CN117037556A - Real-soldier real-load countermeasure training system - Google Patents

Abstract

The application relates to the technical field of countermeasure training, and in particular discloses a real-soldier real-mounted countermeasure training system, which comprises the following components: the system can enable the army to conduct actual combat training in a real environment by means of the system, improve combat skills and coping capacity of soldiers and enhance combat force and combat consciousness of the soldiers.

Description

Real-soldier real-load countermeasure training system

Technical Field

The application relates to the technical field of countermeasure training, and more particularly relates to a real-soldier real-mounted countermeasure training system.

Background

The real-soldier real-installation countermeasure training system is a simulated combat environment and can provide more real combat experience and training effect. Through the system, the army can conduct actual combat training by using simulated weapon equipment in a real environment, so that combat skills and coping capacity of soldiers are improved, and combat force and combat consciousness of the soldiers are enhanced. In addition, the real-soldier practical combat training system can also reduce cost and risk of practical combat exercises and reduce loss of personnel and equipment.

Therefore, developing a real-soldier real-world combat training system is very important for improving the combat power and the combat ability of the army.

Disclosure of Invention

The application aims to provide an actual soldier real-installation countermeasure training system which can enable troops to conduct actual combat training by using simulated weapon equipment in a real environment, improve combat skills and coping ability of soldiers and enhance combat ability and combat consciousness of the soldiers.

In a first aspect, there is provided a real-soldier real-world combat training system, the system comprising: the individual suit comprises a helmet and training clothes; the simulated weapon comprises 92/92G simulated terminals, 95/95-1/03 simulated terminals, 95-B class light machine gun simulated terminals, 88 sniper/high-precision sniper simulated terminals and 35 durene simulated terminals; a data transmission station; a comprehensive display control system; and, auxiliary equipment, wherein the auxiliary equipment comprises a gun calibrating instrument and a conditioning gun; and the actual combat simulation countermeasure unit is connected with the individual harness, the simulated weapon, the data transmission radio station, the comprehensive display control system and the auxiliary equipment.

With reference to the first aspect, the embodiment of the present application provides a first possible implementation manner of the first aspect, wherein the actual combat simulation countermeasure unit includes; the personnel management module is used for managing personnel information and configuration; the training management module is used for displaying and controlling the training state; the achievement statistics module is used for counting battlefield data; the data management module is used for managing the combat scheme and weapon information; the system setting module is used for setting system parameters; and a system push-out module for closing the system.

With reference to the first possible implementation manner of the first aspect, the embodiment of the present application provides a second possible implementation manner of the first aspect, where the actual combat simulation countermeasure unit further includes a training policy module configured with an actual combat simulation countermeasure policy, where the actual combat simulation countermeasure policy includes: starting a training; judging whether to use the existing training program; responsive to determining to use the existing training program, opening an existing training program and clicking on the start to activate the device; creating a training plan in response to not using the existing training plan, and clicking on the start to activate the device; conducting pilot control, situation display and data analysis; and, ending the click.

With reference to the second possible implementation manner of the first aspect, the embodiment of the present application provides a third possible implementation manner of the first aspect, where the actual combat simulation countermeasure unit further includes a historical data management module, where the historical data management module is configured to manage historical data.

With reference to the third possible implementation manner of the first aspect, the embodiment of the present application provides a fourth possible implementation manner of the first aspect, wherein the managing the historical data includes: displaying situation; counting the combat losses of the combat fruits in real time; counting overall achievements; and, individual achievement statistics.

The application provides a real-soldier real-world combat training system, which comprises individual soldier outfits, simulated weapons, a data transmission radio station, a comprehensive display control system, auxiliary equipment and a real-soldier combat simulation combat unit, so that troops can conduct real combat training by using simulated weaponry in a real environment by means of the system, combat skills and coping capacity of soldiers are improved, and combat capacity and combat consciousness of the soldiers are enhanced.

Drawings

FIG. 1 is a schematic diagram of a real-soldier real-world countermeasure training system according to an embodiment of the present application;

FIG. 2 is a schematic diagram of a real-soldier real-world countermeasure training system according to an embodiment of the present application;

FIG. 3 is a schematic diagram of a training process in a real-soldier real-world countermeasure training system according to an embodiment of the present application;

FIG. 4 is a schematic diagram of historical data management in a real-soldier real-world combat training system according to an embodiment of the present application;

FIG. 5 is a schematic flow chart of determining whether a simulated weapon has a fault based on the sound effect of the simulated weapon in a real-soldier real-load combat training system according to an embodiment of the present application;

fig. 6 is a schematic flowchart of feature extraction of the plurality of sound effect sampling windows in the real-soldier practical combat training system to obtain the classified feature vector according to the embodiment of the present application.

In the figure: 1. an individual harness device; 2. 92/92G type handgun; 3. 95/95-1/03 type automatic rifle; 4. 88 sniping/high-precision sniping; 5. 95-B class light machine gun; 6. a data transfer station device; 7. and display control equipment.

