CN112558464A - Aircraft controller gain scheduling method adaptive to strong pneumatic nonlinearity - Google Patents

Aircraft controller gain scheduling method adaptive to strong pneumatic nonlinearity Download PDF

Info

Publication number
CN112558464A
CN112558464A CN202010993584.2A CN202010993584A CN112558464A CN 112558464 A CN112558464 A CN 112558464A CN 202010993584 A CN202010993584 A CN 202010993584A CN 112558464 A CN112558464 A CN 112558464A
Authority
CN
China
Prior art keywords
flight
network
gain scheduling
controller
aircraft
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202010993584.2A
Other languages
Chinese (zh)
Inventor
奚勇
何飞毅
廖幻年
陈光山
徐桂甲
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shanghai Aerospace Control Technology Institute
Original Assignee
Shanghai Aerospace Control Technology Institute
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shanghai Aerospace Control Technology Institute filed Critical Shanghai Aerospace Control Technology Institute
Priority to CN202010993584.2A priority Critical patent/CN112558464A/en
Publication of CN112558464A publication Critical patent/CN112558464A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B11/00—Automatic controllers
    • G05B11/01—Automatic controllers electric
    • G05B11/36—Automatic controllers electric with provision for obtaining particular characteristics, e.g. proportional, integral, differential
    • G05B11/42—Automatic controllers electric with provision for obtaining particular characteristics, e.g. proportional, integral, differential for obtaining a characteristic which is both proportional and time-dependent, e.g. P. I., P. I. D.

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Automation & Control Theory (AREA)
  • Feedback Control In General (AREA)

Abstract

The invention discloses an aircraft controller gain scheduling method adaptive to strong pneumatic nonlinearity, which is characterized by comprising the following steps: s1, selecting characteristic points according to flight state parameters of an aircraft, and designing PD control parameters of a PD controller as samples; s2, training a gain scheduling network of the controller according to the samples in the step 1 to obtain a gain scheduling network model of the controller; and S3, inputting the real-time flight state parameters of the aircraft into a gain scheduling network of the controller, outputting the real-time flight control parameters, and outputting rudder instructions to the rudder system in real time through the PD controller. According to the invention, a flight control network based on a deep neural network is designed according to flight state parameters of the aircraft and corresponding control parameters, and the flight control parameters are calculated in real time through the control network, so that the control quality of the control system under strong nonlinearity, high dynamic and asymmetric pneumatics is effectively improved, and support is provided for full airspace high maneuvering flight of the aircraft.

