A VOIP simulation communication platform based on civil aviation control training business

Abstract

The present invention discloses a VOIP simulation communication platform based on civil aviation control training business. The simulation communication platform adopts a distributed communication service architecture to construct simulated wired and wireless communication channels. Voice media flow control can be realized for different channels. At the same time, different seat roles such as controllers and crews are set in the application scene of civil aviation control business training to simulate the real interactive communication business of aviation control process. In addition, the system supports the connection of voice communication interfaces of multiple simulation systems such as external PABX / PSTN and dedicated wired and wireless communication equipment of VOIP system, that is, the simulation communication platform can be configured according to specific needs and evolved into a real voice communication system, thereby improving the compatibility of the system.

Description

Technical Field

[0001] The invention relates to a simulation communication platform, in particular to a VOIP simulation communication platform based on civil aviation control training business. Background Art

[0002] In the past decade, with the rapid development of the domestic and international aviation transportation industry, air traffic has become increasingly busy, new navigation automation systems have been gradually implemented, and a large number of control equipment have been updated. The business level requirements for controllers have also become higher and higher, driving major civil aviation air traffic control bureaus and civil aviation colleges to improve their training level when delivering control talents and continuously enrich the skills of controllers. Traditional analog communication platforms only have simple voice communication functions, and most voice transmission systems are still based on analog (non-VOIP system). This channel-exclusive communication system is not suitable for high-concurrency or complex business application scenarios. Summary of the invention

[0003] Purpose of the invention: The technical problem to be solved by the present invention is to provide a VOIP simulation communication platform based on civil aviation control training business in view of the deficiencies of the prior art.

[0004] In order to solve the above technical problems, the present invention discloses a VOIP simulation communication platform based on civil aviation control training business, including: a dispatch service unit, a voice communication unit, a system operation and maintenance unit, a recording service unit, a VOIP radio gateway and a VOIP identification unit;

[0005] The dispatch service unit, the system operation and maintenance unit, the recording service unit, the VOIP radio gateway and the VOIP identification unit are interconnected through an IP network bus and exchange control instructions and data;

[0006] The scheduling service unit implements the management of each access user in the platform, the scheduling and processing of wired and wireless services based on the ED137 series standards, the management of wired and wireless media links, and the management of master-slave switching and data synchronization;

[0007] The voice communication unit acts as a human-computer interaction terminal, configures different roles as needed for user login, and supports wired and wireless service functions of civil aviation control services;

[0008] The system operation and maintenance unit is used for remote configuration of the platform parameters and reporting of information on the status of each node unit in the platform;

[0009] The recording service unit is used to realize the media recording storage of each terminal device during the civil aviation control training and the synchronous playback on demand during the training review;

[0010] The VOIP radio gateway is used to simulate a VOIP radio instance, and at the same time, constructs radio transceiver links of different frequencies through internal conference bridge processing, and the simulated crew seats and controller seats realize the ground-to-air communication interactive negotiation process by selecting the same frequency;

[0011] The VOIP recognition unit is used to recognize the voice of ground-to-air communication in real time, recognize the voice issued by the control seat and generate structured instructions, drive the simulated aircraft of the external digital tower, and perform an automated simulation process.

[0012] Beneficial effects:

[0013] The present invention not only provides a communication platform and channel link environment with agile deployment, complete functions and realistic effects for civil aviation control simulation training applications, but also provides the necessary basic support for controller training services, and improves the reliability and effectiveness of civil aviation control simulation training. At the same time, the present invention provides a design idea for the next generation of aviation control simulation training system based on virtual reality (VR) based on intelligent design of voice recognition of structured control instructions. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments, and the above and / or other advantages of the present invention will become more clear.

[0015] Figure 1 The figure is a logical principle diagram of the VOIP simulation communication platform based on civil aviation control business of the present invention.

[0016] Figure 2 It is a schematic diagram of the application of the VOIP simulation communication platform based on civil aviation control business of the present invention.

[0017] Figure 3 Schematic diagram of the platform architecture of the present invention.

[0018] Figure 4 The figure is a schematic diagram of the worker thread design of the scheduling service unit of the present invention.

[0019] Figure 5 The logical architecture diagram of the recording service unit of the present invention. DETAILED DESCRIPTION

[0020] The VOIP simulation communication platform of the present invention provides rich voice support for civil aviation control simulation training business. With the help of complete voice services, recording services, recognition services and operation and maintenance services, it not only builds a basic communication platform for ground-to-air (including ground-to-ground) negotiation and interaction, but also realizes intelligent driving of simulated aircraft in the tower view based on voice recognition technology, providing a display platform for the training effect of control trainees who are undergoing on-the-job training.

[0021] Speech recognition technology based on aviation control instructions is a novel and key technology in the current simulation training field. Its main difficulties lie in the recognition engine and control semantic parsing technology. The speech recognition engine mainly includes: 1) Data preprocessing module, which performs preprocessing such as filtering (VAD processing) and framing on the received RTP data stream; 2) Speech feature extraction module, which mainly calculates the Mel-frequency cepstral coefficient (MFCC) of the signal to convert the waveform signal of each frame into a multi-dimensional feature vector (feature sequence) containing sound information; 3) Acoustic Model (AM), which is obtained by training a large amount of speech data, and the feature sequence can be converted into a phoneme sequence through this model; 4) Dictionary: The dictionary is a mapping set between a phoneme set and a character or word; 5) Language model (Language Model, LM), the language model is a model obtained by training a large amount of text information, which provides the prior probability that characters or words are associated with each other to form a sequence; 6) Decoder: decode the feature sequence (MFCC) of the input sound through the acoustic model, dictionary and language model, and use the Viterbi algorithm to solve the word sequence corresponding to the feature sequence with the maximum posterior probability, that is, recognize the output text information; 7) Control instruction parsing service: the parsed text is semantically analyzed to obtain structured control instructions, and the process includes sentence segmentation, part-of-speech tagging, control term matching, structured instructions, etc.

[0022] The technical problem to be solved by the present invention is how to build a basic communication platform with a distributed architecture, support the construction of simulated wired and wireless communication channels, realize voice media flow control for different channels, and set different seat roles such as controllers and crew members in the civil aviation control business training application scenario, simulate the real aviation control process interactive communication business, support civil aviation control simulation training business, and improve the control skills of trainees.

[0023] The present invention provides a VOIP (Voice Over Internet Protocol, voice transmission based on IP) analog communication platform applied to civil aviation control business, adopting a distributed communication service architecture, the platform includes a dispatch service unit 1, a voice communication unit 2, a system operation and maintenance unit 3, a recording service unit 4, a VOIP radio gateway 5 and a VOIP identification unit 6;

[0024] Among them, except for the voice communication unit 2 deployed and running on a terminal device based on the Windows operating system, the other platform components are deployed on a workstation, and the workstation operating environment is the CentOS 7.4 operating system;

[0025] The dispatch service unit 1, the voice communication unit 2, the system operation and maintenance unit 3, the recording service unit 4, the VOIP radio gateway 5 and the VOIP identification unit 6 are interconnected and communicate with each other through the IP network to exchange control instructions and data;

[0026] The dispatch service unit 1 is used to implement the management of various types of access users in the system, the dispatch processing of wired and wireless services based on the ED137 series standards, the management of wired and wireless media links, the master-slave switching, the data synchronization management and other functions;

