needs more priori knowledge. The method comprises the realizing steps of forming a co-prime array by an antenna receiver; obtaining observation data to spatial signal sampling; receiving data vectors by a virtual array element obtained by observation data; dividing spatial grids to form over-complete bases; receiving a spare relationship between the data vectors and the over-complete bases by the virtual array element to build a spare restraint equation; resolving the spare restraint equation by adopting a convex optimization method to obtain sparest resolution; drawing a magnitude spectrogram by a relative relationship between the sparest resolution and the spatial angle to obtain a DOA angle value. According to the method provided by the invention, the passive direction-finding precision and operating speed can be improved under a condition of low priori knowledge, the number of the recognized information source can be improved, and the estimation precision of a signal direction angle can be improved in a low signal to noise ratio, therefore the estimation method can be used for target reconnaissance and passive location.

The embodiment provides an indoor positioning or navigation method and device, and a map database generation method, and relates to the communication technical field. The method includes the steps: acquiring real-time image information of a position of a user; identifying a preset identification contained in the real-time image information; based on the preset identification, searching in a map database, to obtain position information corresponding to the preset identification, wherein in the map database, the preset identification and the position information are in one-to-one correspondence; and according to the position information, generating user current position information. According to the indoor positioning or navigation method and device, and the map database generation method provided by the embodiment of the invention, positioning is accurate, the operations are simple and easy, maintenance is almost not required, and the method can be used even if other environments are updated as long as the preset identification exists; at the same time, the work of data acquisition during construction of the map database is greatly reduced, and occupied storage space after the map database is constructed and the calculation amount during matching of the map database are also greatly reduced.

The invention discloses a hybrid power assembly performance matching test bench which comprises an engine, a clutch, a speed changer, an engine controller ECU, a power generating / motor-driven integrated motor ISG, a motor controller IPU, a power battery pack, a battery management system ECU, a vehicle controller HCU, a DC converter DC / DC, a shift mechanism, a clutch pedal, a brake pedal, an accelerator pedal, two dynamometers, a bench control system and a rapid control prototype dSPACE. The test bench has the advantages that the entire vehicle state is simulated on an engine test bench, the main parts of the hybrid power system, the power battery pack, the BCU, the HCU, the IPU and the DC / DC are arranged by simulating the entire vehicle state, the sturcture is compact, the testing programis simple, the performance matching of the hybrid power assembly can be easily realized on the test bench, and NEDC and other condition tests can be conducted by simulating the entire vehicle.

The invention discloses a method for achieving industrial robot off-line programming based on three-dimensional modeling software. The principle of the method is that a three-dimensional modeling software application programming interface (API) which is suitable for COM or object linking and embedding (OLE) and the like is used in the three-dimensional modeling software environment, model and space matrix data in the three-dimensional modeling software are obtained, a robot virtual body model and a mathematical model are built, and a virtual robot motion control system is realized; discretization from a curve to a point is conducted through obtained space line data, and therefore a motion trail of a robot is generated, actual robot operation is simulated; and an industrial robot controller operational procedure generating module is built, an actual controller executable file is generated and lead out so as to be operated in an actual robot controller, and industrial robot off-line programming can be achieved. The method can achieve off-line programming operation rapidly and generate optimal gesture of corresponding points automatically, time for generating training data is short, and a space curve trail can be achieved conveniently and rapidly, wherein a basis of the space curve trail is difficult to find in reality to conduct teaching.

A gesture recognition system is shown using a 77 GHz FMCW radar system. The signature of a gesturing hand is measured to construct an energy distribution in velocity space over time. A gesturing hand is fundamentally a dynamical system with unobservable “state” (i.e. the type of the gesture) which determines the sequence of associated observable velocity-energy distributions, therefore a Hidden Markov Model is used to for gesture recognition. A method for reducing the length of the feature vectors by a factor of 12 is also shown, by re-parameterizing the feature vectors in terms of a sum of Gaussians without decreasing the recognition performance.

The invention relates to a heterogeneous cloud wireless access network resource allocation method based on deep reinforcement learning, and belongs to the technical field of mobile communication. Themethod comprises the following steps: 1) taking queue stability as a constraint, combining congestion control, user association, subcarrier allocation and power allocation, and establishing a random optimization model for maximizing the total throughput of the network; 2) considering the complexity of the scheduling problem, the state space and the action space of the system are high-dimensional,and the DRL algorithm uses a neural network as a nonlinear approximation function to efficiently solve the problem of dimensionality disasters; and 3) aiming at the complexity and the dynamic variability of the wireless network environment, introducing a transfer learning algorithm, and utilizing the small sample learning characteristics of transfer learning to enable the DRL algorithm to obtain an optimal resource allocation strategy under the condition of a small number of samples. According to the method, the total throughput of the whole network can be maximized, and meanwhile, the requirement of service queue stability is met. And the method has a very high application value in a mobile communication system.

A probabilistic digital signal processor using data from multiple instruments is described. Initial probability distribution functions are input to a dynamic state-space model, which operates on state and / or model probability distribution functions to generate a prior probability distribution function, which is input to a probabilistic updater. The probabilistic updater integrates sensor data from multiple instruments with the prior to generate a posterior probability distribution function passed (1) to a probabilistic sampler, which estimates one or more parameters using the posterior, which is output or re-sampled in an iterative algorithm or (2) iteratively to the dynamic state-space model. For example, the probabilistic processor operates on fused data using a physical model, where the data originates from a mechanical system or a medical meter or instrument, such as an electrocardiogram or pulse oximeter to generate new parameter information and / or enhanced parameter information.

The invention relates to a dexterous hand teleoperation control method based on Kinect human hand motion capturing. The hardware for realizing the method comprises a Kinect sensor, a computer and a remote robot dexterous hand system. A mixing movement outline model is used for tracking the outline of the hand, hand posture tracking is realized through a particle filter algorithm, the far-end situation is observed through feedback image information transmitted by a robot in real time, and therefore a specific operation task is implemented. Through the movement outline model combining depth information and YCbCr skin color detection space technology, the problem that a real solution can not be obtained easily when an image contains high noise or a target is provided with a weak boundary is solved. A particle filter posture detection method based on template matching is adopted, and the human hand three-dimensional posture with high-dimensional characteristics can be detected and estimated. Through a master-slave operation combining a manual control mode and a program mode, convenience and rapidness are realized, and the guiding performance and usability of human-machine interaction are represented.