Description

Aircraft controller gain scheduling method adaptive to strong pneumatic nonlinearity
Technical Field
The invention relates to the field of flight control of tactical weapons, in particular to an aircraft controller gain scheduling method adaptive to strong pneumatic nonlinearity. The method is characterized by solving the problems that the traditional parameter adjusting method is difficult to adapt to the strong nonlinearity and high dynamic change of aerodynamic characteristics of tactical weapons under the conditions of non-axisymmetric appearance, large airspace and high maneuvering flight, and improving the non-flight control quality of a stable control system.
Background
Traditionally, a tactical weapon flight control system is designed mainly based on a PD control method, and the main design idea is to select a certain number of typical characteristic points to design control parameters according to flight state information of an object in a flight envelope, such as speed, height, dynamic pressure, attack angle and the like, and the corresponding dynamic coefficient change characteristics, and then obtain a corresponding parameter adjusting rule by adopting methods such as linear fitting, polynomial fitting and the like to meet the control performance requirement in a real-time flight environment. Since the linear fitting method cannot realize accurate approximation of the nonlinear state, the control requirement under nonlinear change is difficult to be considered, and the polynomial fitting method is difficult to determine the polynomial form and parameters along with the improvement of the complexity of the nonlinear characteristic. Therefore, the control quality of the method under the condition of severe pneumatic nonlinear change is difficult to guarantee, a more accurate gain scheduling method needs to be adopted to adapt to nonlinear control requirements, and the control performance of the tactical weapon under the conditions of non-axisymmetric appearance, large airspace and high maneuvering flight is effectively improved.
Disclosure of Invention
The invention aims to provide an aircraft controller gain scheduling method adaptive to strong pneumatic nonlinearity, which is characterized in that a flight control network based on a deep neural network is designed according to flight state parameters of an aircraft and corresponding control parameters, and the flight control parameters are calculated in real time through the control network, so that the control quality of a control system under strong nonlinearity, high dynamic and asymmetric pneumatics is effectively improved, and support is provided for full airspace high maneuvering flight of an object.
In order to achieve the purpose, the invention is realized by the following technical scheme: an aircraft controller gain scheduling method adaptive to strong pneumatic nonlinearity is characterized by comprising the following steps:
s1, selecting characteristic points according to flight state parameters of an aircraft and designing PD control parameters of a PD controller as samples;
s2, training a gain scheduling network of the controller according to the samples in the step 1 to obtain a gain scheduling network model of the controller;
and S3, inputting the real-time flight state parameters of the aircraft into a gain scheduling network of the controller, outputting the real-time flight control parameters, and outputting rudder instructions to the rudder system in real time through the PD controller.
Preferably, the flight state parameters in step S1 include the following parameters:
flight time, flight phase, mass center, rotating inertia of the projectile around the X axis, rotating inertia of the projectile around the Y axis, rotating inertia of the projectile around the Z axis, flight Mach number, dynamic pressure, flight speed, thrust, flight altitude, attack angle, sideslip angle, angular speed of the projectile around the X axis, angular speed of the projectile around the Y axis, angular speed of the projectile around the Z axis, rolling rudder deflection, yawing rudder deflection and pitching rudder deflection.
Preferably, the controller gain scheduling network in step S2 is as follows:
adopting a 5-level cascade feedforward neural network model:
Figure RE-GDA0002933615010000021
in the formula, l is the number of network layers, x is network input, namely flight state parameters, and y is network output, namely PD control parameters; [ h ] of(0),h(1),h(2),…,h(4)]For each layer input, [ eta ](0),η(1),η(2),…,η(4)]Outputting for each layer; [ n ] of1,n2,n3,…,n5]Dimension for the input data on each layer; and n is1＝20，n2＝40， n3＝60，n4＝76，n5＝88；
Figure RE-GDA0002933615010000022
An activation function for each layer, and the activation function is a tansig function;
Figure RE-GDA0002933615010000023
is the first layerThe weight value corresponding to the network; b(l)Corresponding bias for the l-th network.
