215results about How to "Reduce sensitivity" patented technology

The invention relates to the technical field of biological medicines, in particular to a soluble skin-adhering microneedle for delivering an inactivated enterovirus vaccine as well as a preparation method and application of the soluble skin-adhering microneedle. Sodium hyaluronate is adopted as a film forming material to wrap an enterovirus 71 type (EV71)-coxsackie virus A16 (CA16) type bivalent inactivated vaccine antigen, and a needle body layer of the soluble microneedle is formed; the substrate layer is prepared from a pullulan material, so that the microneedle patch with a stable overall structure is formed. The technology provided by the invention is completed in a clean environment, and the vaccine patch which is qualified through sterility detection is obtained. The soluble microneedle disclosed by the invention has good antigen stability under the condition of 2-8 DEG C. In animal experiments, the microneedle is attached to the skin surface of a mouse for skin-adhering immunization, a strong immune response reaction can be induced, good immunogenicity is shown, and a novel, safe and effective vaccine delivery way is provided for prevention and control of the infantile hand-foot-and-mouth disease.

This utility model discloses a MEMS capacitive sensor, relating to the field of micro-motor technology. Anchor points and a fixed structure are fixedly arranged. Elastic beams and connecting beams are perpendicular to each other. The anchor points support the inner frame via the elastic beams, making the inner frame suspended. The inner frame, in turn, supports the mass frame via the connecting beams, making the mass fixed structure frame suspended. Movable comb teeth on the mass frame and fixed comb teeth on the fixed structure are spaced and interleaved to form a differential capacitor. The movable and fixed comb teeth use differential output, which can differentially eliminate the capacitance change caused by vibration or acceleration of non-detected axes, preventing signal output. This helps reduce the sensor's cross-axis sensitivity and improve sensor performance.

The embodiment of the application provides a kind of training method of image encoder, image processing method, electronic equipment, storage medium and program product, it is related to image processing technical field, the training method of image encoder includes: training image is input image encoder, visual word symbol for representing image feature is extracted by the image encoder;Clothing text word symbol is generated for training image;The clothing text word symbol is used to describe the clothing feature of the person in the training image;The model parameter of the image encoder is iteratively trained using a combination loss function;The combination loss function is used to make the visual word symbol of the same person of multiple training images be smaller, and make the visual word symbol and the clothing text word symbol of the same training image be larger difference.It reduces the degree of dependence on visual information by introducing text modal, increases the precision of image encoder to identify the same person under different clothing, improves the precision and robustness of person re-identification.

The application relates to a full-mechanical automatic tracking self-balancing variable current voltage stabilizing device and a use method thereof, which comprises a main pipe and two gear connecting rod combined mechanisms arranged on the two sides of the main pipe respectively; the gear connecting rod combined mechanism comprises a bellows, a pressing plate pressed on one end of the bellows far from a water inlet pipe, and a pressing rod pressed on the hose; one end of the bellows is sealingly connected with the pipe wall of the water inlet pipe, the other end is sealingly fixedly connected with the pressing plate, a supporting rod extending away from the bellows is fixed on the pressing plate, a spring is connected with one end of the supporting rod far from the pressing plate, and the other end of the spring far from the supporting rod is fixedly arranged; the gear connecting rod combined mechanism further comprises a driving gear and a driven gear which are mutually meshed. The application changes the flow area of the hose through the gear transmission mechanism of each connecting rod and the expansion and contraction of the bellows when the water pressure changes, so that the flow is stored and the pressure is increased when the pressure is low, the municipal end water pressure is prevented from being lower than the minimum allowable value due to excessive secondary water supply and water taking, no external power is needed, and the automatic tracking and self-balancing functions are achieved.