[0027] The voice communication unit 2 is the main human-computer interaction terminal, which can be configured with different roles for user login as needed to support the wired and wireless service functions of civil aviation control. The main user roles include crew members and control personnel, etc. The crew members and control personnel mainly realize ground-air negotiation through simulated wireless channels, while control personnel realize ground-ground resource sharing and service interaction through wired channels;

[0028] Different roles have different configuration priorities, and different wired and wireless resources can be allocated as needed;

[0029] The system operation and maintenance unit 3 is used for remote configuration of system parameters and reporting of information on the status of each node unit of the system. Remote configuration of system parameters is mainly achieved through a database service based on a B / S architecture. Users access and modify data by entering the IP address of the database service and user authentication;

[0030] The recording service unit 4 is used to realize the media recording storage of each terminal device during the civil aviation control training and the on-demand synchronous playback function during the training review;

[0031] The VOIP radio gateway 5 is used to simulate various VOIP radio instances, and at the same time, construct radio transceiver links of different frequencies through internal conference bridge processing, and the simulated crew seats and controller seats realize the ground-to-air communication interactive negotiation process by selecting the same frequency;

[0032] The VOIP recognition unit 6 is used to recognize the voice of ground-to-air communication in real time, recognize the voice issued by the control seat and generate structured instructions, drive the simulated aircraft of the external digital tower, and perform the automated simulation process including wheel block removal, start, push out, taxi, takeoff, climb, acceleration and deceleration, etc.

[0033] The scheduling service unit (1) comprises a user management module 11, a session management module 12, a media management module 13 and a master-slave management module 14.

[0034] The user management module 11 manages various types of access users in the system and maintains information such as user name, communication address (IP and port number), user type, enabled state, busy state, heartbeat timeout (expire time), timestamp, etc. of each user instance.

[0035] Various types of terminal devices send registration (REGISTER) and heartbeat (KEEP ALIVE) information to the scheduling service unit 1 through the standard SIP protocol. The user management module 11 numbers the terminal device as a user instance according to the registration message and adds it to the system's internal user table, and maintains the user's status in real time through the heartbeat, that is, the user management module 11 calculates the interval between two consecutive timestamps. If it exceeds the user's expiration time, it is considered that the user connection is abnormal. If the connection is abnormal for three consecutive times, it is considered that the user is offline.

[0036] Until the terminal device initiates registration again, the user management module 11 sets the user instance status to online again;

[0037] The session management module 12 is used to implement the service interaction logic control with the external terminal device, mainly including the service logic control of the wired function and the service logic control of the wireless service, so that the system has complete wired and wireless communication capabilities;

[0038] The media management module 13 extracts relevant basic information (source and destination addresses, methods, topics, etc.) and media parameter information (Session Description Protocol, SDP) according to the interactive content of the session management module 12, builds a conference bridge list, and realizes the creation, fusion and destruction of the underlying multimedia link.

[0039] The conference bridge list is a key data structure in the scheduling service unit 1. The main parameters of the conference bridge include:

[0040] 1) Session ID (session-id), which is a unique identifier associated with the current SIP session;

[0041] 2) Conference number (group-id), one conference number is associated with multiple media numbers. Similar to the session ID, the conference number is a unique identifier for the media management module (13) to identify a session. The conference number is a 2-byte short data type;

[0042] 3) Member list: the member list of a common session is the calling number (caller) and the called number (callee); the member list of a multi-party conference is the sponsor number (sponsor), conference number (conference-id), and member list (member); the member list of a scheduling group is the calling number (caller), the answering number (callee), and the scheduling group number;

[0043] 4) Media number (media-id) list. The media number corresponds to the conference member. Each terminal instance that establishes a call connection is assigned a media number. The media number is a 4-byte integer data. The first 2 bytes are the conference number, and the last 2 bytes are the position number of the conference member in the current conference list.

[0044] 5) Media information: a memory is allocated for each media number (media-id) to store the corresponding media information, including the local port number of the scheduling service unit 1, the IP address of the conference member terminal, the port number of the conference member terminal, the current media status of the conference member (idle state, ready state, in conference, kicked out of conference state), etc.

[0045] The media management module 13 implements a dynamic allocation strategy for resources such as the conference number, media number, and media port number of the conference bridge. When a session ends, the resources will be recycled and used for subsequent new conferences.

[0046] For ordinary point-to-point communication applications, the media management module 13 receives the RTP stream from the terminal UA and forwards the RTP stream to the terminal UB in the conference bridge through the media conference bridge. The reverse is also true. The media management module 13 receives the RTP stream from the terminal UB and forwards the RTP stream to the terminal UA in the conference bridge through the media conference bridge to achieve real-time communication between the terminal UA and the terminal UB in the conference bridge.

[0047] In complex voice communication applications, if the number of members in the media conference bridge exceeds two, when multiple RTP streams are received at the same time, they need to be distributed to each member in the conference bridge after stream alignment and stream fusion processing.

[0048] The media management module 13 constructs a media conference bridge including multiple members, and implements the mixing and fusion processing of multiple media in the conference bridge. The method used for mixing and fusion processing is the table lookup method to realize PCM digital voice superposition, and its principle is to linearize and superimpose the streaming media of each RTP link and then delinearize it to obtain the output mixed media stream.

[0049] The media management module 13 will perform the sending and receiving interaction of media streams with the conference bridge members, i.e., the external media terminals (voice communication unit 2, VOIP radio gateway 6, etc.) according to the media information of the conference bridge;

[0050] The master-slave management module 14 is used to implement the state switching of the master and standby servers and the business data synchronization function. First, the state machine is used to realize the master and standby status confirmation and control of the dual services, and then the business data synchronization flow is determined according to the master-standby relationship. Finally, the real-time synchronization of various business data is realized through a distributed synchronization method. The function of the master-standby management module 14 is to ensure that the standby service module of the scheduling service unit 1 can achieve 0-second switching when the main service module of the scheduling service unit 1 in the running process is due to software abnormality or hardware failure, without affecting the normal voice services of the terminals connected to it;

[0051] The media management module 13 can implement functions such as simulated noise superposition of wired and wireless links and simulated communication link discontinuity through database configuration according to the needs of simulated communication services. Simulated noise superposition is to randomly superimpose gain on each frame of RTP stream (20ms per frame), and the distribution of gain is Gaussian white noise; simulated link discontinuity is to cover the RTP stream with a fixed period (the period is configured by the database) of silent packets;

[0052] The voice communication unit 2 includes a signaling interaction module 21, a media transceiver module 22 and a status management module 23. The signaling interaction module 21 interacts with the scheduling service unit 1 using standard SIP messages to complete authentication registration, wired service and wireless service interaction; the media transceiver module 22 is used to implement the RTP stream exchange process between the local RTP stream and the media management module 13; the status management module 23 is used to implement the status management of all terminal number buttons configured in the maintenance interface, and the button status includes idle, ringing, talking, holding and monitoring, etc.

[0053] When the voice communication unit 2 logs in, it can select a corresponding role from the role list (the role configuration is pre-configured in the database configuration module 31) to log in, so that the seat has different application permissions (the permissions are divided into super administrator, teacher seat, student seat, etc.);

[0054] The system operation and maintenance unit 3 includes a database configuration module 31 and a status monitoring module 32;

[0055] The database configuration module 31 is used to configure SIP (Session Initial Protocol) accounts, VCU (Voice Communication Unit) information, user numbers, radio station frequencies, role information, etc., and to load and update parameters in the scheduling service unit 1. The database configuration module 31 adopts a B / S architecture design, which supports users to remotely access and control the parameters configured by the server through a WEB page. While loading parameters, the master and backup synchronization of the database can be achieved.