Preferably, the network training method in step S2 is:
s2.1, performing redundancy removal and normalization processing on 20 flight state parameters in the sample;
s2.2, inputting the processed data into a 5-level cascade feedforward neural network model and calculating to obtain PD control parameter output;
s2.3: and calculating the mean square error between the output control parameters of the network model and the control parameters in the samples, establishing and optimizing a loss function, and updating the model weight through error back propagation.
The specific method for updating the model weight comprises the following steps:
establishing a loss function J (w), solving an error propagation term based on a Levenberg-Marquardt algorithm, and updating a weight w on each layer;
Figure RE-GDA0002933615010000031
Figure RE-GDA0002933615010000032
in the formula, ynFor designing the PD control parameters according to the characteristic points,
Figure RE-GDA0002933615010000033
outputting PD control parameters for the neural network; l (w) is the empirical risk, R (w) is the regularization term,
Figure RE-GDA0002933615010000034
is Frobenius norm, lambda is penalty coefficient, and N is number of training data sets.
S2.4: and repeating the step 22 and the step 23 until the mean square error meets the set precision requirement.
Preferably, in step S3, the rudder instruction output method includes:
Figure RE-GDA0002933615010000035
Figure RE-GDA0002933615010000036
in the formula, KP、KDOutputting control parameters for the trained deep neural network, wherein deltan is the deviation between the expected overload and the feedback overload,
Figure RE-GDA0002933615010000037
as attitude angular velocity, δcIs a rudder command.
The aircraft controller gain scheduling method suitable for strong pneumatic nonlinearity has the following advantages: a gain scheduling network of a control system is designed by adopting a deep neural network, so that a complex calculation process of the traditional parameter adjusting rule under strong nonlinear characteristics can be avoided; by utilizing the accurate approximation capability of the neural network to the nonlinear characteristic, the problem of control quality reduction caused by the consideration of the nonlinear characteristic under the traditional parameter design can be avoided, and the control performance of the control system is fully developed. The method solves the control problem of the tactical weapon under the conditions of non-axisymmetric appearance, large airspace and high maneuvering flight by establishing the control gain scheduling network based on the deep neural network, and can provide support for the full airspace high maneuvering flight of the aircraft.
Drawings
FIG. 1 is a flow chart of an aircraft controller gain scheduling method of the present invention that accommodates strong aerodynamic non-linearities;
FIG. 2 is a neural network training result;
FIG. 3 is a result of fitting the fitting engagement depth neural network tuning contrast piecewise linearly.
Detailed Description
The present invention will now be further described by way of the following detailed description of a preferred embodiment thereof, taken in conjunction with the accompanying drawings.
As shown in fig. 1, an aircraft controller gain scheduling method for adapting to strong pneumatic nonlinearity includes the following steps:
s1, selecting characteristic points according to flight state parameters of an aircraft and designing PD control parameters of a PD controller as samples;
according to the known object flight envelope, selecting characteristic points in the flight airspace according to flight time, flight phases, flight Mach numbers, dynamic pressure, flight speed, thrust, flight altitude, attack angles and sideslip angles, calculating corresponding power coefficients by combining mass, mass centers, rotational inertia of the projectile body around an X axis, rotational inertia of the projectile body around a Y axis, rotational inertia of the projectile body around a Z axis, angular speed of the projectile body around an X axis, angular speed of the projectile body around a Y axis, angular speed of the projectile body around a Z axis, deviation of a rolling rudder, deviation of a yawing rudder and deviation state information of a pitching rudder, and designing PD control parameters meeting the requirements of stability and rapidity according to a pole allocation method.
S2, training a gain scheduling network of the controller according to the samples in the step 1 to obtain a gain scheduling network model of the controller;
adopting a 5-level cascade feedforward neural network model:
Figure RE-GDA0002933615010000041
in the formula, l is the number of network layers, x is network input, namely flight state parameters, and y is network output, namely PD control parameters; [ h ] of(0),h(1),h(2),…,h(4)]For each layer input, [ eta ](0),η(1),η(2),…,η(4)]Outputting for each layer; [ n ] of1,n2,n3,…,n5]Dimension for the input data on each layer; and n is1＝20，n2＝40， n3＝60，n4＝76，n5＝88；