The application discloses a layered continuous inverse kinematics solving method for fiber reciprocating winding of a cylindrical component and belongs to the technical field of robot end control. Based on a fiber winding task of a six-axis robot cylindrical component, trajectory modeling and pose generation are carried out; a kinematics model of the six-axis robot is established; a Jacobian matrix and singularity representation thereof are obtained; end pose error is defined; the fiber winding task of the six-axis robot cylindrical component is described as a continuous inverse kinematics problem; full-pose inverse kinematics iteration is carried out by using a damped least square method; initial value generation based on state prediction is carried out for the fiber winding task of the six-axis robot cylindrical component; a position rollback strategy is set; and the continuous inverse kinematics problem is solved. The application aims at solving the problems of poor continuity of the end trajectory of the six-axis robot, strong attitude constraint and unstable end reversing solving in the fiber winding process of the cylindrical component.

The application discloses a micro light emitting diode based on polarization orientation difference regulation and a preparation method thereof. The micro light emitting diode comprises a substrate layer, and a plurality of first regions and second regions arranged alternately on the substrate layer; the first region comprises a non-doped nitrogen-polar nitride layer, a first n-type ohmic contact layer and an n-type electrode which are sequentially stacked; the second region comprises a nucleation layer, a non-doped metal-polar nitride layer, a second n-type ohmic contact layer, a multi-quantum well active region, an electron blocking layer, a p-type hole injection layer, a p++-type ohmic contact layer and a p-type electrode which are sequentially stacked; wherein the second region is a light emitting unit of the device, and a mesa structure thereof is defined by wet etching; the first region and the adjacent second region form a homopolar junction in the horizontal direction. By constructing nitride epitaxial structures with different polarities on the same substrate, a lateral carrier migration potential barrier is formed, and at the same time, etching damage is eliminated by wet etching, so that side wall non-radiative recombination is avoided.

A method for manufacturing high-strength, hydrogen-embrittlement-resistant bolts based on microsphere shot peening, and the bolts themselves, are disclosed. The manufacturing method includes: S1, cold heading and thread rolling of alloy structural steel billets to obtain bolt forming parts; S2, surface cleaning of the bolt forming parts, followed by overall surface peening with microspheres driven by high-pressure air; after overall peening, targeted strengthening peening treatment is performed on stress concentration areas such as thread roots, undercut grooves, and flange transition corners to obtain reinforced bolt parts; S3, the reinforced bolt parts are cleaned, dried, and phosphated to obtain finished bolt parts. This design, through a "overall + targeted" composite microsphere shot peening strategy, introduces a uniform residual compressive stress field and a micron-level fine-grained plastic deformation layer on the bolt surface, forming a thicker strengthening layer in stress concentration areas, significantly improving the fatigue strength and delayed fracture resistance of high-strength bolts, while ensuring uniform surface roughness and stable assembly friction coefficient.

The invention relates to a three-dimensional optical control method based on a metasurface, and the method comprises the steps: scanning a nano-column unit structure to construct a phase library, selecting high-efficiency half-wave plate nano-columns with phase coverage from-pi to pi from the phase library for the construction of the metasurface, cooperatively introducing a geometric phase and transmission phase regulation and control mechanism, and combining spatial multiplexing arrangement to construct a metasurface. Independent regulation and control of different circular polarization light fields under different wavelengths are realized, so that four switchable optical control focuses are generated in a three-dimensional space; by switching the polarization state or the working wavelength of the incident light, stable capture and controllable transfer of the particles are realized in the space. Compared with the prior art, the method has the advantages that three-dimensional optical control over particles is achieved in an on-chip structure, the control freedom degree is high, and the control adjusting mode is simple and efficient.

The application discloses a multi-mode magneto-rheological actuator, which comprises a magneto-rheological damper, a working cylinder and a control cylinder which are connected in parallel, a piston is arranged in the working cylinder, so that the working cylinder is divided into a first working cavity and a second working cavity by the piston; an adjusting valve core is arranged in the control cylinder, so that the control cylinder is divided into a first control cavity and a second control cavity by the adjusting valve core; the first working cavity, the second working cavity, the first control cavity and the second control cavity are filled with magneto-rheological fluid; the first working cavity is communicated with the first control cavity, and the communication flow in the two cavities is controllable; the second working cavity is communicated with the second control cavity, and the communication flow in the two cavities is controllable; the two cavities are filled with magneto-rheological fluid; the magneto-rheological actuator has two working modes, a mode switching switch of active damping and magneto-rheological semi-active damping is formed by a motor and the adjusting valve core, and model selection is actively carried out for different scenes, so that the damping requirements of different application scenes can be met.

