235results about How to "Improve acceleration" patented technology

According to the self-distillation training method of the convolutional neural network, the performance of the convolutional neural network is significantly enhanced by reducing the size of the convolutional neural network rather than expanding the size of the network. When knowledge is distilled in the network, the network is firstly divided into several parts; knowledge in the deeper portion ofthe network is then squeezed into the shallow portion. Under the condition that the response time is not used as the cost, the performance of the convolutional neural network is greatly improved through self-distillation, and the average precision improvement of 2.65% is obtained; a precision increase of 0.61% in the dataset ResNeXt is used as a minimum value to a precision increase of 4.07% in the VGG19 is used as a maximum value. According to the method, a convolutional neural network with multiple outputs can be regarded as a plurality of convolutional neural networks, and the output of each shallow classifier is utilized according to different requirements by matching with intensified extraction of features of the shallow classifier by an attention layer, so that the precision of the shallow classifier is remarkably improved.

A hybrid input differential engine system comprising a planetary gear set. Preferably, an ICE is connected to the planet gear carrier, the output shaft is connected to the ring gear, and the sun gear is connected to a supercharger / expander and an electric or hydraulic motor / generator. As engine torque increases, the supercharger speeds up, increasing torque still further, enabling a small displacement engine to have very high torque at low RPM. In cruise conditions, the sun gear direction is reversed by the motor / generator, causing the supercharger to act as an expander for efficiently throttling the engine. The motor / generator modulates the speed / torque relationships between the engine and the supercharger / expander. A second motor / generator may be used on the output shaft. The electric machines and electric storage may be downsized because less electrical power is needed for the operation of the system.

A transmission is described, in particular an automatic transmission, which has several shift control elements (A, B, C, D, E) and several gearwheels (2, 3) which can be engaged via the shift control elements (A to E) to form a power flow, in which, to establish a transmission ratio in each case at least one of the shift control elements (A to E) is closed. The shift control elements (B, C, E), which are engaged during an up-shift, are formed as frictional shift control elements, and the shift control elements (A, D), which during up-shifts respectively only constitute a shift control element to be disengaged, are made as positive-locking gear elements.

A diesel fuel, diesel fuel additive concentrate and method for improving the performance of fuel injectors for a diesel engine are provided. The diesel fuel includes a major amount of middle distillate fuel having a sulfur content of 15 ppm by weight or less; and a reaction product of (a) a hydrocarbyl substituted dicarboxylic acid or anhydride, and (b) an amine compound or salt thereof of the formulawherein R is selected from hydrogen and a hydrocarbyl group containing from about 1 to about 15 carbon atoms, and R1 is selected from hydrogen and a hydrocarbyl group containing from about 1 to about 20 carbon atoms, wherein the reaction product contains less than one equivalent of an amino triazole group per molecule of reaction product, and wherein the reaction product is present in an amount sufficient to improve the performance of diesel direct and / or indirect fuel injectors.

The invention relates to a double-motor multi-mode hybrid power driving system and a control method thereof. The double-motor multi-mode hybrid power driving system comprises an engine, an ISG (integrated starter and generator) motor, a main motor, a vehicle-mounted energy source, a drive axle, wheels, a normally closed clutch (2) and a clutch (4), wherein the output end of the engine is directly connected with the input end of the ISG motor through the normally-closed clutch (2); an output shaft of the ISG motor is connected with a driving disc of the clutch (4); a driven disc of the clutch (4) is connected with an input shaft of the main driving motor; the output shaft of the main driving motor is connected with the wheels through the drive axle; and the vehicle-mounted energy source is connected with the ISG motor and the main driving motor. When the double-motor multi-mode hybrid power driving system works, no transmission case is needed for speed change, the high-power main motor is used for direct drive starting, the drive axle is used for directly driving the wheels, and therefore the energy loss is reduced; and in addition, the engine and the ISG motor are directly connected by the normally-closed clutch, and the engine and the ISG motor are integrated with the main motor into a whole, and therefore the double-motor multi-mode hybrid power driving system is more compact in structure and reduced in system cost.