Detailed Description

For the purpose of making the objects, technical solutions and advantages of the embodiments of the present application more apparent, the technical solutions of the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application, and it is apparent that the described embodiments are only some embodiments of the present application, not all embodiments. The components of the embodiments of the present application generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations. Thus, the following detailed description of the embodiments of the application, as presented in the figures, is not intended to limit the scope of the application, as claimed, but is merely representative of selected embodiments of the application. All other embodiments, which can be made by a person skilled in the art without making any inventive effort, are intended to be within the scope of the present application.

The actual fight training system is a new generation of simulation fight technology, adopts the technologies of laser simulation, space positioning, wireless communication, weapon equipment simulation and the like, and a laser transmitter can be hung on an actual automatic rifle to simulate the fire fight process of the automatic rifle and various battlefield targets by transmitting laser; the emission and hit effects are simulated by means of an acoustic, optical, smoke and vibration display device. The training system has the advantages of simple structure, convenient installation, convenient use and operation, high simulated shooting precision, stable and reliable work, can effectively strengthen the training on actual combat consciousness and technology, has no safety risk, does not need logistics and medical care, and is mainly used for carrying out simulated countermeasure training of actual soldiers, actual mounting and non-actual bullets.

The actual combat training system aims at improving the combined combat capability, takes actual combat training as traction, comprehensively utilizes advanced technologies such as laser communication, weapon simulation, digital sensing, wireless communication, computer simulation and the like, adopts a mode of combining laser simulation and digital simulation, and truly reflects the thermal interaction relation by integrally and systematically simulating the combat effectiveness and the battlefield effect of various weapons of the army, comprehensively records the practice information and builds an informationized practice training environment close to actual combat for the actual combat practice.

The actual combat training system consists of six parts including an individual combat tool (helmet and training clothes), an external simulation transmitting terminal, a conditioning terminal device, a data centralized communication workstation and a system background.

In this embodiment, the actual combat training system may be implemented as the actual combat simulation training system 100 shown in fig. 1. Specifically, referring to fig. 1 and 2, the actual combat simulation training system 100 includes: individual harness 110, simulated weapon 120, data transmission station 130, integrated display and control system 140, auxiliary equipment 150.

Wherein the individual harness 110 includes a helmet and a training suit. The helmet is equipment for protecting the heads of soldiers of special forces, and is also provided with various sensors, energy equipment, a communication system and the like, so that the soldiers can observe and communicate conveniently during combat. Helmets for modern special forces are often made of special materials, such as carbon fiber and ceramics, which can withstand high-speed bullets and explosion impact. The training clothes are standard clothes for soldiers of special forces, and have multiple functions of protecting soldiers, hiding identities, storing tools and the like in the course of action. The training clothes of modern special forces are made of light materials and are provided with a plurality of pockets and accessories for use and adjustment by soldiers. The color and style of the training clothes can be selected and adjusted according to different environments and tasks to achieve the best concealing effect, and the individual harness equipment 1 shown in fig. 2 can be seen.

The simulated weapon 120 comprises 92/92G simulated terminals, 95/95-1/03 simulated terminals, 95-B class light machine gun simulated terminals, 88 sniper/high-precision sniper simulated terminals and 35 durene simulated terminals.

The 92/92G analog terminal is used for simulating a 92/92G type pistol, in particular a 92 type pistol, a full-scale QSZ92 type semiautomatic pistol is divided into two calibers of 5.8mm and 9mm, the 92G type pistol is an improved pistol based on the 92 type pistol, and the 92G progress is more obvious on a gun barrel, and the reliability is obviously improved compared with the 92 type pistol. While the accuracy of 92G pistols is greatly improved over 92 pistols due to the extremely high processing levels of the barrels. 92/92G handguns are two common handguns now fully listed by the army, see the 92/92G type handgun 2 illustrated in FIG. 2.

Wherein, the 95/95-1/03 analog terminal is used for simulating a 95/95-1/03 type automatic rifle, and the 95 type automatic rifle is called as 95 for short: the QBZ95 type automatic rifle belongs to a part of a 5.8mm rifle group, is one of standard rifles of the liberation army and armed forces of Chinese people, has the caliber of 5.8mm, is of a bracket-free structure, and has good stability; the precision is high, and the rifle body is shorter, and equilibrium is good, and convenient to carry makes it can fight in all weather, and 5.8 millimeter ammunition can break through 8 millimeter steel sheet in 100 meters and still take weak killing power, and the killing power is big. The 95 series automatic rifle is a common rifle series that is now fully assembled by armies, see the 95/95-1/03 type automatic rifle 3 illustrated in fig. 2.

The 95-B light machine gun simulation terminal is used for the 95-B light machine gun, the 95-type 5.8mm light machine gun is a light machine gun in a 95-type gun group, and the 95-type light machine gun simulation terminal and the 95-type automatic rifle form the gun group for the work, and armies are equipped successively. See the 95-B class light machine gun 5 illustrated in fig. 2.

