CN113239788A - Mask R-CNN-based wireless communication modulation mode identification method - Google Patents
Mask R-CNN-based wireless communication modulation mode identification method Download PDFInfo
- Publication number
- CN113239788A CN113239788A CN202110510807.XA CN202110510807A CN113239788A CN 113239788 A CN113239788 A CN 113239788A CN 202110510807 A CN202110510807 A CN 202110510807A CN 113239788 A CN113239788 A CN 113239788A
- Authority
- CN
- China
- Prior art keywords
- mask
- cnn
- network
- neural network
- training
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Signal Processing (AREA)
- Digital Transmission Methods That Use Modulated Carrier Waves (AREA)
Abstract
The invention discloses a Mask R-CNN-based wireless communication modulation mode identification method, which comprises the following steps: s101: normalization processing of radio frequency signals in different modulation modes; s102: training a Mask R-CNN neural network for identifying a signal modulation mode; s103: and (5) optimizing a Mask R-CNN neural network for signal modulation pattern recognition. In the invention, signals which are mixed and superposed with different modulation models can be input into a Mask R-CNN neural network for training in a uniform dimension by introducing a normalization technology of radio frequency signals, and the modulation modes in the mixed signals are identified by fully utilizing the strong classification capability of the neural network through the training of the Mask R-CNN neural network.
Description
Technical Field
The invention relates to the field of signal modulation pattern recognition in intelligent communication, in particular to a Mask R-CNN neural network-based signal modulation pattern recognition method.
Background
With the development of wireless communication technology, the application range of wireless communication technology is becoming wider and wider, and wireless communication systems carry the task of transmitting information over long distances through wireless channels. Generally, in order to improve the efficiency and reliability of a communication system, different modulation patterns are used for wireless signals, i.e. the modulation signals are converted into modulated signals suitable for wireless channel transmission. In the fields of wireless communication, communication countermeasure and the like, only if the receiver correctly identifies the modulation mode of the transmitter signal, the receiver can perform subsequent work flows of demodulation, information processing, analysis and the like.
In practical application environments, radio frequency signals, channel models and noise of a communication system have unpredictable strong randomness, and in addition, the influence of channel fading, multipath propagation, interference and the like is added, and the identification of a signal modulation mode essentially solves the problem of mode classification with a plurality of unknown parameters. The traditional analog modulation modes such as amplitude modulation, single-side band and frequency modulation mostly adopt manual identification methods, are low in efficiency and poor in precision, and are difficult to be suitable for a wireless communication system in a current large-scale complex scene. As for the data signal modulation pattern recognition method, techniques of automatic recognition, such as feature-based statistical pattern recognition, recognition methods based on the maximum likelihood hypothesis, are currently being increasingly employed.
Currently, in the more flexible and open field of intelligent wireless communication, the above-mentioned traditional digital signal modulation pattern recognition technology based on statistics or traditional shallow machine learning no longer has advantages, mainly expressed in three aspects: firstly, it is difficult to adapt to the modulation mode in which a plurality of modulation identification modes confuse signals with each other; secondly, the method is difficult to adapt to signal modulation mode identification without prior information and with low signal-to-noise ratio; finally, the method is difficult to adapt to the identification of the modulation mode under the condition of signal distortion caused under the multipath channel environment, and therefore a signal modulation pattern identification method based on a Mask R-CNN neural network is provided.
Disclosure of Invention
The invention aims to provide a Mask R-CNN-based wireless communication modulation mode identification method, which solves the problem that the existing modulation mode which is difficult to adapt to various modulation identification modes and confuses signals with each other is solved; the method is difficult to adapt to the signal modulation mode identification without prior information and with low signal-to-noise ratio; it is difficult to adapt to the problem of identification of modulation modes in the case of signal distortions caused in a multipath channel environment.
In order to achieve the purpose, the invention provides the following technical scheme: a wireless communication modulation mode identification method based on Mask R-CNN comprises the following steps:
s101: normalization processing of radio frequency signals in different modulation modes;
s102: training a Mask R-CNN neural network for identifying a signal modulation mode;
s103: and (5) optimizing a Mask R-CNN neural network for signal modulation pattern recognition.
Preferably, the modulation in step S101 mainly matches the signal to be transmitted with a channel, so as to improve the transmission efficiency of the communication system; to accommodate different application domains, in particular different wireless channel media, transmitters often use different modulation models; different modulation modes are expressed in different dimensions or dimension units, and different radio-frequency signals are input into the Mask R-CNN neural network to be trained before being normalized by using data, so that the signals in different modulation modes can be input into the Mask R-CNN neural network to be trained in the same dimension unit.