A system and method are disclosed for controlling the positioning of a movable member of a valve device by an electric actuator. A model is used to generate estimates for the position and velocity of the movable valve member that result from an actuator control signal being applied to drive the electric actuator. The actuator control signal comprises a feedback control signal combined with a feedforward control signal. The feedback control signal is generated based upon a difference between the estimated and actual positions of the movable valve member, while the feedforward control signal is generated based upon the desired position, the estimated position, and the estimated velocity. The feedforward signal is adjusted to cause the estimated velocity to approximately follow a defined maximum deceleration velocity trajectory as the estimated position of the movable valve member moves to the desired position. An exemplary embodiment is presented where the principles of the invention are applied to the control of an electronic throttle valve.

The invention discloses an urban rail maglev train traction converter output device. The device comprises an inverter, an input switch cabinet and an output converting device, wherein the input switch cabinet is in circuit connection with an input end of the inverter; the output converting device is in circuit connection with an output end of the inverter; the output end of the output converting device is in circuit connection with a plurality of external motors; the inverter comprises a plurality of inverter modules; and the output converting device comprises an output transformer which is in circuit connection with the output end of the inverter and an output switch device which is in circuit connection with the output end of the output transformer. The urban rail maglev train traction converter output device is provided with a plurality of same inverter modules and reduces the capacity demand of the traction inverter, thereby reducing the weight and the volume of the inverter and being beneficial to the arrangement of equipment; the voltage resistant grade and the output current demand of an inverter power device are reduced, and the cost of a traction system is reduced; the voltage formed by combining two inverters is high, and the speed-up capability of a train high-speed section can be improved; and the redundancy and the reliability of the system are improved.

The invention discloses a molten steel vacuum refining top gun which comprises a center pipe, an inner pipe arranged outside the center pipe and used for inputting cooling water, an outer pipe arranged outside the inner pipe and used for outputting the cooling water, an intermediate pipe matched and connected with the inner pipe, and a nozzle matched and connected with the center pipe and the outer pipe, wherein at least one surface of the outer side of the nozzle is a tapered surface. According to the invention, the flowing speed of the cooling water of the nozzle is increased, and the cooling is enhanced so that the service life is prolonged. The nozzle is designed into a tapered shape, thus the flow acceleration is facilitated, slag is bonded in a blowing process, bonded slag automatically drops under the action of the gravity, the service life of the nozzle is prolonged, the labor efficiency is increased, and the labor intensity of an operator is lowered.

The invention discloses a full speed domain torque maximizing vector control current distribution method for an induction motor. In the method, voltage limitation, current limitation and physical characteristic limitation of the motor are comprehensively considered, and a given torque maximizing exciting current vector value and a given maximum torque current vector value are obtained by deriving in different speed domains. Due to the adoption of the method, the current utilization ratio and efficiency of an induction motor driving system can be greatly increased, and the full speed domain torque output capability of is enhanced.

The invention belongs to the field of rust surface treatment, in particular relates to a recyclable particle jet descaling device comprising an air compressor, a pressure dividing valve, a mixing chamber, a high pressure particle tank, a screw, a 3D particle amount detector, a speed regulating motor, a particle storage tank, a double-layer vibrating screen, a vertical screw conveyor, an X-direction rail, a Y-direction rail, a Z-direction rail, a nozzle assembly, a nozzle, a dust cover, a blower, a horizontal filter pore plate, an inclined filter pore plate, a dust removal cloth bag, and the like. The recyclable particle jet descaling device can realize rapid particle addition, real-time particle amount monitoring, uniform, stable and controllable gas-solid mixing, injection position three-dimensional movement and angle adjustment, multi-stage particle separation and recycling, effective dust recovery, significant improvement of part surface rust removal efficiency, and enhancing of part surface strength.