[0056] When designing the database configuration module 31, the impact domain of parameter modification is considered, that is, before issuing the update notification, the role affected by the modified parameters is first confirmed, and then the updated notification message is broadcast only to the voice communication unit 2 that selected the role when logging in. For example, the user changes the telephone key number 025xxxx8860 of the "controller" role to 025xxxx8861 through the database configuration module 31. When clicking the "Submit Modification" button, the system detects that the voice communication unit 2 that selected the "controller" role is seat 1 and seat 3, and then sends the "database parameter modification" notification message to seat 1 and seat 3. A "parameter update" notification appears on the interface of seat 1 and seat 3, and click to load the database parameters. The advantage of this design is that it ensures the accuracy of parameter configuration and modification;

[0057] The state monitoring module 32 is used to realize the state collection, wired and wireless communication records, operation logs, abnormal alarms of various devices in the system including the dispatch service unit 1, the voice communication unit 2, the recording service unit 4, the VOIP radio gateway 5, and the VOIP identification unit 6, and supports user access and viewing;

[0058] The recording service unit 4 adopts a three-layer architecture of "low coupling and high cohesion", including a human-computer interaction module 41, a business service module 42 and an interface management module 43;

[0059] Among them, the human-computer interaction module 41 belongs to the presentation layer, which is used for displaying the parameters and status of each channel, recording and playback operations, recording the current memory status, and disk usage management; the business service module 42 belongs to the business logic layer, which is used to implement services such as voice recording, local playback, remote playback, disk alarm processing, record file deletion and transfer; the interface management module 43 belongs to the data access layer, which is used to support interaction with analog recording cards and IP devices, receive PCM (Pulse Code Modulation) data or send RTP (Real-time Transport Protocol) streams, and implement the underlying operations of recording and playback;

[0060] The local and remote playback processes of the recording service unit 4 adopt the ping-pong cache method to realize the pipeline process of reading files, PCM decoding, pushing RTP streams, etc., which improves the efficiency and performance of file reading and playback services;

[0061] The data VOIP radio gateway 5 complies with the ED137 interface series issued by EUROCAE, and is used to simulate ground-to-air communication transceiver channels of different frequencies to support communication training services for controllers and crew members;

[0062] The VOIP radio station gateway 5 includes a radio station access module 51, a conference proxy module 52 and a parameter configuration module 53;

[0063] The radio access module 51 is used to process the interaction with the external real IP radio station to realize the transmission of the sending and receiving bidirectional RTP link. In the application scenario of civil aviation control business training, the radio access module 52 self-loops the receiving link and the sending link belonging to the same frequency, that is, the Rx link of a certain frequency of the VOIP radio gateway 5 receives the RTP stream sent from the outside, searches for the Tx link of the same frequency inside the gateway, and sends the RTP stream out through the Tx link;

[0064] The conference proxy module 52 acts as an intermediate proxy for the radio access module 51 and the media management module 13, and uses a "back-to-back" mode to maintain communication with the radio access module 51 and the media management module 13, so as to realize the scheduling and exchange of signaling and media of the voice communication link;

[0065] After the conference proxy module 52 establishes a conference with the media management module 13, after receiving the RTP stream from the media management module 13, it forwards it to the corresponding receiving channel (according to the signaling scheduling rules) of the radio access module 51. Similarly, after the conference proxy module 52 receives the RTP stream of the transmission channel of the radio access module 51, it forwards it to the corresponding receiving channel (according to the signaling scheduling rules) of the media management module 13;

[0066] The parameter configuration module 52 is used to configure the parameters of each transceiver link of the VOIP radio gateway 5, including channel number, channel frequency, channel attributes (local side tone, remote side tone, closed side tone, etc.), media delay, media additional noise, media additional on / off, etc.;

[0067] The VOIP recognition unit 6 includes a recognition engine module 61 , a semantic analysis module 62 and a recognition processing module 63 .

[0068] Among them, the recognition engine module 61 adopts the popular convolutional neural network (CNN) and continuous temporal classification (CTC) methods based on deep learning theory, uses a large amount of structured air traffic control instruction voice data for training, builds an acoustic model and a language model, and can convert the input RTP streaming media into output text information;

[0069] The recognition engine module 61 is composed of signal processing, acoustic model, language model, dictionary and decoder modules. The acoustic model is obtained by training a large amount of speech data in the corpus, and the language model is obtained by training a large amount of text information to obtain the probability of correlation between single characters or words;

[0070] The semantic analysis module 62 further analyzes the text output by the recognition engine module 61, parses the meaningless recognition text into a sequence containing several structured terms, and then matches it with the control instruction dictionary (established in advance), and finally extracts the instruction sequence with control meaning;

[0071] The recognition processing module 63 performs corresponding branch processing according to the obtained control instructions, interacts with the simulated aircraft of the external digital tower, and issues instructions for controlling the aircraft including unblocking, starting, pushing out, taxiing, taking off, climbing, accelerating and decelerating, so that the simulated aircraft can perform corresponding actions according to the voice requirements of the controller;

[0072] The system performs the following steps:

[0073] Step 1: The system is turned on and the parameters of each software unit are initialized;

[0074] Step 2, the trainee in the control seat interacts with the captain seat through a simulated communication link;

[0075] Step 3: The system recognizes the student's voice and drives the aircraft model to move according to the instructions;

[0076] Step 4: After the control training is completed, the training process is evaluated and reviewed through the playback function;

[0077] Wherein, step 1 comprises the following steps:

[0078] Step 1-1, the simulation training server is turned on, and the system database configuration module 31 implements parameter initialization of the scheduling service unit 1, the recording service unit 4 and the VOIP radio gateway 5 according to MySQL and the local configuration file;

[0079] Step 1-2, the voice communication unit 2 is turned on, the control seat / pilot seat enters the user name and password, and selects the configured user role, logs in to the human-computer interaction interface of the voice communication unit 2, and the four accounts of the voice communication unit 2 are connected to the dispatch service unit 1;

[0080] Step 1-3, the voice communication unit 2, the system operation and maintenance unit 3, the recording service unit 4, the VOIP radio gateway 5 and the VOIP identification unit 6 automatically access the scheduling service unit 1 as a user instance, and maintain the keep-alive heartbeat. When the heartbeat is lost, the abnormal state will be uploaded to the system operation and maintenance unit 3 for display;

[0081] Step 1-4, several simulated radio accounts configured by the system database configuration module 31 establish communication with the dispatch service unit 1, forming a two-way media link (silent package) of several frequencies, and the corresponding radio button on the interface of the voice communication unit 2 is displayed as available;

[0082] Step 2 includes the following steps:

[0083] Step 2-1, the control seat / pilot seat respectively clicks on the wireless display area of ​​the voice communication unit 2 to select a receiving radio station / transmitting radio station of a certain frequency (118.125Hz), and the two establish a communication link through the scheduling service unit 1 and the VOIP radio gateway 5, and the link RTP stream is initially a silent packet;

[0084] Step 2-2, according to the requirements of the control training manual, the control trainee presses PTT (Push-To-Talk) to issue flight control instructions. After the captain understands the control instructions, the flight trainee presses PTT to repeat the instructions and perform corresponding control actions;

[0085] Step 2-3, the voice communication unit 2 collects the voice data of the control trainee's handle, performs A / D conversion and other related processing, generates RTP data, and sends it to the media management module 13;

