CN115173924A - Unmanned aerial vehicle unmanned inspection management system based on cloud edge cooperation technology - Google Patents
Unmanned aerial vehicle unmanned inspection management system based on cloud edge cooperation technology Download PDFInfo
- Publication number
- CN115173924A CN115173924A CN202210789586.9A CN202210789586A CN115173924A CN 115173924 A CN115173924 A CN 115173924A CN 202210789586 A CN202210789586 A CN 202210789586A CN 115173924 A CN115173924 A CN 115173924A
- Authority
- CN
- China
- Prior art keywords
- edge
- data center
- cloud
- data
- aerial vehicle
- 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
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B7/00—Radio transmission systems, i.e. using radiation field
- H04B7/14—Relay systems
- H04B7/15—Active relay systems
- H04B7/185—Space-based or airborne stations; Stations for satellite systems
- H04B7/18502—Airborne stations
- H04B7/18506—Communications with or from aircraft, i.e. aeronautical mobile service
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W16/00—Network planning, e.g. coverage or traffic planning tools; Network deployment, e.g. resource partitioning or cells structures
- H04W16/02—Resource partitioning among network components, e.g. reuse partitioning
- H04W16/10—Dynamic resource partitioning
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- Astronomy & Astrophysics (AREA)
- Aviation & Aerospace Engineering (AREA)
- General Physics & Mathematics (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention discloses an unmanned aerial vehicle unmanned inspection management system based on cloud edge cooperation technology, and particularly relates to the field of unmanned inspection, wherein the unmanned aerial vehicle unmanned inspection management system comprises an edge data center end for storing and processing data for unmanned aerial vehicle inspection; the aggregation cloud data center end distributes the control instruction and the model parameter; an uplink for transmitting the edge data to the central cloud; a downlink for transmitting control aggregation frequency commands and loss function coefficient data to the edge data center; the edge computing network based on the Federal learning applied in the real-time integrated big data center construction introduces a (D, B) -G (t) model to simulate the participation behavior of edge nodes in the network by considering the sample data resource limit condition of the edge nodes; through cloud-edge cooperation, timely response of edge intelligence is achieved, and the federal learning aggregation frequency is dynamically adjusted according to data resource conditions, so that effectiveness and learning quality of edge intelligent learning are maintained, and resource utilization rate is effectively improved.
Description
Technical Field
The invention relates to the technical field of unmanned inspection, in particular to an unmanned inspection management system of an unmanned aerial vehicle based on a cloud edge cooperation technology.
Background
The rapid growth of internet of things (IoT) and internet applications has led to an exponential growth in data generated at the edge of a network. It is predicted that in the near future, data generation rates will exceed the capacity of today's internet. Sending all data to the remote cloud is impractical and often unnecessary due to network bandwidth and data privacy concerns. Thus, research institutions estimate that over 90% of the data will be stored for local processing. Edge computing techniques enable local data storage and processing with global coordination. Where lightweight data centers or servers are deployed locally, local edge data center nodes work in concert with remote host data centers to perform large-scale distributed tasks involving local processing and remote coordination/execution.
The unmanned aerial vehicle meets various problems in engineering practice, particularly, an AI algorithm needs to be executed in real time in unmanned inspection application, the real-time performance of the AI calculation result is the main problem at present, and a new unmanned inspection mode is provided by the aid of edge calculation and particularly the introduction of cloud edge cooperation.
In a federated learning system, a federated network may be composed of a large number of devices. The efficiency of communication in the network has a large impact on overall speed. Therefore, it is important to research methods for improving data utilization efficiency. It is generally possible to proceed from reducing the transmission frequency and reducing the amount of information transmitted per round. Firstly, the transmission frequency is reduced, and the exchange times of the gradient between the client and the central server are mainly reduced, so that the local optimization times of the client in one global iteration can be properly increased. And secondly, the reduction of the information quantity is realized mainly by reducing the exchange times of the client and the central server, and therefore, proper gradient compression or quantization can be carried out. Based on comparing the synchronous and asynchronous aggregation methods of distributed gradient descent, some studies suggest that federal learning uses the synchronous method because it is more efficient than the asynchronous method. Many research methods use a fixed global aggregation frequency. It does not provide theoretical convergence guarantees, nor is the experiment performed in a network environment.
Further research has proposed a mechanism for secure global aggregation. For example, some studies have proposed methods for compressing the information exchanged in a global aggregation step. Some studies have tuned the standard gradient descent procedure in the federal environment to achieve better performance. Further research has proposed a method of sharing a small amount of data with other nodes to improve learning performance of non-i.i.d. data distribution. These studies do not take into account the adaptability of the global aggregation frequency, which in practical scenarios needs to take into account the adaptability to the local data resource-limited case. Current research has not investigated the adaptability of global aggregation frequency to resource-constrained federal learning.