The invention discloses a CNN and LSTM network-based logistical sorting standard detection method, and belongs to the technical field of image processing. The method comprises the following steps of:firstly obtaining a continuous skeleton data frame sequence from somatosensory equipment when an express sorting action is executed by a worker, screening 20 pieces of joint data which can represent actions from the skeleton data, converting the joint data into world coordinates, and dividing the skeleton data into two parts; selecting certain key points from the skeleton data from a part of the skeleton data to form certain geometric features formed by points, lines and surfaces so as to describe the actions, and carrying out time-domain modeling on the actions through an LSTM network; projecting the other part of skeleton data to generate joint graphs, such as a front view, a left view and a top view, at three visual angles and motion trail graphs thereof, and taking the joint graphs andthe motion trail graphs are taken as inputs of a convolutional neural network to carry out feature extraction; and finally fusing the two parts of information to carry out action judgement. The method has good timeliness, robustness and correctness, is convenient and reliable to implement and is suitable for real-time action recognition systems.

A device (10) for feeding bulk material items (12) from a large mass into a randomly re-alignable, individually portioned disentangled position within the reach of a robot (18). A vision system comprising a camera (26) or a sensor records the number, position and alignment of the separated bulk material items (12). A processor (56) processes the sensor signals and generates control commands in actuators. The entire surface area of a one-piece oscillating conveyer surface (16) is supported by an oscillating conveyor arm (44). The surface lies on top of the arm and is horizontally displaceable in the x- or the x- and y-direction, extending together with the oscillating conveyor arm (44) and projecting in a freely supported manner from a storage zone (40) across a distribution zone (42) up to one front end of a selection zone (28) for the bulk material items (12). An oscillating motion (S) is generated in the z-direction with a continuously increasing amplitude (A) in the x-direction. The oscillating motion (S) alternates or is at least partially synchronous and co-ordinated with a forward feed or reverse transport of the bulk material items (12) in the x- or x- and y-direction.

The present invention provides an automatic driving-orientated vehicle longitudinal positioning system and method using variable grid-based image feature detection. According to a forward target obtained from data outputted from a high-precision navigation system and the distance of the target calculated by a road in high-precision navigation, an ORB (Oriented Brief) feature extraction algorithm based on a variable grid region which carries scale information is adopted to search the specific forward target in a vehicle-mounted binocular vision system and output the distance from the vision system to the forward target; and according to the distance and the installation position of the binocular system in a vehicle, a vehicle trajectory in the high-precision navigation can be corrected, and therefore, vehicle longitudinal positioning accuracy in the automatic driving can be improved.

The invention provides a webpage access acceleration method and system, computer equipment, a server and a medium, and belongs to the technical field of the Internet. The method comprises the following steps of: when a starting instruction for a target client is received, starting a target client, according to a target domain name configured on the target client, sending a webpage access request for accessing an address corresponding to the target domain name to a target server, sending the webpage access request to the proxy server by the target server through a target communication link, detecting the target website of the webpage access request by a proxy server, when the target website comprises the target domain name, obtaining and sending webpage data to the terminal, and displayingthe received webpage data by the terminal through the target client. Through communication among the terminal, the target server and the proxy server, the webpage data corresponding to the target website can be obtained, the cost is low, the information transmission speed of the target communication link is high, and the acceleration effect is improved.

A diesel fuel, diesel fuel additive concentrate and method for improving the performance of fuel injectors for a diesel engine are provided. The diesel fuel includes a major amount of middle distillate fuel; and a reaction product of a hydrocarbyl substituted dicarboxylic acid or anhydride and an amine compound or salt. The amine compound has the formulawherein R is selected from the group consisting of a hydrogen and a hydrocarbyl group containing from about 1 to about 15 carbon atoms, and R1 is selected from the group consisting of hydrogen and a hydrocarbyl group containing from about 1 to about 20 carbon atoms. The reaction product contains at least one amino triazole group and is present in the fuel in an amount sufficient to improve the performance of diesel fuel injectors.