Figure RE-GDA0002933615010000042
An activation function for each layer, and the activation function is a tansig function;
Figure RE-GDA0002933615010000043
for layer I network correspondenceThe weight of (2); b(l)Corresponding bias for the l-th network.
Training a neural network according to the following method:
s2.1, performing redundancy removal and normalization processing on 20 flight state parameters in the sample;
s2.2, inputting the processed data into a neural network model and calculating to obtain PD control parameter output;
s2.3: and calculating the mean square error between the output control parameters of the network model and the control parameters in the samples, establishing and optimizing a loss function, and updating the model weight through error back propagation.
The specific method for updating the model weight comprises the following steps:
establishing a loss function J (w), solving an error propagation term based on a Levenberg-Marquardt algorithm, and updating a weight w on each layer;
Figure RE-GDA0002933615010000051
Figure RE-GDA0002933615010000052
in the formula, ynFor designing the PD control parameters according to the characteristic points,
Figure RE-GDA0002933615010000053
outputting PD control parameters for the neural network; l (w) is the empirical risk, R (w) is the regularization term,
Figure RE-GDA0002933615010000054
is Frobenius norm, lambda is penalty coefficient, and N is number of training data sets.
S2.4: and repeating the step 22 and the step 23 until the mean square error meets the set precision requirement.
And S3, inputting the real-time flight state parameters of the aircraft into a gain scheduling network of the controller, outputting the real-time flight control parameters, and outputting rudder instructions to the rudder system in real time through the PD controller.
Writing the trained gain scheduling network into a missile-borne computer, inputting the flight state parameters sensed in real time into the gain scheduling network, and calculating a flight control parameter KP，KDAnd the method is applied to a stable control system and outputs a rudder instruction to a rudder system in real time.
The rudder instruction output method comprises the following steps:
Figure RE-GDA0002933615010000055
Figure RE-GDA0002933615010000056
in the formula, KP、KDOutputting control parameters for the trained deep neural network, wherein deltan is the deviation between the expected overload and the feedback overload,
Figure RE-GDA0002933615010000057
as attitude angular velocity, δcIs a rudder command.
Based on the detailed steps, the control gain scheduling network design and the online control gain scheduling can be completed, and the method is suitable for the strong nonlinear characteristic of the aircraft.
The specific application is as follows: selecting an area with severe nonlinear change on the flight envelope of the object, selecting characteristic points according to 20 flight states, designing control parameters to meet the requirements on stability and rapidity, obtaining 2000 groups of training samples, and dividing the training samples into a training set, a verification set and a test set according to the proportion of 3:1: 1. The feature points and the control parameters are input into a 5-level combined feedforward neural network for training, and the training results (see fig. 2) show that the results of the training set, the verification set and the test set are all converged to be near 0.0001, which shows that a better training effect is achieved.
And adding the trained network into a control system, so that the network can output control parameters according to the flight state on the trajectory. By setting a typical trajectory, changing the static stability of the trajectory from 0.5 to-1.0 within 370S to 390S, calculating a rudder instruction by adopting the method in S3, and performing fixed point simulation on characteristic points on the trajectory, the simulation result shown in FIG. 3 can be seen, and the result shows that the deep neural network fitting can realize higher stability margin and faster response speed of the control system when the nonlinear change is severe.
Therefore, the method can accurately adjust the control parameters in the nonlinear severe change area and improve the control quality.
While the present invention has been described in detail with reference to the preferred embodiments, it should be understood that the above description should not be taken as limiting the invention. Various modifications and alterations to this invention will become apparent to those skilled in the art upon reading the foregoing description. Accordingly, the scope of the invention should be determined from the following claims.
The invention has not been described in detail in part of the common general knowledge of those skilled in the art.