Still other studies have studied the synchronization frequency optimization problem considering the run time for the data center setting, but have not considered the characteristics of the non-i.i.d. data distribution, especially the sample asymmetry problem of the non-i.i.d. data, which is essential in federal learning.
Still further, distributed machine learning across multiple data centers in different geographic locations is investigated, and a threshold-based approach is proposed to reduce communication between different data centers. While the work of these studies is related to the adaptation of synchronization frequency to resource considerations, it focuses on data centers that are point-to-point connected, as opposed to the federal learning architecture, which is not point-to-point. It also allows asynchrony between data center nodes, but this is not the case in federal learning. Furthermore, many of the methods under study are designed empirically, without considering specific theoretical goals, and without considering computational resource constraints, which are important in edge computing systems in addition to constrained communication resources.
Disclosure of Invention
In order to overcome the above defects in the prior art, embodiments of the present invention provide an unmanned aerial vehicle unmanned inspection management system based on cloud-edge coordination technology, so as to solve the problems mentioned in the background art.
In order to achieve the purpose, the invention provides the following technical scheme: the utility model provides an unmanned aerial vehicle management system that patrols and examines based on cloud limit collaborative technology, includes:
the edge data center end is used for an edge data center for sample data processing;
the cloud data center end performs resource allocation by using an algorithm according to the edge model parameters and the sample data;
the uplink is used for transmitting the edge data of the edge data center end to the cloud data center end;
the downlink is used for transmitting the resource aggregation frequency data and the model parameters of the cloud data center end to the edge data center;
wherein: the cloud data center end conducts an aggregation frequency adjustment process through an algorithm according to the model parameters and the resource data uploaded by the edge data center end;
the edge data center receives data, including training data Di of each node; and calculating a loss function fi (w) on each node according to each training sample j belonging to the Di training algorithm:
the cloud data center calculates a global aggregation loss function according to the local loss function of each node
The edge data center calculates the locally updated model parameters wi (t) by using a gradient descent method,
the cloud data center calculates the aggregation model parameter w (t),
the cloud data center calculates the gradient difference,
since F (w) is rho-Lipschitz, beta-smoothen
Wherein F (w) is a constant, then
Passing order in cloud data center
And solving t × = argminG (t) to obtain the aggregation frequency of the optimal learning algorithm under the condition of resource limitation.
The invention has the technical effects and advantages that:
the method is based on an edge computing network applying federal learning in the construction of an integrated big data center in reality, and introduces a (D, B) -G (t) model to simulate the participation behavior of edge nodes in the network by considering the sample data resource limit condition of the edge nodes; aiming at a large-scale edge computing network, an ARE algorithm and an end edge cloud cooperative framework ARE provided to solve the problem, so that the self-adaptive aggregation of the network on the federal learning is realized, and the performance of the federal learning is effectively improved; by dynamically selecting the aggregation frequency for each network in a self-adaptive manner under limited sample data resources, the applicability of the federal learning network can be greatly improved, namely, the network stability is improved, so that the problem that the unmanned aerial vehicle encounters data learning in unmanned inspection mode engineering practice is solved.
Drawings
FIG. 1 is a diagram of an overall system edge computing network topology of the present invention.
FIG. 2 is a schematic diagram of the ARE algorithm of the present invention.
FIG. 3 is a diagram of an edge data center according to the present invention.
FIG. 4 is a schematic flow chart of the ARE algorithm of the present invention.
FIG. 5 is a diagram of the loss function laboratory results of the present invention.