The invention discloses a tendon-rope-driven dexterous hand finger for metacarpophalangeal joint movement decoupling. A flexion and extension rotating joint and a side-sway rotating joint of the tendon-rope-driven dexterous hand finger are both of a bidirectional rotating structure controlled by a rope driver; the side-sway rotating joint is of a rolling meshing structure formed by two toothed plate components, rope guide wheels are arranged on the two toothed plate components, the two rope guide wheels are used for guiding and outputting a flexion and extension driving rope of the flexion and extension rotating joint and a side-sway driving rope of the side-sway rotating joint, and the winding directions of the flexion and extension driving rope on the two rope guide wheels are opposite. The two free ends of the side-sway driving rope are guided by the rope guide wheels to be arranged in a staggered mode. After the scheme is adopted, a transmission link capable of realizing side-sway motion and flexion-extension motion is decoupled, so that the effective length change of the flexion-extension driving rope caused by the side-sway motion is completely compensated geometrically, and the effective length of the flexion-extension link is kept constant at any side-sway angle; and the action stability and the repeated positioning performance of the fingers of the dexterous hand under the multi-posture and multi-load conditions are improved.

The invention relates to the technical field of light field regulation and control, and discloses a static phase distortion compensation system of a spatial light modulator, which comprises a polarization splitting prism arranged in the light emitting direction of a laser and used for adjusting a light beam emitted by the laser into linearly polarized light in the horizontal direction, the first quarter-wave plate adjusts the linearly polarized light into circularly polarized light and irradiates the circularly polarized light to the spatial light modulator; a second quarter-wave plate arranged in the light emitting direction of the spatial light modulator adjusts an emergent light beam of the spatial light modulator into linearly polarized light, and a linear polarizer performs polarization analysis on the emergent light beam of the second quarter-wave plate; the static phase error of the spatial light modulator is determined through the polarization direction included angle of the incident beam of the first quarter-wave plate and the emergent beam of the second quarter-wave plate, and the polarization direction included angle is determined through the Malus law relation of the incident light intensity and the transmission light intensity of the linear polarizer. The static phase distortion of the spatial light modulator is compensated by using the phase error, and the static phase distortion compensation cost is reduced.

The application discloses an anchor point guided clustering based data classification method, device and equipment, relates to the technical field of digital data processing, and comprises the following steps: obtaining an original data set to be classified, converting each sample data into a numerical vector, and constructing a data matrix; initializing a clustering center matrix and an anchor point matrix, and setting a fuzzy coefficient; obtaining the anchor point matrix and the clustering center matrix after joint iterative optimization; calculating the fuzzy membership degree of each sample vector; the category with the maximum fuzzy membership degree is taken as the final category label of the sample vector, and the classification results of all sample vectors are output. The application solves the problems that the existing fuzzy clustering method is sensitive to initial conditions, is easy to fall into a suboptimal solution, leads to unstable and inaccurate classification results, and cannot be directly solved by using a gradient descent algorithm, and is difficult to be applied to large-scale data sets, and realizes the enhancement of complex data classification precision and the applicability in different data scale scenes.

The invention discloses a photoelectric sensor system conversion efficiency calibration system and method based on a reverberation chamber, and belongs to the technical field of photoelectric sensor conversion efficiency calibration, the system comprises an electric wave reverberation chamber, a transmitting antenna, a reference antenna and a photoelectric field sensor system; the photoelectric field sensor system is composed of a photoelectric sensor, a laser source, a photoelectric detector, a polarization maintaining optical fiber and a single-mode optical fiber. In the calibration process, the reference antenna and the photoelectric sensor system are used as receiving ends respectively, receiving power at different stirring positions in the reverberation chamber is sampled to obtain two groups of transfer functions, and the total efficiency of the photoelectric sensor system is determined by comparing the two groups of transfer functions under the condition that the radiation efficiency of the reference antenna is known. The conversion efficiency of the photoelectric sensor system is obtained by combining the relation among the power, the field intensity and the total efficiency in the free space; and the effective averaging of the incident direction and the polarization state is realized by using the uniform and isotropic electromagnetic field characteristics of the reverberation chamber.

