293results about How to "Eliminate accumulated errors" patented technology

The invention discloses a hand movement tracking system and a hand movement tracking method. The invention comprises an attitude and heading reference system based on an accelerometer, a gyroscope and a magnetic sensor, and a hand movement tracking method based on the attitude and heading reference system. The hand movement tracking method comprises the following steps: firstly, obtaining a triaxial acceleration measured by the accelerometer, a triaxial angular velocity measured by the gyroscope and a triaxial magnetic-field component measured by the magnetic sensor, performing error compensation on the magnetic sensor by adopting a least square method to establish an error model after an upper computer receives sensor data, eliminating high-frequency noise of the triaxial acceleration by virtue of a window low-pass filter, and establishing an error model for the gyroscope so as to perform error compensation on random drift of the gyroscope; secondly, effectively integrating the gyroscope, the accelerometer and the magnetic sensor by virtue of an improved adaptive complementary filtering algorithm to obtain an attitude angle and a path angle; and finally, performing gravity compensation and discrete digital integration on acceleration signals to obtain a velocity and a track of a hand movement. The tracking system and the tracking method disclosed by the invention can be applied to a man-machine interactive system, is convenient to operate, and is strong in experience feeling.

The invention relates to a sun orientation automatic tracking method and a device used for photovoltaic power generation. The detection precision of the existing method and the device is lower. The method is a 24-hour 2D sun orientation automatic tracking method which uses a light intensity sensor to combine photoelectric tracking and sun motion trail tracking. The light intensity sensor detects climate and respectively adopts three different working modes to track the sun orientation in sunny days, cloudy days and rainy days, which solves the influences of climate changes and environment interference on the tracking stability to a certain extent, and simultaneously eliminates the accumulated errors through the closed loop control function of the photoelectric tracking mode, thus being capable of stably implementing sun orientation tracking in 24 hours. The device for realizing the method includes a wave-pass drum, the light intensity sensor, protective glass, a four-quadrant photoelectric detector, a signal processing circuit, a computer and an external clock chip. The invention improves the detection precision and solves the problem of unstable detection signal caused as the sunorientation detector is easy to be interfered by environmental light.

The invention provides a longitudinal positioning system and method based on a city road marking map in automatic driving. Through current road marking information collected by an intelligent driving vehicle, and based on locating point coordinates currently given by a GNSS which have an error, longitudinal positioning of the vehicle based on the road marking map is further performed in the vicinity of the current locating point coordinates, thereby correcting and repositioning the current accurate position of the vehicle, and on a structured road in a complex environment of a city, the absolute position of static road markings can be utilized to perform auxiliary correction of GNSS positioning and eliminate an accumulated error of an INS system.

The invention provides a text recognition method and device. The method comprises: determining convolutional neural network (CNN) features of a text image and obtaining an eigenvector sequence of a preset dimension; and using a recurrent neural network (RNN) to decode the eigenvector sequence of the preset dimension, and obtaining texts in the text image to be recognized. According to the technical scheme provided by embodiments of the invention, the method combines the CNN and RNN to achieve integral recognition of the text image to be recognized, so that the method avoids the problem that texts need to be separated character by character and then recognized character by character in the prior art which leads to accumulated errors, and the text recognition rate can be improved.

The invention provides an electric power inspection robot positioning method based on multi-sensor fusion, and the method comprises the steps: firstly carrying out the preprocessing of data collectedby a camera, an IMU and a speedometer and the calibration of a system, and completing the initialization of a robot system; secondly, extracting a key frame, and performing back-end optimization on the position, the speed and the angle of the robot and the bias of a gyroscope in the IMU by utilizing the real-time visual pose of the key frame to obtain the real-time pose of the robot; then, constructing a key frame database, and calculating the similarity between the current frame image and all key frames in the key frame database; and finally, performing closed-loop optimization on the key frames in the key frame database forming the closed loop, and outputting the pose after closed-loop optimization to complete positioning of the robot. According to the back-end optimization method provided by the invention, the positioning precision is improved; and closed-loop optimization is added in the visual positioning process, so that accumulated errors existing in the positioning process areeffectively eliminated, and the accuracy under the condition of long-time operation is guaranteed.

An electric network transformer monitoring device comprises a data acquisition unit, a communication unit, a data processing unit, a control unit and a power supply unit, wherein the data acquisition unit consists of a filter circuit, a sampling circuit and a GPS timing circuit. The monitoring method provided by the invention applies unified time labels on the acquired transformer parameters through the GPS timing circuit of each electric network transformer monitoring device, transmits the monitored transformer parameters to an electric network dispatching center, judges and positions the internal faults of each transformer winding in the electric network, carries out the voltage stability analysis of the whole electric network according to the state of each transformer at the same time, reliably guarantees the safe and stable operation of local electric networks, and realizes the real-time monitoring of the wide-area electric network.