FIG. 6 is a diagram of the learning accuracy laboratory test result of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The utility model provides an unmanned aerial vehicle management system that patrols and examines based on cloud limit collaborative technology, includes:
the edge data center end is used for an edge data center for sample data processing;
the cloud data center end performs resource allocation by using an algorithm according to the edge model parameters and the sample data;
the uplink is used for transmitting the edge data of the edge data center end to the cloud data center end;
the downlink is used for transmitting the resource aggregation frequency data and the model parameters of the cloud data center end to the edge data center;
wherein: the cloud data center end conducts an aggregation frequency adjustment process through an algorithm according to the model parameters and the resource data uploaded by the edge data center end;
the edge data center receives data, including training data Di of each node; and calculating a loss function fi (w) on each node according to each training sample j epsilon Di training algorithm:
the cloud data center calculates a global aggregation loss function according to the local loss function of each node
The edge data center calculates the locally updated model parameters wi (t) by using a gradient descent method,
the cloud data center calculates the aggregation model parameter w (t),
the cloud data center calculates the gradient difference,
since F (w) is rho-Lipschitz, beta-smoothen
Wherein F (w) is a constant, then
Passing order in cloud data center
Solving t × = argmin G (t) to obtain the aggregation frequency of the optimal learning algorithm under the condition of resource limitation;
the invention aims to effectively solve the problem of performance improvement of federal learning under the limitation of sample data resources in edge computing, and provides a (D, B) -argmin G (t) problem and an end edge cloud cooperative frame and an ARE algorithm by considering the relation and the structural relationship of the edge data center under an integrated data center construction system and data space-time information from the aspects of adaptability and stability of training sample data resources of an edge node data center, so that the problem is effectively solved, and the overall stability of an edge computing federal learning network can be effectively maintained under the limitation of the sample data resources through verification, and the computing performance is maximized under the limitation of the sample data resources.
Fig. 4 is a schematic flow chart of a method for optimizing adaptive federal learning aggregation frequency based on an end edge cloud coordination framework in an unmanned aerial vehicle unmanned inspection management system based on a cloud edge coordination technology, which includes the following steps:
(1) Each edge node acquires data records of a data center covered by the edge node, wherein the data records comprise training sample data, network information and the like; preprocessing the data and extracting required effective data, wherein the extracted effective data information comprises the sample data size and network delay information processed by each edge node; processing all service data and network information of each edge node into data and a network sequence according to the time stamp sequence, wherein the pattern is Di = (Di (1), di (2), and.. Di (n)), bi = (Bi (1), bi (2), and.. Di (n)), wherein i is the edge node number, D is a sample data size record, and B is a network resource record;
(2) Each edge node calculates a corresponding loss function F (w) and a parameter w according to an edge learning algorithm model deployed by each edge node;
(3) Then calculating a loss function parameter Wi (t) for simulating the next iteration by using a gradient descent method;
(4) Calculating lipschitz parameters Li and smoothness parameters bi of Fi (w), and sending the Li, bi and Fi (w) to an aggregator;
(5) And the aggregator performs aggregation calculation according to the uploading parameters of the nodes to obtain aggregation parameters. Then calculating the gradient difference of each edge node, and then calculating the gradient difference after aggregation;
(6) Calculating an aggregation frequency value according to a (D, B) -argmin G (t) algorithm;
(7) The cloud allocates aggregation frequency to each edge node according to the aggregation strategy obtained by calculating argmin G (t), resources are issued to a base station corresponding to the network through a downlink, and then the resources are reliably shared to the whole network through anchor points, so that the dependence of the edge nodes on data resources and network resources is effectively reduced, and the utilization efficiency of the resources is improved;
finally, in order to verify the effectiveness of the ARE algorithm provided by the invention, the algorithm is subjected to a room test under four sample data conditions based on a lower edge computing environment, the degree of goodness of the algorithm is represented by a loss function value and computing precision, and the algorithm performance is compared by selecting different edge node total numbers; the 4 data sample resource cases are that each data sample is randomly allocated to a node, all data samples in each node have the same label, each node has the whole data set, and the data samples with the first half labels are distributed to the first half nodes (other samples are allocated to the second part nodes). The ARE algorithm dynamically acquires edge node data and network resource data in each edge data center, and then selects the aggregation frequency corresponding to (D, B) -argmin G (t) as the optimal aggregation frequency; in the invention, under the condition of presetting 5 nodes, the ARE algorithm is compared with a data centralized data center (figure 5), and the result shows that the performance of the ARE algorithm can be effectively improved; (FIG. 6) it was found that the algorithm of the present invention can achieve better results than the direct operation of the data center; in the invention, the learning precision change of the federal learning is checked through 4 kinds of edge node data conditions; finally, the invention compares the algorithm precision and the loss function under the condition of several data, and can see that the ARE algorithm has improved effect along with the change of the number of the nodes; the ARE method is an optimal method for solving the problem of self-adaptive aggregation frequency and improving the network stability and the federal learning reliability by integrating the experimental results and considering the performance and efficiency of the algorithm;