[0086] Step 2-4, the media management module 13 searches for relevant members in the current conference bridge, performs time alignment and de-jitter processing on the received RTP stream, and then sends it to the radio receiving channel of the corresponding frequency of the VOIP radio gateway 5;

[0087] Step 2-5, after receiving the RTP stream, the VOIP radio gateway 5 sends the RTP stream to the media management module 13 through the closed-loop link via the transmission channel of the same frequency of the radio station;

[0088] Step 2-6, the media management module 13 searches for relevant members in the current conference bridge, performs time alignment and de-jitter processing on the received RTP stream, and then sends it to the captain seat of the selected frequency transmitting station;

[0089] Step 2-7, the captain seat plays the received RTP stream, which comes from the control trainee seat, and realizes the control command from the control seat to the captain seat. The reverse link is similar, and the captain seat can repeat the control instructions to the control seat for playback;

[0090] Step 2-8, when the PTT is pressed at the control seat / pilot seat, the recording service unit 4 and the voice communication unit 2 establish a recording channel link, record the RTP stream data sent by the voice communication unit 2 in real time, convert the RTP stream into a PCM (or WAV) file for storage, and record the file index and file detailed information in the database;

[0091] Step 2-9, the recording service unit 4 regularly scans the disk space, and when the space usage rate reaches the set first-level threshold, it sends an alarm status information to the monitoring; and when the second-level threshold is reached, the historical file is automatically deleted;

[0092] Step 3 includes the following steps:

[0093] Step 3-1, when the control seat or the captain seat presses PTT, the voice communication unit processes the sound signal collected by the handle through A / D conversion and other processes to obtain an RTP stream, which is then transmitted to the VOIP recognition unit 6;

[0094] Step 3-2, the recognition engine module 61 of the VOIP recognition unit 6 extracts the Mel Frequency Cepstrum Coefficient (MFCC) of the RTP stream signal as a feature sequence, and the feature sequence is used as an input of the VOIP decoding process;

[0095] Step 3-3, with the help of an acoustic model based on CNN (Convolutional Neural Networks) + CTC (Connectionist Temporal Classification) algorithm, a dictionary, and a language model based on N-gram, the decoder of the recognition engine module 61 uses the Viterbi algorithm to solve the Chinese character sequence with the maximum a posteriori probability corresponding to the current input feature sequence;

[0096] Step 3-4: semantically analyze the parsed text to obtain structured control instructions, including sentence segmentation, part-of-speech tagging, control term matching, and structured instructions;

[0097] Step 3-5, the recognition processing module 63 performs corresponding branch processing according to the control instructions obtained by recognition, interacts with the simulated aircraft of the external digital tower, and issues instructions for controlling the aircraft including unblocking, starting, pushing out, taxiing, taking off, climbing, accelerating and decelerating, so that the simulated aircraft can perform corresponding actions according to the voice requirements of the controller;

[0098] Step 4 includes the following steps:

[0099] Step 4-1, after the control training is completed, the training officer selects the training subject, training start and end time, control trainee seat, captain seat and other parameters on the training control seat interface, initiates scene playback to the digital tower, and sends the parameters to the recording service unit 4;

[0100] Step 4-2, the recording service unit 4 searches the database and searches for the locally stored audio file (PCM / WAV) and its storage location by using parameters such as the start and end time, the control trainee seat, and the captain seat;

[0101] Step 4-3, the recording service unit 4 establishes a unidirectional communication link with the control trainee seat and the captain seat, that is, the RTP stream is unidirectionally transmitted from the recording service unit 4 to the control trainee seat and the captain seat for playback, which is synchronized with the tower scene;

[0102] Step 4-4, the recording service unit 4 locally establishes a "ping-pong cache" and reads RTP streaming media data for a certain period of time (the length of time can be set) from the local disk;

[0103] Step 4-5, "ping cache" starts playback, and sends RTP stream to the control trainee seat and the captain seat for playback at a fixed time (20ms period). The RTP stream playback is synchronized with the digital tower scene playback in time;

[0104] Step 4-6: Ping Cache ends playback and starts to read RTP stream from local disk. Meanwhile, Pong Cache starts playback and sends RTP stream to the trainee and pilot seats for playback at a fixed time (20ms period).

[0105] Step 4-7, repeating steps 4-5 and 4-6 until all the speech data in the training process are played back, and the recording service unit (4) releases the current "ping-pong buffer".

[0106] The application software described in the present invention includes dispatch service software, voice communication unit software, system monitoring software, database management software, analog radio gateway software, control voice recognition software, etc., which are responsible for the system's core communication exchange processing, voice communication terminal access and control, master-slave switching and data synchronization, analog terminal radio communication, voice recognition based on control services, equipment status collection and reporting, message generation, routing selection control and management and maintenance.

[0107] The system platform of the present invention is suitable for application scenarios of simulated training and assessment of air traffic control trainees. It is currently the first domestic air traffic control simulation communication platform with control voice recognition. The platform has distributed dual redundant backup capabilities, supports complex functional requirements such as instant communication, review and commentary, and real-time recognition, and provides a function-rich, agilely deployed communication platform support for the cultivation, training and assessment of civil aviation control talents.

[0108] The present invention is proposed for the first time based on the application requirements of users, and currently there is no product with the same or similar functions on the market.

[0109] In combination with the business needs, business characteristics, existing communication resources and technologies of the simulated communication platform in civil aviation control training for various users at home and abroad, the present invention has developed a design of a VOIP simulated communication platform with control voice recognition capability for civil aviation control training business. The system adopts a distributed flexible architecture design and an agile and flexible deployment method, integrating core dispatching services, media control, recording and playback services, voice recognition services, system operation and maintenance capabilities, and can delete and expand the above capabilities as needed, providing an instant, effective, efficient and intelligent basic communication platform for complex civil aviation control business.

[0110] The VOIP (Voice over Internet Protocol, abbreviated as VOIP) simulation communication system based on civil aviation control services of the present invention can not only realize wired and wireless voice communication services, but also simulate a real ground-to-air interactive communication environment in complex application scenarios by superimposing interference noise on the media, intermittent voice and other technologies, and at the same time use deep learning and speech recognition and other technologies to achieve the automation of civil aviation control services, greatly improving the intelligent level of simulation training.

[0111] Example:

[0112] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0113] Combine the following Figure 1 , Figure 2 , Figure 3 and Figure 4 The present invention is described in further detail.

[0114] like Figure 3 As shown, the VOIP simulation communication platform based on civil aviation control business of the present invention includes a dispatch service unit 1, a voice communication unit 2, a system operation and maintenance unit 3, a recording service unit 4, a VOIP radio gateway 5 and a VOIP identification unit 6. Except that the voice communication unit 2 is deployed and operated on a terminal device based on the Windows operating system, the remaining platform components are deployed on a workstation, and the workstation is deployed with the CentOS7.4 operating system. The units are interconnected through an IP network and exchange control instructions and data.

[0115] The dispatch service unit 1 is the core control and processing unit of the entire VOIP analog communication platform. According to the system function, the dispatch service unit 1 can be divided into four functional modules: user management 11, conference management 12, media management 13, and master and backup management 14. According to the service function, the dispatch service unit 1 is mainly divided into two service modules: signaling dispatch (SIP) and media management (RTP) (corresponding to Figure 1The conference management 12 and the media management 13) work together to provide support services for ground-to-air and ground-to-ground voice communications.