The utility model discloses a printed circuit board, a communication device and a vehicle. The printed circuit board comprises a dielectric related interface, an electrostatic impedor, a common-mode resistor, an alternating-current coupling capacitor, a common-mode filter and a control chip, and a differential signal line is connected to the control chip through the electrostatic impedor, the common-mode resistor, the alternating-current coupling capacitor and the common-mode filter in sequence; wherein the differential signal lines are arranged based on a cross-layer reference mode; wiring of the positive signal line and the negative signal line of the differential signal line, pin bonding pads on wiring paths and via holes in the wiring paths are all symmetrically arranged, and the positive signal line and the negative signal line are matched in equal length. Through the symmetrical arrangement and equal-length matching of the differential signal lines and the use of the electrostatic impedor, the common-mode resistor, the alternating-current coupling capacitor, the common-mode filter and other elements, noise and interference are effectively suppressed, and the integrity of signals is improved. The cross-layer reference and the reasonable wiring strategy help to reduce electromagnetic radiation and sensitivity, and improve the electromagnetic compatibility of the printed circuit board.

The invention provides a control method and system for local gluing of the outer surface of a metal round piece, and belongs to the technical field of intelligent gluing control. The method comprises the following steps: firstly, acquiring a current outer surface image and process task data of a to-be-glued metal round piece, and determining a target local gluing area, a state detection area and a target glue line parameter; releasing detection glue points in a state detection area, acquiring a detection dynamic image sequence, and extracting spreading length, edge expansion rate, necking rate, broken wire time delay and retraction displacement to form a dynamic morphology feature sequence; inputting the dynamic morphology feature sequence into a glue solution rheological state inversion model, and solving a current glue solution rheological state parameter; local gluing is carried out according to the formal gluing control instruction, and a local glue line forming result is output in combination with an outer surface result detection record. According to the method, the current flowing state of the glue solution can be recognized before formal gluing, gluing control parameters are synchronously adjusted according to the current flowing state, and the forming stability of local gluing glue lines on the outer surface of the metal round piece is improved.

This invention discloses a high-precision clock synchronization method between two wireless devices, comprising: Initially, a first wireless device sends a first message containing a transmission timestamp to a second wireless device. Upon receiving the message, the second wireless device overwrites its local reception timestamp with this first message, completing initial clock alignment. Subsequently, after a turnaround time, the second wireless device returns a second message to the first wireless device and calculates the one-way time of flight based on the time interval and turnaround time. At the next predefined superframe moment, the first wireless device sends a third message. The second wireless device calculates and corrects its local clock frequency offset based on this message and the timestamp of the previous message, while simultaneously correcting the local reception timestamp of the third message using the one-way time of flight. Thereafter, the above message exchange and calculation process is repeated at each superframe moment to continuously maintain high-precision synchronization. This invention fundamentally improves the clock synchronization accuracy between wireless devices.

The application relates to a wireless charging device wake-up method and circuit based on double detection, the wake-up method comprising the following steps: performing a digital Q value detection process, outputting a corresponding first detection result, judging whether the first detection result exceeds a first preset threshold range, if the first detection result does not exceed the first preset threshold range, performing an analog PING detection process, outputting a corresponding second detection result, judging whether the second detection result exceeds a second preset threshold range, if the second detection result does not exceed the second preset threshold range, re-entering a deep sleep mode, and if the second detection result exceeds the second preset threshold range, performing a corresponding wake-up process; the two detection results jointly act, so that the system only needs to perform a digital detection with extremely low power consumption to maintain the wake-up judgment in most cases, and only enters a high-precision detection stage when there is uncertainty, thereby realizing the intelligent wake-up effect with high reliability, low false triggering rate and extremely low standby power consumption in a complex environment.