Preferably, the different modulation modes include amplitude Shift Keying (Ask), Binary Phase Shift Keying (BPSK), Quadrature Phase Shift Keying (QPSK), and the like.
Preferably, the step S102 is different from the traditional neural networks such as CNN and Faster R-CNN, and the Mask R-CNN neural network realizes more accurate and fine-grained example segmentation by adding a Mask prediction score; in the convolutional layer, signal feature extraction is carried out, the signal feature extraction is carried out rapidly and accurately and consists of a residual error network ResNet and a feature Pyramid network FPN (feature Pyramid network); the extracted features are delivered to a Region generation Network (RPN), the RPN is a lightweight neural Network, the features are scanned by using a sliding window, a Region with a target is searched, the dimensions of candidate frames are unified through a RoI Align layer, and radio frequency signals are detected by using a full connection layer to obtain categories.
Preferably, the proper representation of the loss function in step S103 plays a crucial role in improving the accuracy of deep learning network training, learning and recognition, and a uniform loss function is defined for the RPN network loss, Fast R-CNN network loss and Mask segmentation network loss included in the training process of the Mask R-CNN, so that a uniform scale is provided in the training and optimizing processes of the Mask R-CNN network, and the training efficiency of the network is improved.
Compared with the prior art, the invention has the following beneficial effects:
in the invention, signals which are mixed and superposed with different modulation models can be input into a Mask R-CNN neural network for training in a uniform dimension by introducing a normalization technology of radio frequency signals, and the strong classification capability of the neural network is fully utilized by training the Mask R-CNN neural network to identify the modulation modes in the mixed signals; in addition, the accuracy of modulation mode identification and classification is further improved through an end-to-end optimization strategy based on the same loss function in a Mask R-CNN neural network.
Drawings
FIG. 1 is a schematic flow chart of the present invention;
FIG. 2 is a flow chart of multi-modulation signal pattern recognition in the present invention;
FIG. 3 is a schematic diagram of a Mask R-CNN neural network used in the present invention;
fig. 4 is a schematic diagram of a ResNet residual network structure used in the present invention.
Detailed Description
In order to make the technical solutions and advantages of the present invention more apparent, the following further detailed description of exemplary embodiments of the present invention is provided with reference to the accompanying drawings, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and are not exhaustive of all embodiments. And the embodiments and features of the embodiments described herein may be combined with each other without conflict.
The technical scheme of the invention is as shown in fig. 1-4, firstly, the radio frequency signals of different system modes are normalized and preprocessed, and are input into a Mask R-CNN deep neural network for training in a uniform dimension unit, and the training is stopped until a specific threshold value is met in the training process.
S101: normalization processing of radio frequency signals of different modulation modes
The purpose of the modulation is to match the signal to be transmitted with a channel and convert the baseband signal which is not beneficial to transmission into a high-frequency signal which is suitable for remote transmission and has anti-interference capability through the carrier, so that the communication system can efficiently transmit the signal. Therefore, in order to adapt to different wireless channel mediums and application environments, the transmitter uses different modulation models, such as ASK, BPSK, QPSK, etc.
Different modulation modes are expressed in different dimensions or dimension units, specifically, modulation information can be identified by characteristics such as instantaneous amplitude, instantaneous frequency, instantaneous phase and the like, the characteristics are reasonably selected to be the key for constructing a mode identification network, in the field of digital communication, a constellation diagram contains the structure and the state of a digital modulation signal and is an important characteristic capable of reflecting the signal modulation mode, and the relationship is fully utilized to construct the basis of efficient signal modulation identification classification.
Therefore, different radio frequency signals are input into the Mask R-CNN neural network before training, data standardization or normalization is needed, so that the signals under different modulation modes can be input into the Mask R-CNN neural network to be trained under the same dimension unit.
After the radio frequency signals of different modulation modes are subjected to normalization processing, the convergence speed and the modulation mode identification precision in the subsequent Mask R-CNN neural network training process can be improved.
S102: training of Mask R-CNN neural network for signal modulation pattern recognition
The Mask R-CNN neural network model is shown in FIG. 3, and frequency modulation pattern recognition is carried out through two stages. The training process of the Mask R-CNN neural network by using the radio frequency signal is as follows:
(1) normalization pretreatment: the input layer is a radio frequency signal gray image subjected to normalization processing.
