236results about How to "Reduce hardware requirements" patented technology

The invention discloses a BP neutral network heavy machine tool thermal error modeling method optimized through a genetic algorithm. Through the establishment of the structure of a BP neutral network, global optimization is conducted on the initial weight and threshold of each layer of the BP neutral network through a training sample. After the error objective is set, global optimization is conducted on the initial weight and threshold of the BP neutral network structure through the genetic algorithm, and the optimal weight and threshold found by the genetic algorithm is substituted into the BP neutral network to be conducted with sample training. Based on the decline principle of the error gradient, quick search is conducted near the extreme point until the training is end and thermal error prediction model is obtained. Finally, robustness testing is conducted on the obtained thermal error prediction model. The global optimization is conducted on the initial weight and threshold of the BP neutral network structure through the utilization of the genetic algorithm, the self-characteristics of the BP neutral network is overcome, and the quickness, the accuracy and the robustness of convergence when the optimal weight and threshold is trained can be improved.

The invention discloses a block chain-based data processing method and system and related equipment, which are used for reducing block chain data required to be stored by block chain node equipment and reducing hardware requirements of the node equipment for joining a block chain network. The method provided by the embodiment of the invention comprises the following steps: sending a request message for requesting to join in a target block chain to a metadata chain in a block chain network; wherein the metadata of each block in the block chain is stored in the metadata chain, and the metadata comprises the id of the block chain, the block height, the block hash value and the block validity certificate; receiving metadata of the target block chain returned by the metadata chain; obtaining ablock head and block snapshot data of a target block corresponding to the current block height from a distributed data storage system; and verifying the correctness of the block head and the block snapshot data, and if the verification is passed, adding the block head and the block snapshot data into the target block chain.

The invention discloses a communication module, a communication method and a communication system for realizing eSIM remote configuration. The communication module is integrated in the terminal equipment and comprises a module unit and an eUICC unit. The module unit comprises a profile remote configuration application program and is used for communicating with a server to download a new profile file and controlling the eUICC unit to realize communication network access of the eUICC unit through an instruction; and the eUICC unit is an integrated circuit card in which user identity informationand card authentication information are stored, and is used for storing and installing a new profile file, and activating and accessing the new profile file according to an instruction of the module unit. By adopting the communication module provided by the invention, profile remote configuration can be directly carried out with a server, and hardware requirements on a terminal and development ofterminal software are reduced.

The invention provides a data transmission method, input equipment, a cloud server and a cloud game system. The method comprises the steps that an external input device is connected to a cloud server,responds to game control operation of a user to generate a game running instruction after a game is started, and sends the game running instruction to the cloud server; to-be-output data generated byexecuting the game operation instruction and sent by a cloud server is received, and the to-be-output data is decoded; and the decoded to-be-output data is released to the target terminal display device through the established data release channel with the target terminal display device, so that the target terminal display device outputs the to-be-output data after obtaining the to-be-output data. In this way, the hardware requirement for the terminal display device can be further reduced.

Disclosed is a method of implementing a redundant network. Ring network nodes in an initial state are set as master stations, ring ports are set in a half-blocked state, a master station election notification message is broadcast to the ring ports within a preset time interval, the nodes receive the master station election notification message, a node quality comparison vector is compared with node quality comparison vectors of the nodes to elect a master and standby stations, one of the ring ports is in the half-blocked state, the other ring ports are in a forwarding state, the master station broadcasts the master station election notification message to the ring ports, the standby master station stops sending the master station election notification message, the ring ports are in the forwarding state, and the message sent by the master station is received.

The invention discloses a nailing robot system and a nailing control method thereof. The nailing robot system comprises a mechanical arm, wherein a force sensor, a registration device and a surgical instrument are mounted at the front end of the mechanical arm; the force sensor is mounted between the mechanical arm and the surgical instrument and used for detecting the change of force in a drilling and nailing process; the registration device is installed in a mechanical interface position of the front end of the mechanical arm and used for registration between the mechanical arm and an imageacquisition device; an image distortion correcting device is connected with the image acquisition device and used for distortion correction of the image acquisition deice; the image acquisition deviceis used for acquiring surgical images; a remote workstation is connected with the mechanical arm and the image acquisition device through data lines, and is used for processing the surgical images acquired by the image acquisition device, planning a nailing path and controlling the mechanical arm to move. The system has a low hardware equipment requirement, and can reduce the surgical cost, makeadjustment according to image information fed back in a surgery in real time and improve the nailing accuracy.

The invention discloses a neural network system and method suitable for portable power inspection. The system comprises an image acquisition module, a convolutional neural network operation module anda peripheral function unit. The image acquisition module comprises a thermal infrared camera, a visible light camera and an anti-shake holder. The convolutional neural network operation module is mainly used for the operation of a convolutional neural network for processing image information. The convolutional neural network operation module is mainly composed of a CPU and hardware of a GPU, an NPU, a TPU and an FPGA or a system-on-chip composed of the hardware. The invention discloses a neural network system and method suitable for portable power inspection. A thermal infrared camera is adopted to determine the position of the power equipment; the visible light camera is used for studying and judging the state of the power equipment at the position, the influence of a complex backgroundon target recognition is effectively reduced, the infrared image information is less than visible light image information, the image processing speed is higher when the convolutional neural network isused for processing the image, and therefore the requirements for model precision and hardware are reduced.