Claims (6)

1. An aircraft controller gain scheduling method adapting to strong pneumatic nonlinearity is characterized by comprising the following steps:
step 1: selecting characteristic points according to flight state parameters of the aircraft, and designing PD control parameters of a PD controller as samples;
step 2: training a gain scheduling network of the controller according to the samples in the step 1 to obtain a gain scheduling network model of the controller;
and step 3: and inputting the real-time flight state parameters of the aircraft into a gain scheduling network of the controller, outputting the real-time flight control parameters, and outputting rudder instructions to the rudder system in real time through the PD controller.
2. The aircraft controller gain scheduling method of claim 1 wherein said flight state parameters of step 1 comprise one or more of the following parameters:
flight time, flight phase, mass center, rotating inertia of the projectile around the X axis, rotating inertia of the projectile around the Y axis, rotating inertia of the projectile around the Z axis, flight Mach number, dynamic pressure, flight speed, thrust, flight altitude, attack angle, sideslip angle, angular speed of the projectile around the X axis, angular speed of the projectile around the Y axis, angular speed of the projectile around the Z axis, rolling rudder deflection, yawing rudder deflection and pitching rudder deflection.
3. The aircraft controller gain scheduling method of claim 1 wherein the controller gain scheduling network model is obtained in step 2 by:
adopting a 5-level cascade feedforward neural network model:
Figure RE-FDA0002933613000000011
in the formula, l is the number of network layers, x is network input, namely flight state parameters, and y is network output, namely PD control parameters; [ h ] of(0),h(1),h(2),…,h(4)]For each layer input, [ eta ](0),η(1),η(2),…,η(4)]Outputting for each layer; [ n ] of1,n2,n3,…,n5]Dimension for the input data on each layer; and n is1＝20，n2＝40，n3＝60，n4＝76，n5＝88；
Figure RE-FDA0002933613000000012
An activation function for each layer, and the activation function is a tansig function;
Figure RE-FDA0002933613000000013
the weight value corresponding to the first layer network; b(l)Corresponding bias for the l-th network.
4. The aircraft controller gain scheduling method of claim 3 wherein the network training method in step 2 is:
step 21: carrying out redundancy removal and normalization processing on the flight state parameters in the sample;
step 22: inputting the processed data into a 5-level cascade feedforward neural network model, and calculating to obtain PD control parameter output;
step 23: and calculating the mean square error between the output control parameters of the network model and the control parameters in the samples, establishing and optimizing a loss function, and updating the model weight through error back propagation.
Step 24: and repeating the step 22 and the step 23 until the mean square error meets the set precision requirement.
5. The aircraft controller gain scheduling method of claim 4, wherein in step 23, the specific method for updating the model weights comprises:
establishing a loss function J (w), solving an error propagation term based on a Levenberg-Marquardt algorithm, and updating a weight w on each layer;
Figure RE-FDA0002933613000000021
Figure RE-FDA0002933613000000022
in the formula, ynFor designing the PD control parameters according to the characteristic points,
Figure RE-FDA0002933613000000023
outputting PD control parameters for the neural network; l (w) is the empirical risk, R (w) is the regularization term,
Figure RE-FDA0002933613000000024
is Frobenius norm, lambda is penalty coefficient, and N is number of training data sets.
6. The aircraft controller gain scheduling method of claim 1, wherein the rudder instruction output method in step 3 is:
Figure RE-FDA0002933613000000025
Figure RE-FDA0002933613000000026
in the formula, KP、KDOutputting control parameters for the trained deep neural network, wherein deltan is the deviation between the expected overload and the feedback overload,
Figure RE-FDA0002933613000000027
as attitude angular velocity, δcIs a rudder command.
CN202010993584.2A 2020-09-21 2020-09-21 Aircraft controller gain scheduling method adaptive to strong pneumatic nonlinearity Pending CN112558464A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010993584.2A CN112558464A (en) 2020-09-21 2020-09-21 Aircraft controller gain scheduling method adaptive to strong pneumatic nonlinearity

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010993584.2A CN112558464A (en) 2020-09-21 2020-09-21 Aircraft controller gain scheduling method adaptive to strong pneumatic nonlinearity

Publications (1)

Publication Number Publication Date
CN112558464A true CN112558464A (en) 2021-03-26

Family

ID=75041049

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010993584.2A Pending CN112558464A (en) 2020-09-21 2020-09-21 Aircraft controller gain scheduling method adaptive to strong pneumatic nonlinearity

Country Status (1)

Country Link
CN (1) CN112558464A (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106507982B (en) * 2009-12-31 2014-04-23 清华大学 Method using neural fusion on-line amending gain scheduling
CN104270041A (en) * 2014-09-26 2015-01-07 广州航海学院 Rimer motor synchronous speed regulating control system based on active disturbance rejection control technology
CN105652880A (en) * 2016-02-24 2016-06-08 中国人民解放军海军航空工程学院 Non-linear anti-saturation height instruction generating method for aircraft large airspace flight
CN107065571A (en) * 2017-06-06 2017-08-18 上海航天控制技术研究所 A kind of objects outside Earth soft landing Guidance and control method based on machine learning algorithm
CN110488852A (en) * 2019-08-28 2019-11-22 北京航空航天大学 A kind of hypersonic aircraft complete section surface self-adaption control method

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106507982B (en) * 2009-12-31 2014-04-23 清华大学 Method using neural fusion on-line amending gain scheduling
CN104270041A (en) * 2014-09-26 2015-01-07 广州航海学院 Rimer motor synchronous speed regulating control system based on active disturbance rejection control technology
CN105652880A (en) * 2016-02-24 2016-06-08 中国人民解放军海军航空工程学院 Non-linear anti-saturation height instruction generating method for aircraft large airspace flight
CN107065571A (en) * 2017-06-06 2017-08-18 上海航天控制技术研究所 A kind of objects outside Earth soft landing Guidance and control method based on machine learning algorithm
CN110488852A (en) * 2019-08-28 2019-11-22 北京航空航天大学 A kind of hypersonic aircraft complete section surface self-adaption control method