A positioning method for indoor mobile robots belongs to the technical field of mobile robot positioning. The invention provides a multisensor fusion asynchronous time of arrival (ATOA) indoor positioning method based on a single ultrasonic sensor, an encoder and an electron compass, and provides a double-layer Kalman filtering (DLKF) data fusion model. Based on the DLKF model, the ATOA method can effectively eliminate accumulated error in dead-reckoning to obtain high precision of positioning, greatly reduces the quantity of ultrasonic sensors for positioning, and effectively overcomes the disadvantage of blind areas in ultrasonic sensor positioning. Based on the DLKF model, the ATOA method can effectively eliminate accumulated error in dead-reckoning to obtain good precision of positioning, and greatly reduces the quantity of ultrasonic sensors for positioning.

A three-dimensional (3D) pointing device capable of accurately outputting a deviation including yaw, pitch and roll angles in a 3D reference frame and preferably in an absolute manner is provided. Said 3D pointing device comprises a six-axis motion sensor module including a rotation sensor and an accelerometer, and a processing and transmitting module. The six-axis motion sensor module generates a first signal set comprising angular velocities and a second signal set comprising axial accelerations associated with said movements and rotations of the 3D pointing device in the 3D reference frame. The processing and transmitting module utilizes a comparison method to compare the first signal set with the second signal set to obtain an updated state of the six-axis motion sensor module based on a current state and a measured state thereof in order to output the resulting deviation in the 3D reference frame and preferably in an absolute manner.

The invention requests the protection of an inertial navigation high-precision positioning method based on binocular, acceleration and gyroscope. The three-dimensional coordinates of the characteristic points of an inertial navigation reference position are determined based on binocular positioning and high-precision GPS information, and the three-dimensional coordinates of inertial navigation areupdated in real time by using the coordinate values of the characteristic points as reference points to eliminate the cumulative error caused by the integration of inertial navigation, and assist inhigh precision positioning of vehicles in an indoor parking lot based on inertial navigation and binocular positioning. A structure block diagram of the invention includes the following parts: I, thecomposition of a high precision inertial navigation system and the structure of a reference coordinate system; II, an acquisition method of the initial coordinates of inertial navigation; III, a selection method of binocular high precision positioning reference points; IV, the selection and design of a high precision inertial navigation coordinate system; V, a conversion method of a high precisioninertial navigation coordinate system; and VI, an elimination method of the cumulative error of inertial navigation.

A 3D pointing device utilizing an orientation sensor, capable of accurately transforming rotations and movements of the 3D pointing device into a movement pattern in the display plane of a display device is provided. The 3D pointing device includes the orientation sensor, a rotation sensor, and a computing processor. The orientation sensor generates an orientation output associated with the orientation of the 3D pointing device associated with three coordinate axes of a global reference frame associated with the Earth. The rotation sensor generates a rotation output associated with the rotation of the 3D pointing device associated with three coordinate axes of a spatial reference frame associated with the 3D pointing device itself The computing processor uses the orientation output and the rotation output to generate a transformed output associated with a fixed reference frame associated with the display device above. The transformed output represents a segment of the movement pattern.

The invention discloses a camera, imu and laser radar combined robot positioning method, comprising the steps: maintaining the more accurate depth of a surrounding environment through the camera and the laser radar, and constructing a speedometer through the continuous tracking of feature points. According to the camera, imu and laser radar combined robot positioning method, the advantages of thecamera, the imu and the laser radar can be combined; an accurate and dense point cloud map is established for the robot to perform path planning; the precision of the interframe odometer is improved through tight coupling of the camera and the laser radar; high robustness is achieved; through a point cloud map maintained by laser and a camera in real time, local optimization of the odometer is carried out in laser scanning; and a loop is judged through word information provided by a camera and imu data, and closed-loop optimization is achieved on a point cloud map provided by vision and laserjointly.

The invention discloses a photovoltaic array maximum power point tracking servo system comprising a clock circuit, a double-shaft mechanical tracking and locating system and a photovoltaic detecting device respectively connected with a singlechip control system. The double-shaft mechanical tracking and locating system comprises a first tracking component and a second tracking component connected with a battery panel bracket. A battery panel and the photovoltaic detecting device are provided on the battery panel bracket. The singlechip control system respectively controls and drives the first tracking component and the second tracking component to drive the battery panel and the photovoltaic detecting device to move in the vertical direction and the horizontal direction. The driving system adopted by the invention is a double-shaft active solar servo system which can realize the sun tracking under all weather conditions, has high precision, can always align the solar battery panel to the sun and maintain the maximum generating efficiency. Furthermore, the invention also discloses a control method of the system.