(2) Feature extraction: convolution feature extraction and Feature Pyramid (FPN) fusion are realized in a convolution network of a first layer, specifically, convolution feature extraction is performed through a ResNet101, the ResNet101 network is shown in fig. 3, wherein the network is composed of a first convolution layer, a second convolution layer, a third convolution layer and a fifth convolution layer, and a sample is subjected to feature extraction in three steps of convolution, regularization and activation function at each layer to obtain feature maps of different scales, and the feature maps enter an FPN neural network. The FPN neural network utilizes the feature map to establish a feature map pyramid, and the feature map after fusion of all scales is obtained through operations such as convolution, pooling and fusion.
In the ResNet101 deep residual error network shown in FIG. 4, a cross-layer connection design is adopted, and the output of the network isF(x)+xWhereinF(x) And the residual error is 0 when the depth of the network layer number reaches a certain degree, so that the performance can not be degraded while the depth of the network layer number is ensured, and the problem of the depth and performance degradation of the ResNet101 network is balanced.
(3) And (3) generating a candidate frame: and the region generation network RPN carries out extraction operation of the candidate frame through foreground and background classification and frame regression, and generates a corresponding candidate frame.
The RPN realizes the selection of the candidate frames in a sliding window mode, generates candidate windows with different scales and different widths and heights at the position of each sliding window, extracts the characteristics of the corresponding candidate windows and is used for target classification and frame regression.
(4) The process of pooling: and performing pooling operation on a pooling layer (ROI Align), namely, corresponding pixels of the feature map to the original map, and then corresponding the feature map to the fixed features.
(5) And (3) ROI classification: the pooling result is divided into two parallel lines to enter a head network, frame regression and softmax multi-classification are achieved through a bounding box and classes respectively, a deconvolution method is used in mask score, the feature graph of the last layer is up-sampled, a heat map which is consistent with the size of the original image is obtained, namely category probability corresponding to the point is output at each position, and a mask which is the same as the class result is output.
In addition, in the training process of the Mask R-CNN neural network, as an improved pooling layer, the ROI Align layer adopts a bilinear difference method to obtain an accurate value of each position, so that errors caused by rough quantization operation of the traditional neural network are eliminated, and the accuracy of the extracted signal characteristics is ensured.
In the training process, according to different radio frequency signal data sets, Mask R-CNN network parameters are initialized to accelerate the training speed, then the training set is input into a Mask R-CNN neural network to be transmitted in a forward direction through the network, corresponding errors can be obtained in a first convolution layer, an RPN layer, an ROI Align layer, a second convolution layer and the like, then weights of the corresponding parameters are updated through gradient calculation by utilizing backward transmission until the errors are within a preset threshold range, and finally the training of the whole Mask R-CNN neural network is completed.
S103: adjusting and optimizing Mask R-CNN neural network for signal modulation mode identification
Theoretically, the reasonable loss function plays a vital role in improving the accuracy of deep learning network training, learning and recognition. Therefore, aiming at the RPN network loss, Fast R-CNN network loss and the loss function related to the Mask segmentation network in the process of training the radio frequency signal data set by the Mask R-CNN, the invention defines a uniform loss function representation, so that a uniform scale is provided for training and optimizing in the process of training and optimizing the Mask R-CNN network, and the training efficiency of the network is further improved.
The feature extraction network, the RPN neural network, the Fash R-CNN neural network and the Mask neural network can be integrated into a unified neural network by using a unified loss function, so that the end-to-end joint training and optimization process of the Mask R-CNN neural network is realized, and the unified training and optimization of the whole Mask R-CNN network are realized.
In addition, the pre-training can also obviously provide the performance of the neural network, so that before the actual data training of the radio frequency signals, a SigPreText training set with equal signal sample quantity and size of different systems can be adopted to carry out preliminary training on the Mask R-CNN neural network, and after a network model with the best classification effect on the SigPreText data set is obtained, the parameters in the network are used as initial parameters to be estimated in the actually-trained Mask R-CNN network.
After the Mask R-CNN neural network is trained, a Mask R-CNN model needs to be evaluated on a test data set by using a confusion matrix, and the accuracy of signal system identification is further improved through accuracy, precision, recall rate, cross-over ratio (IOU) and the like.