under the limitation of limited resources, based on a large-scale federal learning network, the invention enables the performance of edge nodes in all networks to be exerted to the maximum by dynamically aggregating the data uploaded by each edge node, namely the final federal learning performance is maximized, so as to adapt to the changes of network and data resource transmission; dynamic aggregation of aggregation frequency is realized through end edge cloud cooperation, and the federal learning performance under resource limitation is effectively improved;
aiming at the existing federated learning aggregation frequency optimization problem, the invention provides an aggregation frequency problem based on large-scale federated learning under the consideration of the edge sample data factors, all edge nodes ensure the maximum calculation performance by dynamically acquiring the resource data of each edge node, the overall stability of the federated learning system is maintained, and an ARE algorithm is provided to effectively solve the problem; meanwhile, dynamic acquisition of network resources and automatic adjustment of federal learning aggregation frequency are realized by combining the proposed end edge cloud cooperation model;
in order to solve the problem of adaptability of the federal learning edge node under the condition of resource limitation, the invention provides a self-adaptive algorithm which dynamically depends on the data of the node and the network resource condition; the invention relates to a side cloud cooperative federated learning network architecture, which realizes dynamic allocation of resources based on a gradient descent simulation algorithm under resource limitation; the edge base station collects and calculates the resource information of the federal learning sample data covered by the edge base station and sends the resource information to the cloud, and the cloud executes a gradient descent algorithm to formulate an aggregation frequency strategy so as to realize the performance optimization of the edge node under the resource limitation;
according to the invention, the optimization of the edge intelligent application performance under the condition of limited resources is ensured through the cooperation of computing and storing resources of the cloud and the edge, the reliability and the availability of the system are improved, and the high-efficiency computation of edge nodes is realized; the method is mainly based on an edge computing technology, a cloud edge cooperation technology and a data mining technology, and belongs to the field of cloud computing and edge computing; for federal learning in edge computing systems, research and discussion dynamically determines a global aggregation frequency to optimize a learning algorithm given data and communication resource budgets, thereby achieving maximized optimized edge node computing performance under data resource constrained conditions.
The points to be finally explained are: first, in the description of the present application, it should be noted that, unless otherwise specified and limited, the terms "mounted," "connected," and "connected" should be understood broadly, and may be a mechanical connection or an electrical connection, or a communication between two elements, and may be a direct connection, and "upper," "lower," "left," and "right" are only used to indicate a relative positional relationship, and when the absolute position of the object to be described is changed, the relative positional relationship may be changed;
secondly, the method comprises the following steps: in the drawings of the disclosed embodiments of the invention, only the structures related to the disclosed embodiments are referred to, other structures can refer to common designs, and the same embodiment and different embodiments of the invention can be combined with each other without conflict;
and finally: the above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that are within the spirit and principle of the present invention are intended to be included in the scope of the present invention.
Claims (7)
1. The utility model provides an unmanned aerial vehicle management system that patrols and examines based on cloud limit collaborative technology which characterized in that includes:
the edge data center end is used for an edge data center for sample data processing;
the cloud data center end performs resource allocation by using an algorithm according to the edge model parameters and the sample data;
the uplink is used for transmitting the edge data of the edge data center end to the cloud data center end;
the downlink is used for transmitting the resource aggregation frequency data and the model parameters of the cloud data center end to the edge data center;
wherein: and the cloud data center end performs an aggregation frequency adjustment process through an algorithm according to the model parameters and the resource data uploaded by the edge data center end.
2. The unmanned aerial vehicle unmanned inspection management system based on cloud edge cooperation technology according to claim 1, characterized in that: the edge data center receives data, including training data Di of each node; and calculating a loss function fi (w) on each node according to each training sample j epsilon Di training algorithm:
3. the unmanned aerial vehicle unmanned inspection management system based on cloud edge cooperation technology according to claim 2, characterized in that: and the cloud data center calculates a global aggregation loss function according to the local loss function of each node:
4. the unmanned aerial vehicle unmanned inspection management system based on cloud edge cooperation technology according to claim 3, characterized in that: the edge data center calculates the locally updated model parameters wi (t) by using a gradient descent method:
5. the unmanned aerial vehicle unmanned inspection management system based on cloud edge cooperation technology according to claim 4, characterized in that: and the cloud data center calculates the aggregation model parameter w (t):
6. the unmanned aerial vehicle unmanned inspection management system based on cloud edge cooperation technology according to claim 5, characterized in that: calculating the gradient difference through the cloud data center,
since F (w) is rho-Lipschitz, beta-smoothen
7. The unmanned aerial vehicle unmanned inspection management system based on cloud edge collaborative technology according to claim 6, characterized in that: wherein F (w) is a constant, then
Passing order in cloud data center
And solving t × = argminG (t), and obtaining the aggregation frequency of the optimal learning algorithm under the condition of resource limitation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210789586.9A CN115173924A (en) | 2022-07-06 | 2022-07-06 | Unmanned aerial vehicle unmanned inspection management system based on cloud edge cooperation technology |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210789586.9A CN115173924A (en) | 2022-07-06 | 2022-07-06 | Unmanned aerial vehicle unmanned inspection management system based on cloud edge cooperation technology |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115173924A true CN115173924A (en) | 2022-10-11 |
Family
ID=83491378
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210789586.9A Pending CN115173924A (en) | 2022-07-06 | 2022-07-06 | Unmanned aerial vehicle unmanned inspection management system based on cloud edge cooperation technology |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115173924A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112532451A (en) * | 2020-11-30 | 2021-03-19 | 安徽工业大学 | Layered federal learning method and device based on asynchronous communication, terminal equipment and storage medium |
| CN113010305A (en) * | 2021-02-08 | 2021-06-22 | 北京邮电大学 | Federal learning system deployed in edge computing network and learning method thereof |
| CN113971461A (en) * | 2021-10-26 | 2022-01-25 | 南京航空航天大学 | Distributed federal learning method and system for unmanned aerial vehicle ad hoc network |
| CN114357676A (en) * | 2021-12-15 | 2022-04-15 | 华南理工大学 | Aggregation frequency control method for hierarchical model training framework |
| CN114466309A (en) * | 2021-12-20 | 2022-05-10 | 上海科技大学 | Efficient communication wireless federal learning architecture construction method based on unmanned aerial vehicle |
-
2022
- 2022-07-06 CN CN202210789586.9A patent/CN115173924A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112532451A (en) * | 2020-11-30 | 2021-03-19 | 安徽工业大学 | Layered federal learning method and device based on asynchronous communication, terminal equipment and storage medium |
| CN113010305A (en) * | 2021-02-08 | 2021-06-22 | 北京邮电大学 | Federal learning system deployed in edge computing network and learning method thereof |
| CN113971461A (en) * | 2021-10-26 | 2022-01-25 | 南京航空航天大学 | Distributed federal learning method and system for unmanned aerial vehicle ad hoc network |
| CN114357676A (en) * | 2021-12-15 | 2022-04-15 | 华南理工大学 | Aggregation frequency control method for hierarchical model training framework |
| CN114466309A (en) * | 2021-12-20 | 2022-05-10 | 上海科技大学 | Efficient communication wireless federal learning architecture construction method based on unmanned aerial vehicle |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110191148B (en) | Statistical function distributed execution method and system for edge calculation | |
| Shan et al. | A survey on computation offloading for mobile edge computing information | |
| Almutairi et al. | Delay-optimal task offloading for UAV-enabled edge-cloud computing systems | |
| CN115278708B (en) | Mobile edge computing resource management method oriented to federal learning | |
| Huang et al. | Enabling dnn acceleration with data and model parallelization over ubiquitous end devices | |
| Yan et al. | Data offloading enabled by heterogeneous UAVs for IoT applications under uncertain environments | |
| CN116755867B (en) | Satellite cloud-oriented computing resource scheduling system, method and storage medium | |
| Ganjalizadeh et al. | Interplay between distributed AI workflow and URLLC | |
| CN111770133B (en) | Multi-body intelligent-connection cloud control platform | |
| CN115173924A (en) | Unmanned aerial vehicle unmanned inspection management system based on cloud edge cooperation technology | |
| Zheng et al. | Joint optimization of the deployment and resource allocation of UAVs in vehicular edge computing and networks | |
| CN116347522A (en) | Task unloading method and device based on approximate computation multiplexing under cloud edge cooperation | |
| CN116684472A (en) | Service deployment system and service deployment method for terminal-side computing network | |
| CN115665859A (en) | Multi-agent task unloading and resource scheduling combined optimization method | |
| Drainakis et al. | On the distribution of ML workloads to the network edge and beyond | |
| Duan et al. | Lightweight federated reinforcement learning for independent request scheduling in microgrids | |
| Han et al. | Task offloading strategies for cloud-side cooperation in compute-intensive scenarios based on edge computing | |
| CN111784029A (en) | Fog node resource allocation method | |
| CN116634388B (en) | Electric power fusion network-oriented big data edge caching and resource scheduling method and system | |
| CN113485718B (en) | Context-aware AIoT application program deployment method in edge cloud cooperative system | |
| CN117042051B (en) | Task unloading strategy generation method, system, equipment and medium in Internet of vehicles | |
| CN114928893B (en) | Architecture based on intelligent reflecting surface and task unloading method | |
| CN117724853B (en) | Data processing method and device based on artificial intelligence | |
| CN114531193B (en) | Bridge state monitoring method based on unmanned aerial vehicle cellular topology networking and mobile edge calculation | |
| CN117196014B (en) | Model training method and device based on federal learning, computer equipment and medium |
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 |