The invention relates to a method for measuring track direction and horizontal irregularity based on plot points, belonging to the fields of rail transit, and railway conservation and maintenance. The method is characterized by comprising the steps of measuring data by mounting an inertia sensor trolley firstly; and then setting a reference coordinate system; calculating the coordinate of a measured point at the position of point zero; then calculating the coordinate of a measured point at the position of point one; then calculating the coordinates of measured points at positions in a measurement interval from the position of point two, thus the curve plot points of the integral measured rail is realized; and calculating the irregularity degree value according to the obtained plot points.The method has the beneficial effects that the rail direction and horizontal irregularity are obtained through setting a two-dimensional coordinate for the measured rail and carrying out geometric calculation, the cumulative errors of the calculation method for the rail direction and the horizontal irregularity degree by the existing measuring method are eliminated, the measurement accuracy is improved, the method has very good adaptability, and can be widely applied to rail irregularity detection devices which are measured by utilizing inertial elements.

The invention relates to a synchronous phasor calculation method based on DFT recursion of FPGA hardware. An FPGA is configured through a central processing unit (CPU) and controls an analog-digital (AD) converter to complete a sampling process under the 1 pulse per second signal (1PPS) synchronization, and the cycle DFT calculation and the recursion DFT calculation are utilized for eliminating accumulated errors of the recursion DFT calculation; and a complex phasor compensation algorithm is completed by the CPU. The FPGA and the CPU are matched to complete an acquisition calculation compensation process of synchronous phasors, and both the high-speed parallel computing capability of the FPGA and the flexible floating-point operation function of the CPU are used. The recursion DFP operation which is long in consumed time is completed by the FPGA, and the CPU loads are small, so that the communication response instantaneity of the CPU is guaranteed, and further, the communication reliability of a power management unit (PMU) is improved.

The invention relates to a robot positioning method and a system with monocular vision and encoder data fused, which comprise real-time calibration on external parameters and visual scale factors based on a sliding window method and pose loose coupling fusion based on the sliding window method. According to the real-time calibration on the external parameters and the visual scale factors based on the sliding window method, the monocular vision and the pose information outputted by the encoder are used to build a sliding window pose diagram, the relative pose between a vision sensor and the encoder is automatically estimated, and the scale factor for the monocular vision is calculated. The built sliding window pose diagram is used to fuse the monocular vision positioning information and the encoder positioning information, problems of poor precision of long-time positioning of the encoder and poor robustness of vision positioning are overcome, the method based on the sliding window pose diagram can effectively ensure the calculation amount of the fusion method is not added along with time, and the algorithm is thus applicable to mobile robot embedded equipment.

The invention relates to a method for detecting the video traffic parameters of a road intersection. The method for detecting the video traffic parameters of the road intersection comprises the following steps: ROI (Region of Interest) setting, selective background modeling, accurate vehicle detection, optimization of traffic flow statistics and error detection. According to the method, the selective background modeling is adopted, and a modeling background picture has high stability and high accuracy, so that a traffic flow parameter can be calculated accurately during real-time passage of vehicles; on the basis of the selective background modeling, the accurate vehicle detection step is performed, so that detection information on whether vehicles press lines can be obtained rapidly, and traffic parameter information is obtained dynamically in real time; and moreover, the detection method is suitable for being applied to complicated environments such as intersections, and has the advantages of high stability, low error rate and high accuracy.

The invention discloses a bus rapid transit arrival information release system and method accurate to the second, and belongs to the technical field of intelligent transportation application. When bus rapid transit is dispatched, a bus rapid transit vehicle provided with a vehicle-mounted GPS positioning unit and a vehicle-mounted mileage counting unit is used for driving once in advance according to a bus rapid transit line, information determined every second by the vehicle-mounted GPS positioning unit and the vehicle-mounted mileage counting unit serves as reference, a current actual position of the vehicle is directly determined in an actual running process, under the conditions that the vehicle travels to a tunnel, below an overpass and the like, GPS signals are shielded, and position data can not be output, the system determines the current position of the vehicle according to traveling mileages read by a mileage counting submodule, and accuracy of bus rapid transit vehicle dispatching is greatly improved. An electronic information display unit on a platform displays the traveling position of the bus rapid transit vehicle and simultaneously displays arrival remaining time of a latest bus rapid transit vehicle, and the remaining time can be accurate to the second.

The invention discloses a mapping system based on a laser navigation substation patrol robot as well as a method. The mapping system comprises a laser radar, an industrial personal computer, a GPS module and a vision module, wherein the laser radar and the GPS module are arranged on a patrol robot body; the vision module is arranged at the front end of the patrol robot body; the industrial personal computer acquires environment screenshots of the vision module, compares two adjacent images and analyzes the course angle of the robot; the laser radar transmits acquired distance information to the industrial personal computer, the GPS module transmits acquired coordinates information to the industrial personal computer, and the industrial personal computer transforms the distance information into a two-dimensional plane picture and rotates the plane picture to the same course angle as the robot so as to acquire the map of the current perimeter zone of the robot. With adoption of the mapping method for the substation on the basis of GPS, vision and laser radar of the laser navigation substation patrol robot, defects of the conventional substation mapping method can be overcome, the mapping precision can be improved, the mapping speed can be increased, and labor consumption can be reduced.