[0116] The signaling dispatching mainly provides transit dispatching services for dispatching and connecting wired and wireless communication services between various terminal devices (including voice communication units (1), VOIP radio gateways (5), etc.).

[0117] The media service mainly receives the RTP stream sent by each terminal, and implements media-related business services within the conference bridge, including integration (such as multi-party conference requirements, radio conferences, etc.), RTP stream exchange, distribution, stream competition, etc.

[0118] Take the simplest communication between two parties (UA and UB) as an example.

[0119] 1) As the initiator of the session, the UA sends a session request (INVITE) message to the conference management 12. The message contains SDP information, namely, the IP address of the UA terminal, the port number of the RTP receiving port, the media encoding method (PCMA, PCMU, G729, etc.), etc.;

[0120] 2) Conference management 12 replies a status response of 100 TRYING to UA. At the same time, a conference bridge is established (the members in the bridge are UAs), the media information of the INVITE message is extracted, and a media instance PmA is established for UA in the server through media management 13, including allocating a media number, RTP local port number, and storing the IP address and remote RTP port number of UA;

[0121] 3) Media management 13 establishes an incomplete media instance PmB for the called party UB, including allocating media number, RTP local port number, reserving UB's IP address and remote RTP port number. At the same time, conference management 12 initiates a session request (INVITE) to the called party UB, and the message contains SDP information, that is, the local port number in the UB media instance, the IP address of the scheduling service unit 1, and the media encoding method are consistent with the request (INVITE) message sent by UA;

[0122] 4) After receiving the session request (INVITE) message, UB replies with 100TRYING and 180Ringing responses in succession, and sends an incoming ring. After the attendant picks up the phone, it replies with a 200OK status message to the conference manager 12. The status message contains SDP information, i.e., the IP address of the UB terminal, the port number for RTP reception, the media encoding method (PCMA, PCMU, G729, etc.), etc.;

[0123] 5) After receiving the 200OK message, the conference management 12 extracts the SDP information, adds the member UB to the conference bridge, and at the same time supplements and improves the media instance PmB through the media management 13, filling in the IP address and remote RTP port number of the UB;

[0124] 6) The conference management 12 replies 200OK to the initiator UA of the session. The message contains SDP information, that is, the local port number in the UA media instance, the IP address of the scheduling service unit 1, and the media encoding method are consistent with those in the application (INVITE) message sent by the UA;

[0125] 7) After receiving the 200OK, the UA replies with an ACK message to the conference management 12, and the conference management 12 forwards the ACK message to the UB, indicating that the communication between the UA and the UB is officially established;

[0126] 8) Media management 13 implements media exchange based on the member relationship in the conference bridge, and establishes two RTP links connecting UA and UB based on the information of media instances PmA and PmB, and sends the exchanged media streams to the terminal devices of UA and UB to achieve normal media communication between the two. Of course, in addition to this, it also includes time alignment, anti-shake, gain addition, and stream buffering and other related processing to ensure clear and continuous communication quality between the two.

[0127] According to the characteristics of VOIP communication, it can be found through the description of a simple example that for a specific session, the signaling scheduling process before establishing media communication is continuous, that is, the message interaction needs to meet a certain order, and similarly, the message message processing also needs to meet a certain order. On the other hand, for high-concurrency communication service scenarios, the scheduling service unit 1 may process dozens or hundreds of sessions and their related messages at the same time, so the task processing must be multi-threaded, that is, the scheduling service must open multiple threads to process the received messages at the same time. This is bound to lead to a contradiction, that is, two message messages of a session may be randomly assigned to two scheduling service threads, and the processing speed of the later message is faster than that of the earlier message. The accumulation effect that occurs in this way leads to a disorder in the order of session processing.

[0128] The present invention takes into account the above-mentioned contradictory situations and designs the following Figure 4 The worker thread processing flow of the scheduling service unit 1 is shown. This flow can maintain the order of session message processing during high-concurrency message processing, which not only improves the performance of the scheduling service unit 1, but also ensures the reliability and stability of system processing. The worker thread processing flow of the scheduling service unit 1 is described as follows:

[0129] 1) For a number of received message messages, the scheduling service unit 1 first classifies the messages (by session ID) to form a task producer queue;

[0130] 2) Scheduling service unit 1 concurrently extracts messages from the task producer queue and sends them to the task consumers (in this example, the four worker threads of scheduling service unit 1). This operation does not delete the message from the task producer queue;

[0131] 3) For each worker thread, after receiving the task, it first goes back to the message queue, searches and confirms that the message is the first message to be processed in the message queue corresponding to a certain session (call-id), otherwise it will wait and confirm in a loop;

[0132] 4) After confirming that the pending message is the task that should be consumed currently (i.e., the first pending message in the message queue corresponding to a session), the worker thread immediately processes the business of the message, and after the processing is completed, deletes the message from the task producer queue;

[0133] 5) After the message is deleted from the task producer message queue in step 4, another worker thread will jump out of the loop waiting and immediately process the received task (the task is already the first message to be processed in the current message queue);

[0134] 6) Repeat steps 1 to 5 to meet the business requirements of high concurrency of multiple sessions and sequential processing of a single session in complex scenarios.

[0135] The voice communication unit 2 is the main human-computer interaction terminal, which carries the important tasks of user command input, sound collection, ground-to-air and ground-to-ground communication interaction, and communication status display. The unit is mainly composed of signaling interaction 21, media transceiver 22, and status management 23. It supports users to configure different roles as needed to log in, providing basic support for the wired and wireless business functions of civil aviation control. The priorities of different role configurations are different, and different wired and wireless resources can be allocated as needed. The user interface includes 3 functional areas and 2 status areas. The functional areas include wired button areas, wireless button areas, and advanced function button areas. The status area includes the terminal connection status area and the current communication status area. In the civil aviation control simulation training, according to the user role, the voice communication unit 2 can be divided into control instructor seats, control trainee seats, captain seats, etc. During the training process, personnel of each role realize voice conferences through handles and gooseneck headsets to complete the communication tasks of control training.

[0136] The system operation and maintenance unit 3 is used for remote configuration of system parameters and reporting of information on the status of each node unit in the system. The unit consists of two components: database parameter configuration 31 and status monitoring 32. Database parameter configuration 31 is mainly implemented through a database service based on a B / S architecture. Users access and modify data by entering the IP address of the database service and user authentication; status monitoring 32 obtains key information such as the connection status, communication status, configuration status, and operation log of each device in the system through interactive processing with the scheduling service unit 1, which can be used for display and storage.

[0137] The recording service unit 4 is a key module for realizing the review and comment process in the control training business, and is used to realize the media recording storage of each terminal device during the civil aviation control training process and the on-demand synchronous playback function during the training review. Figure 5 As shown, the unit adopts a three-layer architecture of “low coupling and high cohesion”, including a human-computer interaction module 41 , a business service module 42 and an interface management module 43 .

[0138] Among them, the human-computer interaction module 41 belongs to the interface display layer, which is used for displaying the parameters and status of each channel, recording and playback control operations, recording the current memory status, and disk usage management; the business service module 42 belongs to the business logic layer, which is used to implement voice recording, local playback, remote playback, disk alarm processing, record file deletion and transfer and other services; the interface management module 43 belongs to the data access layer, which is used to support interaction with analog recording cards and IP devices, receive PCM data or send RTP streams, and implement the underlying operations of recording and playback.