In conclusion, thanks to computer computing power and the development of artificial intelligence technology represented by deep learning, more efficient and accurate signal modulation pattern identification through deep learning is an important step towards intelligent communication. In the invention, the characteristics and the rules of the modulation model contained in the radio frequency signal sample can be autonomously found by utilizing the self-organizing and self-learning capabilities of the Mask R-CNN deep neural network, the method is a non-parametric model modulation mode identification method, and a series of complex processes such as modeling, parameter estimation, inspection, model reconstruction and the like in the traditional mode identification are omitted. In addition, the Mask R-CNN neural network is adopted to apply modulation mode identification, the radio frequency signal is not required to be marked manually, the neural network is used for automatically extracting the radio frequency signal characteristics, and classification identification is carried out.
The foregoing shows and describes the general principles and broad features of the present invention and advantages thereof. It will be understood by those skilled in the art that the present invention is not limited to the embodiments described above, which are given by way of illustration of the principles of the present invention, and that various changes and modifications may be made without departing from the spirit and scope of the invention as defined by the appended claims. The scope of the invention is defined by the appended claims and equivalents thereof.
Claims (5)
1. A method for identifying a wireless communication modulation mode based on Mask R-CNN is characterized by comprising the following steps:
s101: normalization processing of radio frequency signals in different modulation modes;
s102: training a Mask R-CNN neural network for identifying a signal modulation mode;
s103: and (5) optimizing a Mask R-CNN neural network for signal modulation pattern recognition.
2. The method for identifying the wireless communication modulation mode based on Mask R-CNN according to claim 1, wherein: in the step S101, the modulation mainly matches the signal to be transmitted with the channel, so as to improve the transmission efficiency of the communication system; to accommodate different application domains, in particular different wireless channel media, transmitters often use different modulation models; different modulation modes are expressed in different dimensions or dimension units, and different radio-frequency signals are input into the Mask R-CNN neural network to be trained before being normalized by using data, so that the signals in different modulation modes can be input into the Mask R-CNN neural network to be trained in the same dimension unit.
3. The method for identifying the wireless communication modulation mode based on Mask R-CNN as claimed in claim 2, wherein: the different modulation modes include amplitude Shift Keying (Ask), Binary Phase Shift Keying (BPSK), Quadrature Phase Shift Keying (QPSK), and the like.
4. The method for identifying the wireless communication modulation mode based on Mask R-CNN according to claim 1, wherein: different from the traditional neural networks such as CNN and Faster R-CNN in the step S102, the Mask R-CNN neural network realizes more accurate and fine-grained example segmentation by adding a Mask prediction score; in the convolutional layer, signal feature extraction is carried out, the signal feature extraction is carried out rapidly and accurately and consists of a residual error network ResNet and a feature Pyramid network FPN (feature Pyramid network); the extracted features are delivered to a Region generation Network (RPN), the RPN is a lightweight neural Network, the features are scanned by using a sliding window, a Region with a target is searched, the dimensions of candidate frames are unified through a RoI Align layer, and radio frequency signals are detected by using a full connection layer to obtain categories.
5. The method for identifying the wireless communication modulation mode based on Mask R-CNN according to claim 1, wherein: the proper representation of the loss function in the step S103 plays a crucial role in improving the accuracy of deep learning network training, learning and recognition, and a uniform loss function is defined aiming at the RPN network loss, Fast R-CNN network loss and Mask segmentation network loss contained in the training process of the Mask R-CNN, so that a uniform scale is provided in the training and optimizing processes of the Mask R-CNN network, and the training efficiency of the network is improved.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110510807.XA CN113239788A (en) | 2021-05-11 | 2021-05-11 | Mask R-CNN-based wireless communication modulation mode identification method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110510807.XA CN113239788A (en) | 2021-05-11 | 2021-05-11 | Mask R-CNN-based wireless communication modulation mode identification method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113239788A true CN113239788A (en) | 2021-08-10 |
Family
ID=77133324