A composite power supply. Including micro control unit, battery voltage detection module, super capacitor voltage detection module, DC conversion module, current detection module, signal amplification module, voltage regulator module, super capacitor and battery. The micro-control unit processes the status signal according to the preset parameters, sends control commands to the DC conversion module, adjusts the output frequency and duty cycle of the DC conversion module, and controls the output power of the super capacitor and the output power of the battery; its purpose is to improve the electric motor Bicycle launch, acceleration, climbing performance. Avoid high-current charging and discharging of the battery alone, reduce damage to the battery, and improve the service life of the battery. Provided is a method for controlling the output power of the composite power supply of the present invention, which controls the composite power supply to provide electric power for electric bicycles to achieve high-speed and long-term running.

In a hybrid car, or an electric motor-driven four-wheel-drive vehicle, in which an electric motor is driven directly by an output from a generator, vehicle running performance is enhanced. The output from the generator is controlled so as to maintain a generator torque in a range, in which the vehicle does not stall. Specifically, a driving apparatus for the hybrid car includes a second generator 2 driven by an engine 1 for driving one pair of either front or rear wheels 14, 15 , an electric motor 5 driven through directly receiving the output from the second generator 2 and driving the other pair of either the front or rear wheels 14, 15 , and a four-wheel-drive control unit 6 controlling driving of the second generator 2 and the electric motor 5 . When the engine 1 is in an overloaded condition, the four-wheel-drive control unit 6 limits the output of the second generator 2.

A foam generator is provided. According to a foam generation principle, after air enters a foam generator cavity body, for avoiding conditions of single-phase outflowing of a foaming agent and uneven foaming as much as possible, firstly, a spray nozzle structure arranged at the cavity body is utilized for transforming pressure energy of a foaming agent solution at an inlet into kinetic energy to a greater extent; the foaming agent solution is dispersed and atomized by an atomizer after being sprayed; compressed air and the atomized foaming agent solution are fully mixed and then flow to a foam outlet of the foam generator; a dense metal filter screen arranged at the foam outlet ensures that the foaming effect is further improved, and the problem that the air or the foaming agent flows out of the foam generator in a single phase manner does not exist.

The embodiment of the invention discloses an SSD (Solid State Drive) data migration method and device, wherein the method comprises the following steps of: establishing a caching queue and a pre-caching queue, and arranging objects corresponding to the access popularities of data blocks in corresponding queues; according to the cache hit ratio of cached data, adjusting the improvement spans of levels of the various objects, and then, updating the pre-caching queue; and, judging whether the current levels of the various objects in the updated pre-caching queue are up to the pre-set levels or not, and, writing the corresponding data blocks in a cache after the current levels of the objects are up to the pre-set levels. By means of the technical scheme provided by the invention, when hotspot data is changed, the good acceleration effect is provided; and the system response time and the access rate are increased. The embodiment of the invention further provides the corresponding implementation device; therefore, the method has relatively high practicability; and the device has corresponding advantages.

The invention discloses an integrated acceleration factor based electromechanical product reliability index evaluation method. The method is characterized by comprising the steps that a basic reason for a fault is searched for from many aspects with the focus on a main fault mode, and a sensitive stress set is determined; a test scheme is designed, and key factors of the acceleration life test scheme are determined; acceleration life test is carried out; test data is analyzed statistically; and fault data of full life periods of client products are collected continuously, and an integrated acceleration factor is corrected iteratively and improved gradually so that the factor satisfies the engineering practicality more. The problems that an electromechanical product is not sensitive to theacceleration stress, the product is single in the acceleration stress and low in the test efficiency, and error of the acceleration factor is large are solved, the feasibility of engineering practiceis taken into full consideration, and the method is more suitable for evaluation of the reliability index of the high-reliability long-life product.