[0139] The VOIP radio gateway 5 serves as the terminal medium for ground-to-air communication and is used to simulate various VOIP radio instances. At the same time, it constructs radio transceiver links of different frequencies through internal conference bridge processing. The simulated crew seats and controller seats realize the ground-to-air communication interactive negotiation process by selecting the same frequency.

[0140] The VOIP recognition unit 6 is used to recognize the voice of ground-to-air communication in real time, recognize and generate structured instructions according to the voice issued by the control seat, drive the simulated aircraft of the external digital tower, and perform automated simulation processes including wheel block removal, start, push out, taxi, takeoff, climb, acceleration and deceleration. The unit is mainly composed of a recognition engine 61, a semantic analysis 62, and a recognition process 63. Among them, the recognition engine 61 adopts a convolutional neural network (CNN) and a continuous temporal classification (CTC) method, uses a large amount of structured aviation control instruction voice data for training, builds an acoustic model and a language model, and can convert the input RTP streaming media into output text information; the semantic analysis module 62 further parses the output text, parses the meaningless recognition text into a sequence containing several structured terms, and then matches it with the control instruction dictionary (established in advance), and finally extracts the instruction sequence with control meaning.

[0141] Combine the following Figure 2 The working process of the present invention is further described.

[0142] There are four steps to complete the simulation control communication command on the VOIP simulation communication platform.

[0143] Step 1: System initialization

[0144] First, when the simulation training server is started, the system database configuration module initializes the parameters of the scheduling service unit, recording service unit and VOIP radio gateway according to MySQL and local configuration files.

[0145] Secondly, the voice communication unit is turned on, the control seat / pilot seat enters the user name and password, selects the configured user role, logs in to the human-computer interaction interface of the voice communication unit, and the four accounts of the voice communication unit are connected to the dispatch service unit. The voice communication unit, system operation and maintenance unit, recording service unit, VOIP radio gateway and VOIP identification unit are automatically connected to the dispatch service unit as user instances, and maintain the keep-alive heartbeat. When the heartbeat is lost, the abnormal state will be uploaded to the system operation and maintenance unit for display.

[0146] At the same time, several simulated radio accounts configured by the system database configuration module establish communication with the dispatch service unit, forming two-way transceiver media links (silent packages) of several frequencies, and the corresponding radio buttons on the voice communication unit interface are displayed as available.

[0147] Step 2: Regulate analog communications

[0148] The control seat / pilot seat respectively clicks on the wireless display area of ​​the voice communication unit to select a receiving radio / transmitting radio of a certain frequency (118.125Hz). The two establish a communication link through the dispatch service unit and the VOIP radio gateway. The link RTP stream is initially a silent packet (payload part is 0xd5). According to the requirements of the control training manual, the control trainee presses PTT to issue flight control instructions. After the captain seat understands the control instructions, the flight trainee presses PTT to repeat the instructions and perform corresponding control actions. The voice communication unit collects the voice data of the control trainee's handle, performs A / D conversion and other related processing, and generates RTP data, which is sent to the media management module.

[0149] The media management module searches for relevant members in the current conference bridge, performs time alignment, de-jittering, and gain-adding on the received RTP stream, and then sends it to the radio receiving channel of the corresponding frequency of the VOIP radio gateway. After receiving the RTP stream, the VOIP radio gateway sends the RTP stream to the media management module through the closed-loop link through the transmission channel of the same frequency of the radio. The media management module searches for relevant members in the current conference bridge, performs time alignment, de-jittering, and gain-adding on the received RTP stream, and then sends it to the captain's seat of the selected frequency transmitting radio. The captain's seat plays the received RTP stream, which comes from the control trainee seat, realizing the control command from the control seat to the captain's seat.

[0150] The reverse link is similar, allowing the captain's seat to repeat control instructions to the control seat for playback.

[0151] At the same time, when the PTT is pressed at the control seat / captain seat, the recording service unit and the voice communication unit establish a recording channel link, record the RTP stream data sent by the voice communication unit in real time, convert the RTP stream into a PCM (or WAV) file for storage, and record the file index and file detailed information in the database.

[0152] The recording service unit scans the disk space regularly. When the space usage reaches the set first-level threshold, it sends an alarm status message to the monitoring system. When the second-level threshold is reached, the historical files are automatically deleted.

[0153] Step 3: Speech Recognition Processing

[0154] As described in step 2, during the control training process, both the control seat and the captain seat need to transmit voice by pressing PTT. When the control seat or the captain seat presses PTT, the voice communication unit obtains the RTP stream after A / D conversion and other processing of the sound signal collected by the handle, and transmits it to the VOIP recognition unit. The recognition engine module of the VOIP recognition unit extracts the Mel frequency cepstral coefficient (MFCC) of the RTP stream signal as a feature sequence, which is used as the input of the VOIP decoding process. With the help of the acoustic model based on the CNN+CTC algorithm, the word (word) dictionary, and the language model based on the N-gram grammar, the decoder of the recognition engine module uses the Viterbi algorithm to solve the Chinese character sequence with the maximum a posteriori probability corresponding to the current input feature sequence, and performs semantic analysis on the parsed text to obtain structured control instructions. The process includes sentence segmentation, part-of-speech tagging, control term matching, structured instructions, etc.

[0155] The recognition and processing module performs corresponding branch processing according to the control instructions obtained by recognition, interacts with the simulated aircraft of the external digital tower, and issues instructions to control the aircraft including removing wheel blocks, starting, pushing out, taxiing, taking off, climbing, accelerating and decelerating, so that the simulated aircraft can perform corresponding actions according to the voice requirements of the controller, and the specific actions of the aircraft can be seen in the digital tower view.

[0156] Step 4: Assessment and Review

[0157] After the control training is completed, the training officer selects the training subject, training start and end time, control trainee seat, captain seat and other parameters on the training control seat interface, initiates scene playback to the digital tower, and sends the parameters to the recording service unit. The recording service unit searches the database and searches for the locally stored audio files (PCM / WAV) and their storage locations through parameters such as start and end time, control trainee seat number, and captain seat number. At the same time, the recording service unit will establish a one-way communication link with the control trainee seat and the captain seat, that is, the RTP stream is transmitted from the recording service unit to the control trainee seat and the captain seat for playback in sync with the tower scene.

[0158] The recording service unit establishes a "ping-pong cache" locally, and reads RTP streaming media data for a certain period of time (the length of time can be set) from the local disk. When the training officer clicks to start playback on the training control seat interface, the "ping-pong cache" of the recording and revisiting unit starts playback, and sends the RTP stream to the control trainee seat and the captain seat for playback at a fixed time (20ms as a cycle). The RTP stream playback is synchronized with the digital tower scene playback in time. When the "ping-pong cache" finishes playback, it starts to continue reading the RTP stream from the local disk, and the "pong cache" starts playback at the same time, and sends the RTP stream to the control trainee seat and the captain seat for playback at a fixed time (20ms as a cycle). Repeat the above-mentioned stream reading and stream playback pipeline operation until all the voice data of the training process are played back, and the recording service unit releases the current "ping-pong cache". The playback of the voice is consistent with the playback of the digital tower scene, and the training officer and the trainees can analyze and evaluate the flight training process.

[0159] It can be seen from the above four steps that during the normal operation of the platform system, the present invention effectively supports the entire life cycle of civil aviation control training through complete core scheduling services, media control, recording and playback services, voice recognition services, system operation and maintenance and other technical services, including system initialization, ground-to-ground and ground-to-air communications during control training, control voice recognition and system status operation and maintenance.