The invention discloses a CNN convolution kernel hardware design method based on a cascade form. The method comprises the first step of defining a convolution fundamental matrix, and adopting hardware implementation to obtain a convolution fundamental module; the second step of presenting linear convolution using a fast convolution algorithm in a form of matrix multiplication, and conducting cascading on the convolution fundamental module according to the presenting mode to form a cascade convolution module; the third step of adding an addition matrix which is achieved by a summator into the cascade convolution module to achieve parallel processing, addition operation of an offset value and two-dimensional convolution sum so that the final fast convolution kernel can be obtained. According to the CNN convolution kernel hardware design method based on the cascade form, compared with a traditional two-dimensional convolution kernel, the hardware demand is lowered, and the system velocity is increased at the same time; secondly, the shortcoming that when fast convolution kernels are different in size, a hardware architecture needs to be redesigned is avoided.

A robotic system is integrated with one or more mobile computing devices. Physical configurations of individual components of the system in physical space, or agents, under control of a user or users, are duplicated in a representation in virtual space. Some degree of real-time parity is maintained between the physical and virtual spaces, so as to implement a virtual environment that mirrors the physical one. Events occurring within one environment can directly influence and bear consequence on the course of events occurring within the other environment. Elements of virtual space thereby become truly interdependent and unified on a peer footing with elements in physical space. In at least one embodiment, the system of the present invention is implemented as an application in entertainment, such as the manifestation of a video game in physical space.

The invention relates to a deep learning-based quick dynamic human body action extraction and identification method. At present, existing human body identification technology and application have the following deficiencies in multiple aspects: a human skeleton is a complicated structure body, and different human action habits correspond to different human action modes, so that human body identification is difficult generally. The method comprises the following steps of firstly describing whole information of size, color, edge, contour, shape and depth of a human body target; providing useful clues for action identification; extracting effective motion features from a video sequence; performing trace analysis by utilizing a motion trace of the target under the long-distance condition; and performing two-dimensional or three-dimensional modeling on four limbs and trunk of the target by utilizing information extracted from an image sequence under the short-distance condition. The method is used for extracting and identifying quick dynamic human body actions based on deep learning.

The invention discloses a novel least square support vector machine modeling method for a thermal error of a numerical control machine. The method comprises the following steps of: (1) selecting a kernel function and determining parameters; and (2) according to a principle of the least square support vector machine, establishing a machine thermal error model. A compensating system in the invention is simple in structure and reliable in application; and by means of the least square support vector machine modeling method, the model precision and the generalization capability are improved, and the defects of low precision, low generalization capability and the like of the conventional predicting method are overcome.

The invention discloses a fault-tolerant modulation method for co-busbar double-end cascade type five-level inverters. The method includes the steps that a stator winding of a motor is opened, the two ends of the stator winding are powered up by the two inverters which are completely the same, the two inverters share one set of direct-current busbar, and accordingly the co-busbar double-end cascade type five-level topology is formed; in order to ensure that the reference voltage vector track of each inverter is a circle, when short-circuit trouble happens to a switching device, the maximum output voltage of the inverter is lowered to be half of the voltage before, and only the inside hexagon available switching state is used for synthesizing the reference voltage vector; through rotation and per-unit treatment of a coordinate system, the five-level modulation method is effectively simplified, and the hardware requirement for system control is reduced; by the adoption of equilibrium factors, midpoint potential balance is achieved on the premise that fault-tolerant control is guaranteed. Based on the high-fault-tolerant characteristic of the double-end cascade type five-level inverters, when the switching device breaks down, fault-tolerant operation of the inverters is achieved by reselecting the voltage vector. An optimized 60-degree modulation method is provided according to the double-end cascade type system.

A satellite positioning system and method of maintaining the positioning and clock data of a satellite using as few as one or two monitoring stations is disclosed. Each satellite of the positioning system determines its own positioning and clock data relative to other satellites in a constellation. One or more satellite signal monitoring stations receive relative clock and positioning data from a plurality of satellites in the constellation. A constellation solution which is applicable to all of the satellites in the constellation is calculated by treating the constellation as a rigid body at a fixed point in time. The constellation solution is transmitted to a satellite of the constellation which disseminates the solution through a crosslink network within the constellation for use by each satellite to update its individual positioning and clock data as needed. Due to the stability of the position and clock measurements using the method of the present invention, satellite uploads need only occur for one satellite in a constellation once each two or more days rather than every day for every satellite as is conventionally done. The use of mobile monitoring stations and the capacity of the positioning system to assist in updating the position and clock data of satellites passing the constellation are disclosed. In particular embodiments of the invention, the satellite positioning system is configured as a global positioning system (GPS).

The invention discloses a method for measuring laser linewidth of an ultra-narrow linewidth laser. The method is creatively characterized in that by use of an automatic heterodyne detection system, a power spectrum of the ultra-narrow linewidth laser under an automatic heterodyne conduction is obtained; and a peak and a valley which are adjacent to each other in terms of positions are selected from an area beyond high-peak pulses on the power spectrum, a difference delta S between the contrast of the peak and the valley on the power spectrum are respectively calculated, and the laser linewidth delta f is calculated according to the delta S. The method for measuring the laser linewidth of the ultra-narrow linewidth laser, brought forward by the invention, has the following technical advantages that the method can realize measurement of ultra-narrow linewidth under the condition of quite short delay fibers, greatly reduces hardware demands, can effectively prevent the problem of noise brought by large-length delay fibers and is quite good in measurement precision.

An edge intensification device of Bel image is prepared as storing original Bell image data into buffer storage and carrying out brightness character value calculation and edge character analysis by directly fetching out original Bell image data from buffer storage, outputting edge character data and intensifying edge of original Bell image data according to analyzed edge character data.