Non-Patent Citations (9)

* Cited by examiner, † Cited by third party
Title
M. THILAGAVATHI等: "Cascade-Forward Neural Network in Identification of Plant Species of Desert Based on Wild Flowers", 《2018 IEEE INTERNATIONAL CONFERENCE ON SYSTEM, COMPUTATION, AUTOMATION AND NETWORKING (ICSCA)》 *
刘建斌等: "基于神经网络的导弹倾斜变增益控制", 《战术导弹技术》 *
张玄武等: "基于级联前向网络的翼型优化设计", 《浙江大学学报(工学版)》 *
彭博等: "滚转导弹解耦过载驾驶仪及其BP自适应调度法", 《固体火箭技术》 *
计明军等: "《高等学校物流工程与物流管理专业系列规划教材 预测与决策方法》", 31 August 2018, 大连海事大学出版社 *
许东等: "《基于MATLAB6.X的系统分析与设计 神经网络》", 30 September 1998, 西安电子科技大学出版社 *
郭庆等: "三轴稳定质量矩拦截器的末制导律设计", 《系统仿真学报》 *
陈辛等: "卫星姿态控制系统执行器的故障诊断方法研究", 《航空兵器》 *
黄祚继等: "《多源遥感数据目标地物的分类与优化》", 31 May 2017, 中国科学技术大学出版社 *

Similar Documents

Publication Publication Date Title
CN109725644B (en) Linear optimization control method for hypersonic aircraft
US6092919A (en) System and method for adaptive control of uncertain nonlinear processes
CN114281092B (en) Hypersonic aircraft coordination attitude control method based on sliding mode interference observer
CN111290421A (en) Hypersonic aircraft attitude control method considering input saturation
CN106444807A (en) Compound attitude control method of grid rudder and lateral jet
CN111367182A (en) Hypersonic aircraft anti-interference backstepping control method considering input limitation
CN110276144B (en) Online identification method for pneumatic parameters of vertical take-off and landing carrier
CN109164708B (en) Neural network self-adaptive fault-tolerant control method for hypersonic aircraft
CN107515612B (en) Elastic vibration suppression method based on side jet flow control
CN109062055A (en) A kind of Near Space Flying Vehicles control system based on Back-stepping robust adaptive dynamic surface
CN110488852A (en) A kind of hypersonic aircraft complete section surface self-adaption control method
CN107065544A (en) Hypersonic vehicle neural network control method based on angle of attack power function
CN110209192A (en) Fighter plane course augmentation control design method
CN107831653B (en) Hypersonic aircraft instruction tracking control method for inhibiting parameter perturbation
CN114611416B (en) LS-SVM modeling method for nonlinear unsteady aerodynamic characteristics of missile
CN112231835A (en) Thrust performance and deflection efficiency integrated vectoring nozzle outlet area optimization method
CN112198888A (en) Self-adaptive PID control method considering autonomous take-off and landing of unmanned aerial vehicle on motor-driven platform
CN116560232A (en) RBF neural network-based adaptive inversion control method for variant aircraft
CN114815878B (en) Hypersonic aircraft collaborative guidance method based on real-time optimization and deep learning
CN112558464A (en) Aircraft controller gain scheduling method adaptive to strong pneumatic nonlinearity
CN116699994A (en) Aircraft control method based on reinforcement learning
CN111176117A (en) Fuzzy self-adaptive elastic control method of unmanned helicopter
Reigelsperger et al. Robust control law design for lateral‐directional modes of an F‐16/MATV using μ‐synthesis and dynamic inversion
CN116482971A (en) Control method of high-maneuvering aircraft
CN114815625A (en) Intelligent learning control method for composite control aircraft

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication

Application publication date: 20210326

RJ01 Rejection of invention patent application after publication