[0160] In a specific implementation, the present application provides a computer storage medium and a corresponding data processing unit, wherein the computer storage medium can store a computer program, and when the computer program is executed by the data processing unit, the invention content of a VOIP simulation communication platform based on civil aviation control training business provided by the present invention and some or all steps in each embodiment can be executed. The storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.

[0161] Those skilled in the art can clearly understand that the technical solutions in the embodiments of the present invention can be implemented by means of computer programs and their corresponding general hardware platforms. Based on this understanding, the technical solutions in the embodiments of the present invention are essentially or partly contributed to the prior art can be embodied in the form of a computer program, i.e., a software product, which can be stored in a storage medium and includes several instructions for enabling a device including a data processing unit (which can be a personal computer, a server, a single-chip microcomputer, a MUU or a network device, etc.) to execute the methods described in various embodiments of the present invention or certain parts of the embodiments.

[0162] The present invention provides a thought and method for a VOIP simulation communication platform based on civil aviation control training business. There are many methods and ways to implement the technical solution. The above is only a preferred implementation of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the protection scope of the present invention. All components not specified in this embodiment can be implemented by existing technologies.

Claims

1. A VOIP simulation communication platform based on civil aviation control training business, adopting a distributed communication service architecture, characterized in that: include: A dispatch service unit (1), a voice communication unit (2), a system operation and maintenance unit (3), a recording service unit (4), a VOIP radio station gateway (5) and a VOIP identification unit (6); The dispatching service unit (1), the system operation and maintenance unit (3), the recording service unit (4), the VOIP radio station gateway (5) and the VOIP identification unit (6) are interconnected through an IP network bus and exchange control instructions and data; The scheduling service unit (1) implements the management of each access user in the platform, the scheduling and processing of wired and wireless services based on the ED137 series standards, the management of wired and wireless media links, and the management of master-slave switching and data synchronization; The voice communication unit (2) serves as a human-computer interaction terminal, configures different roles as needed for user login, and supports wired and wireless service functions of civil aviation control services; The system operation and maintenance unit (3) is used for remote configuration of the parameters of the platform and reporting of information on the status of each node unit in the platform; The recording service unit (4) is used to realize the media recording storage of each terminal device during the civil aviation control training and the synchronous playback on demand during the training review; The VOIP radio gateway (5) is used to simulate a VOIP radio instance, and at the same time, construct radio transceiver links of different frequencies through internal conference bridge processing, and the simulated crew seat and controller seat realize the ground-to-air communication interactive negotiation process by selecting the same frequency; The VOIP recognition unit (6) is used to recognize the voice of ground-to-air communication in real time, recognize the voice issued by the control seat and generate structured instructions, drive the simulated aircraft of the external digital tower, and perform an automated simulation process.

2. A VOIP simulation communication platform based on civil aviation control training business according to claim 1, characterized in that: The scheduling service unit (1) comprises a user management module (11), a session management module (12), a media management module (13) and a master-slave management module (14); The user management module (11) manages various types of access users in the system and maintains the user name, communication address, user type, enabled state, busy state, heartbeat timeout and timestamp of each user instance; The session management module (12) is used to implement business interaction logic control with external terminal equipment, including business logic control of wired functions and business logic control of wireless services; The media management module (13) extracts basic information and media parameter information based on the interactive content of the session management module (12), constructs a conference bridge list, and realizes the creation, fusion and destruction of the underlying multimedia link; at the same time, based on the address information extracted by the session management module (12), it interacts with the voice communication unit (2) and the VOIP radio gateway to send and receive media streams; The master-slave management module (14) is used to implement the state switching of the master and standby servers and the synchronization of business data; firstly, the master and standby states of the dual services are confirmed and controlled by the state machine, then the business data synchronization flow direction is determined according to the master-standby relationship, and finally, the real-time synchronization of each business data is achieved through a distributed synchronization method; The media management module (13) realizes the simulated noise superposition of wired and wireless links and the interruption of simulated communication links through database configuration according to the needs of the simulated communication business.

3. A VOIP simulation communication platform based on civil aviation control training business according to claim 2, characterized in that: The voice communication unit (2) comprises a signaling interaction module (21), a media transceiver module (22) and a state management module (23); the signaling interaction module (21) interacts with the scheduling service unit (1) using standard SIP messages to complete authentication registration, wired service and wireless service interaction; the media transceiver module (22) is used to implement the RTP stream exchange process between the local RTP stream and the media management module (13); the state management module (23) is used to implement the state management of all terminal number buttons configured in the maintenance interface, and the button states include idle, ringing, talking, holding and monitoring.

4. A VOIP simulation communication platform based on civil aviation control training business according to claim 3, characterized in that: The system operation and maintenance unit (3) comprises a database configuration module (31) and a status monitoring module (32); The database configuration module (31) is used to configure the SIP account, VCU information, user number, radio frequency and role information, and to load and update the parameters in the scheduling service unit (1); the database configuration module (31) adopts a B / S architecture design to support users to remotely access and control the parameters configured by the server; while loading the parameters, the master and standby synchronization of the database is realized; The status monitoring module (32) is used to realize status collection, wired and wireless communication records, operation logs, and abnormal alarms of various devices in the system, including the scheduling service unit (1), the voice communication unit (2), the recording service unit (4), the VOIP radio gateway (5), and the VOIP identification unit (6), and supports user access and viewing.

5. A VOIP simulation communication platform based on civil aviation control training business according to claim 4, characterized in that: The recording service unit (4) adopts a three-layer architecture, including a human-computer interaction module (41), a business service module (42) and an interface management module (43); Among them, the human-computer interaction module (41) belongs to the presentation layer, and is used for displaying the parameters and status of each channel, recording and replaying operations, recording the current memory status, and managing disk usage; the business service module (42) belongs to the business logic layer, and is used to implement voice recording, local and remote playback, disk alarm processing, record file deletion and transfer services; the interface management module (43) belongs to the data access layer, and is used to support interaction with analog recording cards and IP devices, receive PCM data or send RTP streams, and implement the underlying operations of recording and playback.

6. A VOIP simulation communication platform based on civil aviation control training business according to claim 5, characterized in that: The local and remote playback processes of the recording service unit (4) adopt a ping-pong cache method to implement a pipeline process of reading files, PCM decoding, and pushing RTP streams.

7. A VOIP simulation communication platform based on civil aviation control training business according to claim 6, characterized in that: The VOIP radio gateway (5) complies with the ED137 interface standard and is used to simulate ground-to-air communication transceiver channels of different frequencies to support communication training for controllers and crew members.

8. A VOIP simulation communication platform based on civil aviation control training business according to claim 7, characterized in that: The VOIP radio station gateway (5) comprises a radio station access module (51), a conference proxy module (52) and a parameter configuration module (53); The radio station access module (51) is used to process the interaction with the external real IP radio station and realize the transmission of sending and receiving bidirectional RTP links; The conference proxy module (52) serves as an intermediate proxy for the radio access module (51) and the media management module (13), and uses a back-to-back mode to maintain communication with the radio access module (51) and the media management module (13), thereby realizing the scheduling and exchange of signaling and media of the voice communication link; After the conference proxy module (52) establishes a conference with the media management module (13), after receiving the RTP stream from the media management module (13), the conference proxy module (52) forwards it to the corresponding receiving channel of the radio access module (51); similarly, after receiving the RTP stream of the transmission channel of the radio access module (51), the conference proxy module (52) forwards it to the corresponding receiving channel of the media management module (13); The parameter configuration module (52) is used to configure the parameters of each transceiver link of the VOIP radio gateway (5), including channel number, channel frequency, channel attribute, media delay, media additional noise and media additional on-off.