The invention discloses a CNN acceleration calculation method and system based on a low-precision floating-point number data representation form, and relates to the field of CNN acceleration calculation. The acceleration calculation method comprises the following steps: a floating-point number function module receives an input activation value and a weight from a storage system according to a control signal, and distributes the input activation value and the weight to different processing units PE for convolution calculation to complete CNN acceleration calculation; wherein the convolution calculation comprises forward calculation of a convolution layer completed by performing dot product calculation on MaEb floating-point numbers quantized through a low-precision floating-point number representation form. By calculating and using a low-precision floating-point number representation form MaEb, the accuracy of the quantized CNN is ensured under the condition that retraining is not needed; Nm MaEb floating-point multipliers are realized through low-precision floating-point multiplication and a DSP (Digital Signal Processor), so that the acceleration performance of a customized circuit or a non-customized circuit is greatly improved under the condition of ensuring the accuracy, the customized circuit is ASIC (Application Specific Integrated Circuit) or SOC (System On Chip), and the non-customized circuit comprises an FPGA (Field Programmable Gate Array).

The invention provides a hardware acceleration calculation power distribution method and system for a cloud center, and the cloud center. The method comprises the steps of presetting at least two FPGAacceleration cards; recording hardware calculation power of each FPGA acceleration card; receiving an FPGA hardware resource application sent by at least one external client; for each external client, according to estimated calculation power in the FPGA hardware resource application and each piece of the hardware calculation power, distributing the hardware acceleration calculation power of the cloud center to the FPGA hardware resource application; determining at least one optimal FPGA acceleration card, wherein the sum of the hardware calculation power of the optimal FPGA acceleration cardsis equal to the hardware acceleration calculation power of the cloud center; and accelerating a service corresponding to the hardware resource application by utilizing each optimal FPGA accelerationcard. According to the scheme, the service acceleration capability of a server can be improved.

A self rotating deployed film solar batteries array is composed of film solar batteries, electrical cables, a center axle and connecting lines: many long strip shaped film solar batteries radiate to form a circle or a polygon-array, when linking two adjacent film solar batteries with connecting lines, each film solar battery links to a center axle with electrical cables; film batteries deployed and maintained in an array shape by self rotating centrifugal force in space; film batteries are coiling outside the surface of a center axle before their deployment. The unite weight of thin film battery is only 1 / 30 of traditional solar cell paddle. The film solar batteries array can be used in various satellite, space station and solar power rocket. It will greatly reduce the launching cost, enhance the power for spacecrafts and double the length of service life.

The invention relates to a driving hanger bracket device and a control method thereof. The device comprises two front wing plates, two tail wings and four axle wing plates, each front wing plate is arranged at the lower end of the front portion of a vehicle body through a front wing plate attack angle adjusting mechanism, the front wing plates are in symmetry about the middle axis of the vehicle body, each tail wing is arranged at the rear portion of the vehicle body through a tail wing attack angle adjusting mechanism, the tail wings are in symmetry about the middle axis of the vehicle body,and each axle wing plate is arranged on a lower swing arm of a wheel through an axle wing plate attack angle adjusting mechanism. Compared with the prior art, the acting force of a force generator ofa driving hanger bracket is not necessarily identical in magnitude and opposite in direction, the adjusting mode of the driving hanger bracket device is more flexible, and the driving hanger bracket device has the advantages of being simple in structure, greatly improving the safety of vehicles and the like.

A hydraulic control device of an automatic transmission is provided, which includes a hydraulic control circuit with a pressure reducing valve including a first friction element for achieving a first gear position and a second friction element for achieving a second gear position that produces a lower speed than the first gear position. The hydraulic control circuit further includes a switch valve formed with an input port which is inputted with the line pressure in a forward gear position range, and which is switched between a first state where the input port communicates with a first oil path communicating with the first friction element in a failure state of power distribution to the plurality of hydraulic control valves, and a second state where the input port communicates with the second friction element in the same power distribution failure state. The switch valve includes various control ports to control its function.