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110510807.XA Pending CN113239788A (en) | 2021-05-11 | 2021-05-11 | Mask R-CNN-based wireless communication modulation mode identification method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113239788A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114553648A (en) * | 2022-01-26 | 2022-05-27 | 嘉兴学院 | Wireless communication modulation mode identification method based on space-time diagram convolutional neural network |
| CN116055270A (en) * | 2023-01-16 | 2023-05-02 | 中国科学院计算技术研究所 | Modulation recognition model, training method thereof and signal modulation processing method |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110163282A (en) * | 2019-05-22 | 2019-08-23 | 西安电子科技大学 | Modulation Mode Recognition method based on deep learning |
| CN111340784A (en) * | 2020-02-25 | 2020-06-26 | 安徽大学 | Image tampering detection method based on Mask R-CNN |
| CN111369540A (en) * | 2020-03-06 | 2020-07-03 | 西安电子科技大学 | Plant leaf disease identification method based on mask convolutional neural network |
| CN111695417A (en) * | 2020-04-30 | 2020-09-22 | 中国人民解放军空军工程大学 | Signal modulation pattern recognition method |
| CN112308133A (en) * | 2020-10-29 | 2021-02-02 | 成都明杰科技有限公司 | Modulation identification method based on convolutional neural network |
-
2021
- 2021-05-11 CN CN202110510807.XA patent/CN113239788A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110163282A (en) * | 2019-05-22 | 2019-08-23 | 西安电子科技大学 | Modulation Mode Recognition method based on deep learning |
| CN111340784A (en) * | 2020-02-25 | 2020-06-26 | 安徽大学 | Image tampering detection method based on Mask R-CNN |
| CN111369540A (en) * | 2020-03-06 | 2020-07-03 | 西安电子科技大学 | Plant leaf disease identification method based on mask convolutional neural network |
| CN111695417A (en) * | 2020-04-30 | 2020-09-22 | 中国人民解放军空军工程大学 | Signal modulation pattern recognition method |
| CN112308133A (en) * | 2020-10-29 | 2021-02-02 | 成都明杰科技有限公司 | Modulation identification method based on convolutional neural network |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114553648A (en) * | 2022-01-26 | 2022-05-27 | 嘉兴学院 | Wireless communication modulation mode identification method based on space-time diagram convolutional neural network |
| CN114553648B (en) * | 2022-01-26 | 2023-09-19 | 嘉兴学院 | Wireless communication modulation mode identification method based on space-time diagram convolutional neural network |
| CN116055270A (en) * | 2023-01-16 | 2023-05-02 | 中国科学院计算技术研究所 | Modulation recognition model, training method thereof and signal modulation processing method |
| CN116055270B (en) * | 2023-01-16 | 2024-06-11 | 中国科学院计算技术研究所 | Modulation recognition model, training method thereof and signal modulation processing method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110855591B (en) | QAM and PSK signal intra-class modulation classification method based on convolutional neural network structure | |
| CN110086737B (en) | Communication signal modulation mode identification method based on graph neural network | |
| CN108234370B (en) | Communication signal modulation mode identification method based on convolutional neural network | |
| CN109660297B (en) | Physical layer visible light communication method based on machine learning | |
| CN112702294B (en) | Modulation recognition method for multi-level feature extraction based on deep learning | |
| He et al. | Learning-based signal detection for MIMO systems with unknown noise statistics | |
| CN110113288A (en) | A kind of design and demodulation method of the ofdm demodulator based on machine learning | |
| CN113239788A (en) | Mask R-CNN-based wireless communication modulation mode identification method | |
| CN109120563B (en) | Modulation recognition method based on neural network integration | |
| Perenda et al. | Learning the unknown: Improving modulation classification performance in unseen scenarios | |
| CN113723556B (en) | Modulation mode identification method based on entropy weighting-multi-mode domain antagonistic neural network | |
| Lin et al. | A real-time modulation recognition system based on software-defined radio and multi-skip residual neural network | |
| CN110659684A (en) | Convolutional neural network-based STBC signal identification method | |
| Daldal et al. | Deep long short-term memory networks-based automatic recognition of six different digital modulation types under varying noise conditions | |
| Park et al. | Learning how to demodulate from few pilots via meta-learning | |
| Hao et al. | Automatic modulation classification via meta-learning | |
| Chen et al. | Automatic modulation classification using multi-scale convolutional neural network | |
| Almohamad et al. | Dual-determination of modulation types and signal-to-noise ratios using 2D-ASIQH features for next generation of wireless communication systems | |
| CN114398931A (en) | Modulation recognition method and system based on numerical characteristic and image characteristic fusion | |
| CN114595729A (en) | Communication signal modulation identification method based on residual error neural network and meta-learning fusion | |
| CN114615118A (en) | Modulation identification method based on multi-terminal convolution neural network | |
| Wang et al. | Residual learning based RF signal denoising | |
| Brandes et al. | RF waveform synthesis guided by deep reinforcement learning | |
| CN116070136A (en) | Multi-mode fusion wireless signal automatic modulation recognition method based on deep learning | |
| CN115409056A (en) | Automatic modulation identification method for large dynamic signal-to-noise ratio |
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 |