9. A VOIP simulation communication platform based on civil aviation control training service according to claim 8, characterized in that: The VOIP recognition unit (6) comprises a recognition engine module (61), a semantic analysis module (62) and a recognition processing module (63); The recognition engine module (61) uses a convolutional neural network and a continuous time series classification method based on deep learning theory, uses structured air traffic control instruction voice data for training, builds an acoustic model and a language model, and converts the input RTP streaming media into output text information; The recognition engine module (61) is composed of a signal preprocessing, an acoustic model, a language model, a dictionary and a decoder module; wherein the acoustic model is obtained by training the speech data in the corpus, and the language model is obtained by training the text information to obtain the probability of the correlation between single characters or words; The semantic analysis module (62) further analyzes the text output by the recognition engine module (61), parses the meaningless recognition text into a sequence containing a number of structured terms, and then matches it with the control instruction dictionary, and finally extracts the instruction sequence with control meaning; The recognition processing module (63) performs corresponding branch processing according to the obtained control instructions, interacts with the simulated aircraft of the external digital tower, and issues instructions for controlling the aircraft's movements, so that the simulated aircraft performs corresponding actions according to the controller's voice requirements.

10. A VOIP simulation communication platform based on civil aviation control training service according to claim 9, characterized in that: The platform performs the following steps: Step 1, initialization of parameters of each unit of the platform; Step 2, the trainee in the control seat interacts with the captain seat through a simulated communication link; Step 3: The platform recognizes the student's voice and drives the aircraft model to move according to the instructions; Step 4: After the control training is completed, the training process is reviewed and reviewed through playback; Wherein, step 1 comprises the following steps: Step 1-1, the database configuration module (31) implements parameter initialization of the scheduling service unit (1), the recording service unit (4) and the VOIP radio station gateway (5) according to the configuration file; Step 1-2, the voice communication unit (2) is turned on, the control seat or the captain seat enters the user name and password, and selects the configured user role, logs into the human-computer interaction interface of the voice communication unit (2), and the four accounts of the voice communication unit (2) are connected to the dispatch service unit (1); Step 1-3, the voice communication unit (2), the system operation and maintenance unit (3), the recording service unit (4), the VOIP radio gateway (5) and the VOIP identification unit (6) automatically access the scheduling service unit (1) as user instances and maintain a keep-alive heartbeat. When the heartbeat is lost, the abnormal state will be uploaded to the system operation and maintenance unit (3) for display; Step 1-4, the simulated radio account configured by the database configuration module (31) establishes communication with the scheduling service unit (1), forming a two-way media link of different frequencies, i.e., a silent packet, and the corresponding radio button on the interface of the voice communication unit (2) is displayed as available; Step 2 includes the following steps: Step 2-1, the control seat or the captain seat respectively clicks on a receiving radio station or a transmitting radio station of a frequency in the wireless display area of ​​the voice communication unit (2), and the two establish a communication link through the dispatch service unit (1) and the VOIP radio gateway (5), and the link RTP stream is initially a silent packet; Step 2-2, according to the requirements of the control training manual, the control trainee presses PTT to issue flight control instructions. After the captain understands the control instructions, the flight trainee presses PTT to repeat the instructions and perform corresponding control actions; Step 2-3, the voice communication unit (2) collects the voice data of the control trainee's handle, generates RTP data after processing, and sends it to the media management module (13); Step 2-4, the media management module (13) searches for relevant members in the current conference bridge, performs time alignment and de-jitter processing on the received RTP stream, and then sends it to the radio receiving channel of the corresponding frequency of the VOIP radio gateway (5); Step 2-5, after receiving the RTP stream, the VOIP radio gateway (5) sends the RTP stream to the media management module (13) through the closed-loop link via the transmission channel of the same frequency of the radio station; Step 2-6, the media management module (13) searches for relevant members in the current conference bridge, performs time alignment and de-jitter processing on the received RTP stream, and then sends it to the captain's seat of the selected frequency transmitting station; Step 2-7, the captain seat plays the received RTP stream, which comes from the control trainee seat, to realize the control command from the control seat to the captain seat; the reverse link is the same, to realize the captain seat repeating the control instructions to the control seat for playback; Step 2-8, when the PTT is pressed at the control seat or the captain seat, the recording service unit (4) and the voice communication unit (2) establish a recording channel link, record the RTP stream data sent by the voice communication unit (2) in real time, convert the RTP stream into a PCM or WAV file for storage, and record the file index and file detailed information in the database; Step 2-9, the recording service unit (4) periodically scans the disk space, and when it detects that the space usage rate reaches the set first-level threshold, it sends an alarm status information to the monitoring; and when it reaches the second-level threshold, it automatically deletes the historical files; Step 3 includes the following steps: Step 3-1, when the control seat or the captain seat presses the PTT, the voice communication unit processes the sound signal collected by the handle to obtain the RTP stream and transmits it to the VOIP recognition unit (6); Step 3-2, the recognition engine module (61) of the VOIP recognition unit (6) extracts the Mel frequency cepstral coefficients MFCC of the RTP stream signal as a feature sequence, and the feature sequence is used as an input of the VOIP decoding process; Step 3-3, with the aid of an acoustic model based on a CNN+CTC algorithm, a dictionary and a lexicon, and a language model based on an N-gram grammar, the decoder of the recognition engine module (61) uses a Viterbi algorithm to decode a Chinese character sequence with a maximum a posteriori probability corresponding to the current input feature sequence; Step 3-4, semantic analysis of the parsed text to obtain structured control instructions, including sentence segmentation, part-of-speech tagging, control term matching and structured instructions; Step 3-5, the recognition processing module (63) performs corresponding branch processing according to the control instructions obtained by recognition, interacts with the simulated aircraft of the external digital tower, issues instructions to control the aircraft's movements, and enables the simulated aircraft to perform corresponding actions according to the voice requirements of the controller; Step 4 includes the following steps: Step 4-1, after the control training is completed, the training officer selects the training subject, training start and end time, control trainee seat and captain seat parameters on the training control seat interface, initiates scene playback to the digital tower, and sends the parameters to the recording service unit (4); Step 4-2, the recording service unit (4) searches the database and searches for the locally stored audio file and its storage location according to the start and end time, the control trainee seat and the captain seat parameters; Step 4-3, the recording service unit (4) establishes a unidirectional communication link with the control trainee seat and the captain seat, that is, the RTP stream is unidirectionally transmitted from the recording service unit (4) to the control trainee seat and the captain seat for playback, which is synchronized with the tower scene; Step 4-4, the recording service unit (4) locally establishes a ping-pong cache and reads RTP streaming media data of a set time period from the local disk; Step 4-5, ping the buffer to start playback, and send the RTP stream to the control trainee seat and the captain seat for playback at a fixed time. The RTP stream playback is synchronized with the digital tower scene playback in time. Step 4-6, the ping cache ends playback and starts to read the RTP stream from the local disk. At the same time, the pong cache starts playback and sends the RTP stream to the control trainee seat and the captain seat for playback at a fixed time; Step 4-7, repeating steps 4-5 and 4-6 until all the speech data in the training process are played back, and the recording service unit (4) releases the current ping-pong buffer.

Images