The present invention is a Low-Drag Fin and Foil System for Surfboards (10), particularly including cambered fin foils (40; 42). The invention (10) also discloses low-drag, directionally unstable fin positions wherein the lesser of negative angle of attack of a trailing fin (50), versus the higher or positive angle of attack of a forward fin (48), makes the board (12) highly maneuverable by creating a yawing moment that aids the rotation of the board (12) as it is turned. The system particularly utilizes fins (40) having foil (42) shapes in which either the cambered side (74) or the non-cambered side (76) is provided with a combination of a convex curvature (68) and a concave curvature (70) to result in an oscillating curvature (72) which has a positive effect on control and acceleration.

A system and method are disclosed for controlling the positioning of a movable member of a valve device by an electric actuator. A model is used to generate estimates for the position and velocity of the movable valve member that result from an actuator control signal being applied to drive the electric actuator. The actuator control signal comprises a feedback control signal combined with a feedforward control signal. The feedback control signal is generated based upon a difference between the estimated and actual positions of the movable valve member, while the feedforward control signal is generated based upon the desired position, the estimated position, and the estimated velocity. The feedforward signal is adjusted to cause the estimated velocity to approximately follow a defined maximum deceleration velocity trajectory as the estimated position of the movable valve member moves to the desired position. An exemplary embodiment is presented where the principles of the invention are applied to the control of an electronic throttle valve.

The invention discloses a CDVS extraction process accelerating method based on a GPGPU platform. For the two stages, feature detection and local descriptor computation, of CDVS extraction process, arithmetic logic and parallel strategy of inter-pixel parallel substages and inter-feature-point parallel substages are realized by means of an OpenCL universal parallel programming frame, and acceleration is achieved by means of the parallel computation capacity of a GPU. The method comprises the steps of dividing GPU and CPU computing tasks, reconstructing a scale space pyramid storage model of an image, assigning a parallel strategy to each substage on the GPU, and weakening memory access bottleneck by means of local memory. By the adoption of the technical scheme, CDVS extraction process can be accelerated, and extraction performance is improved remarkably.

The invention discloses a heterogeneous accelerating genetic calculating method based on FPGA in a cloud platform. By the calculation method, genetic calculation is divided into a slow calculation part in a computer processor and a fast calculation part in the computer processor, hardware is adopted to accelerate according to the slow calculation part, software is adopted to accelerate according to the fast calculation part, the software accelerating part is achieved by the cloud platform, the hardware accelerating part is achieved by being downloaded to the FPGA provided by the cloud platform, and finally the acceleration is achieved by synthesizing results of heterogeneous acceleration. According to the method, heterogeneous calculation used by both a universal calculating platform and the FPGA is adopted, compared with a single processing unit, the FPGA can achieve the optimal combination of a hardware structure and calculation by changing the connecting relations of the hardware, and the speed is faster; compared with calculation which only uses the FPGA, the heterogeneous calculation enables processing units of two structures to exert respective strengths and has a better effect compared with the single processing unit, data quantity of single transmission is increased after reinforcing communication interfaces, transmission times are reduced, and the integral acceleration effect is improved.

The invention relates to an internal combustion engine intake gate and exhaust gate control method, belonging to the fields of internal combustion engine intake and exhaust systems and brake. In the control method, two drive modes of high compression ratio and low expansion ratio or low compression ratio and high expansion ratio, and a brake mode of air intake from the intake gate and air exhaustfrom the exhaust gate of a turbocharging internal combustion engine, and a brake mode of a naturally aspirated internal combustion engine are proposed, so as to achieve the control method of the intake gate and exhaust gate in these modes. By using the high compression ratio and low expansion ratio mode for low speed and high torque, tableland environment and acceleration condition, and low compression ratio and high expansion ratio mode for high speed and heavy load condition of the turbocharging internal combustion engine, in combination, the dynamic property, accelerating ability and high speed performance for low speed and high torque and tableland condition are simultaneously improved. By using the brake mode to convert the internal combustion engine with 720 degree per cycle to an air compressor with 360 degree per cycle, an auxiliary brake system of a vehicle is replaced, thus the brake of the internal combustion engine is realized, and the effect of adjustable brake by regulation of the starting parameters of air gates is realized.