The invention discloses a direct current charging interface identification and positioning method, and belongs to the technical field of automatic charging of electric vehicles. The method comprises a coarse positioning stage and a fine positioning stage. In the coarse positioning stage, the initial position of the charging port is quickly determined through image preprocessing, contour screening and ellipse fitting. In the fine positioning stage, a multi-factor cluster matching algorithm is adopted, and firstly, a matching preselected area of each feature hole is determined according to a preset geometric cluster relationship among a plurality of feature holes in the charging port; secondly, dynamically calculating scaling and rotation coefficients of the template to adapt to different postures; and finally, constructing a matching degree evaluation function in combination with a central gravitation and edge gravitation principle, and performing matching calculation on an original image and a gradient image template so as to realize accurate positioning of the feature hole. The method solves the problem that the recognition precision and efficiency of the charging port are difficult to consider at the same time in a complex light field environment, and has the advantages of high positioning precision, high processing speed, high environmental adaptability and easy engineering deployment.

The invention relates to the technical field of intelligent sensing systems, and discloses a formaldehyde and total volatile organic compound detection device and detection method.The formaldehyde and total volatile organic compound detection device comprises a pretreatment filter screen with a hydrolyzable hydrophilic coating on the surface; the pressure difference monitoring module is used for outputting a pressure difference monitoring value reflecting the permeability of the pretreatment filter screen; the reference gas providing module is used for continuously generating and conveying clean air; the differential sensing module is used for providing a background reference signal for the processor; and the processor is used for dynamically compensating the detection value of the gas sensor according to the physical blockage state of the pretreatment filter screen. According to the invention, hydrated dust is actively captured and removed, an error is judged by analyzing a physical signal provided by the differential pressure monitoring module and an electrical signal provided by the differential sensing module, and finally, a compensation coefficient is dynamically generated to correct the reading of the detection device in real time, so that the credibility of a detection result of the detection device is improved.

This invention, entitled "A Distributed Hybrid Interleaved Batch Scheduling Method and System for Parallel Production Lines," belongs to the field of production line scheduling. It addresses the problem of minimizing inventory and stockout costs. The method includes: complete delivery based on vehicle model bill of materials; each production line comprises multiple sequentially connected stages, with the same set of dies only allowed on one production line at any given time; during the stamping stage, only one production line can utilize the stamping die to process the corresponding sub-batch; a consistent sub-batch partitioning strategy is adopted for the same batch, ensuring that the quantity and corresponding dimensions of each sub-batch remain unchanged in subsequent processing stages after partitioning; the dimensions between different sub-batches can be determined by the selected sub-batch partitioning rules; alternating processing of sub-batches from different batches is allowed on the same production line; each sub-batch can only be processed by one production line in each stage; the scheduling objective is to minimize the total sum of inventory and stockout costs while achieving complete delivery.

This utility model discloses an anti-leakage plating structure for a plastic metallized waveguide antenna, relating to the field of waveguide antenna technology. The anti-leakage plating structure for a plastic metallized waveguide antenna includes a waveguide groove disposed on the waveguide antenna body. Each right-angle section inside the waveguide groove has a guiding surface to direct the flow of electroplating solution. This utility model, by setting guiding surfaces in the form of arcs, slopes, or a combination of both at the right angles inside the waveguide groove, completely eliminates dead zones in the flow of electroplating solution, guiding the plating solution smoothly and comprehensively covering the bottom and corners of the waveguide groove. This effectively solves the plating leakage problem that easily occurs in traditional inner right-angle structures, ensuring that the plating layer is complete and of uniform thickness, thus guaranteeing the stability and reliability of the antenna's core conductivity. Simultaneously, it eliminates the need for multiple electroplating operations and increased plating solution flow rates, simplifying the production process, reducing sensitivity to equipment parameters and debugging difficulty, improving production efficiency, and reducing electroplating costs.