The invention discloses a system and method for improving bus rapid transit dispatch precision and belongs to the field of intelligent transportation applied technologies. A bus rapid transit vehicle is used for driving one time in advance according to a bus rapid transit line, wherein the bus rapid transit vehicle is provided with a vehicle-mounted GPS positioning unit and a vehicle-mounted mileage counting unit. Position information determined by the vehicle-mounted GPS positioning unit and the vehicle-mounted mileage counting unit at every second serves as reference positions, and in the process of driving, the position information detected by the direct adoption of a GPS positioning sub-module serves as actual position information of the vehicle. When the vehicle is driven under a tunnel, a flyover and the like, a GPS signal is sheltered and position data can not be output, the system determines the current position of the vehicle through the mileage counting sub-module, so that bus rapid transit is corrected to seconds per second, and the situation that issued arrival information is corrected to seconds is further guaranteed.

The invention provides a dynamic compensation method of the abrasion of a grinding wheel of a numerical control horizontal shaft surface grinding machine and relates to the technical field of numerical control grinding machines, particularly to a dynamic compensation technology of the abrasion of the grinding wheel of the numerical control horizontal shaft surface grinding machine. The active power of a main shaft or grinding head motor and for grinding is measured to indirectly calculate the abrasion of the grinding wheel and implement the dynamic compensation of the abrasion of the grinding wheel of the numerical control horizontal shaft surface grinding machine. The invention also provides a calculation method of the dynamic compensation of the abrasion of the grinding wheel of the numerical control horizontal shaft surface grinding machine. According to the dynamic compensation method of the abrasion of the grinding wheel of the numerical control horizontal shaft surface grinding machine, the dynamic compensation method is not affected by the shape of workpieces and limitations are avoided; only a main shaft motor active power transducer and a digital transmission device need to be added to a numerical control system and accordingly the cost of the added hardware is low; the machining efficiency of the numerical control surface grinding machine can be greatly improved.

Disclosed is a system for measuring the azimuth angle of a wind wheel of a wind driven generator. The system comprises a rotating speed measuring device, a zero position detection device and a controller. The rotating speed measuring device is used for detecting the rotating speed of a driving shaft. The zero position detection device outputs a first switching value signal when the gauge point of the driving shaft rotates to the predetermined position. The controller controls the rotating speed measuring device to detect the rotating speed of the driving shaft according to the preset time interval, the value of the rotating angle of the driving shaft at the preset time interval is calculated according to the rotating speed of the driving shaft, the calculated values of the rotating angles of the driving shaft at the preset time intervals are accumulated one by one to obtain an accumulated value, the accumulated value serves as the azimuth angle of the wind wheel, and the controller enables the accumulated value to be reset when receiving the first switching value signal outputted by the zero position detection device. The system for measuring the azimuth angle of the wind wheel of the wind driven generator can accomplish measurement of the azimuth angle of the wind wheel by being free of utilization of a large-diameter metal rotary table, has the high real-time performance, the high accuracy and the high continuity, and can eliminate accumulated errors.

The invention provides a face key point positioning method and device. The method comprises the steps of obtaining initial image feature vectors corresponding to face key point initial positions; inputting the initial image feature vectors into a circular neural network, obtaining updated face key point positions through utilization of the circular neural network, and taking the updated face key point positions as the face key point positions corresponding to the image feature vectors obtained at the next time; and when the face key point positions output by the circular neural network satisfy a convergence condition, obtaining the face key point final positions. According to the method and the device, the face key point positions in the image are iterated and updated through utilization of the circular neural network, the accumulation error of every-time iteration can be eliminated, and the accuracy of positioning the face key point positions is improved.

The invention discloses a plate trimming machine, which comprises a base, a workbench, a saw beam and disk saw devices, wherein one movable disk saw device is arranged on the right side of the saw beam; two movable disk saw devices are arranged on the left side of the saw beam. The plate trimming machine is characterized in that the saw beam is connected with the top ends of two upright posts to construct a gantry; the base below the saw beam extends from the left side and the right side of the saw beam, and is separated from the two upright posts; a movable workbench and a main driving device connected with the workbench are arranged on the base; workpiece clamping tongs for clamping in the horizontal direction are arranged on front side, back side, left side and right side of the workbench; a plurality of gantry hole devices for supporting workpieces in a rolling way are arranged in an arrayed manner on the workbench respectively; each gantry hole device consists of a ball and a ball base. The plate trimming machine is high in trimming accuracy, high in stability and convenient for machining.