The invention provides a power hardware-in-the-loop interface compensation method based on power amplifier parameter identification. Parameter identification is carried out on a power amplifier system, parameter information of a passive device in the power amplifier is estimated, and the robustness of the system to parameter changes is improved. Due to the existence of a filter in a power amplifier, deviation of actual power transmission and ideal power transmission characteristics is generally caused, and the amplitude and phase angle deviation caused by the filter is compensated through parameters identified in an interface method, so that the power transmission precision is improved, and the reliability of power hardware in the loop is improved. In addition, the compensated system is closer to an ideal system, is more suitable for power hardware in-loop stability analysis and verification, and particularly has remarkable advantages under the condition of parameter aging offset of the power amplifier.

The application belongs to the field of automatic control technology of aircraft, and proposes a quadrotor double pendulum finite time swing elimination method based on full drive mapping and active anti-disturbance, which is applied to a quadrotor unmanned aerial vehicle double pendulum hoisting system. The method adopts a hierarchical control structure composed of an outer ring position and swing elimination control module and an inner ring bottom layer flight control module. The method comprises the following steps: establishing a quadrotor double pendulum horizontal plane block matrix dynamics model; full drive state mapping and dynamic reconstruction facing the controller; Y-axis full drive space isomorphic expansion; UDE-based finite time control law design; closed loop system stability and finite time convergence analysis based on Lyapunov theorem. The application can improve the waypoint tracking accuracy, swing angle suppression effect and robust flight capability of the quadrotor double pendulum hoisting system under the working condition that discrete waypoints are switched and external wind disturbance exists simultaneously.

This application discloses a low-power temperature-compensated RC oscillator, comprising a low-voltage linear buffer, an RC core circuit, a comparator, a Schmitt trigger, a non-overlapping clock generator, and a level shifter. The low-voltage linear buffer uses intrinsic transistors to reduce operating voltage and power consumption; the RC core circuit employs temperature compensation technology, including a series low-temperature drift resistor and a temperature-dependent current source, causing the equivalent resistance to increase with temperature; the comparator compares a triangular wave signal with a reference voltage and outputs a reference clock; the Schmitt trigger filters out clock glitches; the non-overlapping clock generator generates a two-phase clock with dead time to control charging and discharging; and the level shifter converts the low-voltage domain clock to the high-voltage domain. This oscillator eliminates the influence of comparator offset voltage through two-phase switching control and employs low-voltage and temperature compensation technology to achieve a high-precision, low-temperature drift reference frequency in a small area with low power consumption, making it suitable for compact sensing and measurement systems.

The invention discloses an all-solid-state battery positive electrode material. The all-solid-state battery positive electrode material comprises a positive electrode active material and a composite sulfide solid electrolyte, the composite sulfide solid electrolyte comprises a sulfide solid electrolyte matrix and a metal fluoride layer coated on the surface of the sulfide solid electrolyte matrix, and the mass ratio of the positive active material to the composite sulfide solid electrolyte is 7: 3. The sulfide solid electrolyte is coated with a layer of metal fluoride, on one hand, the stable metal fluoride effectively isolates sulfide from being in direct contact with air, the air stability of the sulfide solid electrolyte is greatly improved, and the problem that the sulfide solid electrolyte is sensitive to air is relieved; on the other hand, the metal fluoride coating layer can prevent sulfide solid electrolyte from being in direct contact with the positive electrode material, the instability between the positive electrode material and the electrolyte is relieved, and an effective thought is provided for industrialization of the all-solid-state battery.

The invention relates to semiconductor manufacturing equipment which comprises an equipment body, a vacuum pressure gauge and a heat dissipation piece, a vacuum cavity is formed in the equipment body, the vacuum pressure gauge is used for measuring the pressure of the vacuum cavity and is provided with a measuring cavity, the heat dissipation piece comprises a shell connected between the equipment body and the vacuum pressure gauge, and a heat dissipation cavity is formed in the shell. And the heat dissipation cavity is used for communicating the vacuum cavity with the measurement cavity so as to dissipate heat of the polymer entering the measurement cavity from the vacuum cavity. According to the semiconductor manufacturing equipment disclosed by the invention, the polymer entering the vacuum pressure gauge is subjected to heat dissipation through the heat dissipation piece, so that the polymer is gathered in the heat dissipation cavity after entering the heat dissipation cavity by utilizing the characteristic that the gathering speed of the polymer in a low-temperature environment is higher than that in a high-temperature environment; the polymer entering the measuring cavity of the vacuum pressure gauge can be reduced to a great extent, it is guaranteed that the polymer does not cause reduction of the precision and sensitivity of the vacuum pressure gauge, the service life of the vacuum pressure gauge is prolonged, and the frequency of maintaining the vacuum pressure gauge is reduced.

This invention provides a method and system for assembling a genome stack, a method and system for assembling a genome architecture, and a method and system for assembling a genome sequence, belonging to the field of bioinformatics. This invention optimizes a global loss function using robust regression techniques to determine the stacking relationships between sequencing sequences, eliminating interference from false positive alignments, thereby obtaining a more accurate assembled stack set and avoiding copy number loss or misassembly in the assembled genome. The estimation of stack arrangement based on regression matrix calculations has a higher tolerance for outliers; unless the number of ambiguous sample information exceeds the number of true samples, a "collapse" will not occur. By using resampling techniques to independently sample, assemble, improve, and evaluate sequencing data multiple times, and then integrating them into the final assembled genome, the uncertainty of the assembly results caused by noise in the sequencing data can be reduced, and the sensitivity of the assembly results to the selection of assembly parameters can be decreased.

The invention relates to the field of atomic interference measurement, in particular to an atomic Raman interferometer based on a flat-topped beam. Comprising two Raman interference subsystems arranged in parallel, each of the two Raman interference subsystems comprises an ultrahigh vacuum cavity and a light path system, and the light path system comprises a flat-topped beam shaping module located on one side of the high vacuum cavity and a Raman light Doppler sensitive configuration module located on the other side of the high vacuum cavity; in the flat-topped beam shaping module, the holographic algorithm principle is adopted, the actual size of atoms in an interference area is combined to determine the size of a target flat-topped beam, the initial complex amplitude is precisely constrained, and a final diffraction phase is obtained through iterative optimization. After the Raman light passes through the diffraction phase, the Raman light is shaped into a flat-topped light beam with a target size and uniform intensity distribution at the ultrahigh vacuum cavity, so that the coherence efficiency under the same interference time is remarkably improved.

The application discloses a multispectral radiation temperature measurement inversion method based on an improved Bayesian optimization generalized regression neural network, and comprises the following steps: carrying out pretreatment on an infrared multispectral image sequence, corresponding the pretreated data to a target real temperature, and constructing a data set; establishing a generalized regression neural network model for real temperature regression inversion of a target temperature measurement area; based on the data set, adopting a bionic optimization algorithm as a sampling method of Bayesian optimization, and optimizing hyperparameters in the generalized regression neural network model by using an improved Bayesian optimization algorithm, so as to inversely calculate the target real temperature by using the optimized generalized regression neural network model. The method does not need to assume an emissivity model, has higher temperature measurement efficiency and temperature measurement precision, and has wider applicability.

This invention discloses a collaborative training method and system for autonomous driving planning and control, belonging to the field of autonomous driving technology. The method includes: acquiring the current state of the vehicle, the states of surrounding traffic participants, and reference road information; using a Transformer-based planning network to output a control sequence, which is then used to generate a planned trajectory via a vehicle kinematics model; inputting the planned trajectory into a model predictive control-based neural network controller for dynamic tracking optimization to obtain the executed control quantity and predicted vehicle dynamic state; constructing a control loss and backpropagating its gradient to the planning network; and, during training, layering constraints on the planning cost term, using safety and traffic rule-related terms as hard constraints and comfort-related terms as soft constraints, and employing a multi-stage cost scheduling strategy to adjust the constraint strength or weight. This invention achieves collaborative optimization of planning and control, improves the dynamic executability and safety compliance of the trajectory, and enhances training stability and closed-loop robustness.