The 88 sniper/high-precision sniper simulation terminal is used for simulating an 88 sniper/high-precision sniper, and the 88 sniper full scale QBU-88 sniper rifle is an active sniper rifle for the Chinese army. The caliber is 5.8mm, and the ammunition is universal with a 95-type automatic rifle and is of a bracket-free design. Full weight 4.1kg, full length 920mm, magazine capacity 10, effective range 800m. See fig. 2 for a schematic diagram of 88 sniping/high-precision sniping 4.

Among them, a 35 grenade simulation terminal is used for simulating a 35 mm grenade launcher, and the 35 mm grenade launcher has been developed and installed in parallel in the army, and is one of the most important fire weapons in the army.

The data transceiver 130 is configured to perform wireless communication. Here, the data transmission station may be an individual soldier station, which is capable of connecting individual infantries of a hiking combat in a class laterally, and is therefore also referred to as an "intra-class speaker". The airlines can enter a higher level command network to join the entire class into a digital battlefield. See the data transfer station apparatus 6 illustrated in fig. 2.

The integrated display control system 140 is used for monitoring and controlling a training process, and specifically, the integrated display control system may be used for monitoring running state information of participating equipment and key link information of equipment running in real time, so as to provide real-time accurate equipment state data for a decision and evaluation personnel and provide a decision and evaluation basis. See the display control device 7 illustrated in fig. 2.

The auxiliary equipment 150 includes a gun calibration instrument and a conditioning gun, and various auxiliary equipment is important in practical simulation besides fire equipment, for example, the gun calibration instrument can be used for calibrating and adjusting the accuracy and performance of the weapon. The conditioning gun is used for laser real soldier to resist and guide, controls individual soldiers and weapons and is handheld equipment of a training site instructor. These devices are all designed to allow soldiers to more realistically simulate the situation of a real battle field during training, improving their combat and coping capacities.

Preferably, the real-soldier actual combat training system further comprises a real-soldier combat simulation combat unit, wherein the real-soldier combat simulation combat unit comprises; the personnel management module is used for managing personnel information and configuration; the training management module is used for displaying and controlling the training state; the achievement statistics module is used for counting battlefield data; the data management module is used for managing the combat scheme and weapon information; the system setting module is used for setting system parameters; and a system push-out module for closing the system.

Specifically, the personnel management module is used for managing personnel information and configuration of personnel participating in training, including personal information of soldiers, units of the soldiers, positions of the soldiers and the like. The training management module is used for displaying and controlling training states, including operations such as training start, suspension, recovery and ending. The achievement statistics module is used for counting battlefield data, including soldier performances, weapon use cases and tactical response. The data management module is used for managing the combat scheme and weapon information, including combat plans, weapon performances and the like. The system setting module is used for setting system parameters including training sites, weather, time and the like. The system push module is used for closing the system, including data storage, system exit and other operations. The modules cooperate with each other to form a complete actual combat simulation countermeasure unit, which can help soldiers to perform more real combat simulation training and improve their combat capacity and coping capacity.

Referring to fig. 3, fig. 3 is a schematic diagram of a training process in a real-soldier real-world countermeasure training system according to an embodiment of the present application. As shown in fig. 3, the actual combat simulation countermeasure unit further includes a training strategy module configured with an actual combat simulation countermeasure strategy including: starting a training; judging whether to use the existing training program; responsive to determining to use the existing training program, opening an existing training program and clicking on the start to activate the device; creating a training scheme in response to not using an existing training program, and clicking on the device to activate the device, the creating the training scheme including entering programming equipment data, entering training programming data, associating the training device; conducting pilot control, situation display and data analysis; and, ending the click. Here, a complete training process is equal to a training stage in which soldiers perform actual combat simulation training, plus an evaluation stage, by performing combat simulation using various equipment and weapons, to improve their combat and coping abilities. During the evaluation phase, the soldiers' performance will be evaluated to determine how they perform in the maneuver and to determine the aspects they need further improvement. The assessment phase may also be used to determine the effectiveness and effect of the training in order to make the necessary adjustments and improvements. Through the training process, soldiers can improve fight skills and coping capacity of the soldiers in a more realistic training environment, and fight force and fight consciousness of the soldiers are enhanced.

The actual combat simulation countermeasure unit further comprises a historical data management module, wherein the historical data management module is used for managing historical data. Referring to fig. 4, fig. 4 is a schematic diagram of historical data management in a real-soldier real-world countermeasure training system according to an embodiment of the present application. The managing the history data includes: situation playback assessment, brief data assessment, and deletion. Wherein the situational playback assessment comprises: situation display, real-time combat damage statistics, overall score statistics and individual score statistics. The profile assessment includes overall performance statistics and individual performance statistics. Specifically, the situation display may show past combat situations, the real-time combat loss statistics may help the troops to learn past combat results and loss situations, the overall performance statistics may help the troops evaluate the performance of the overall combat process, and the individual performance statistics may help the troops evaluate the performance of each soldier. The overall performance statistics and individual performance statistics may help the troops learn past performance in order to better improve training and combat ability. Meanwhile, the outdated historical data is deleted, so that the army can be helped to manage the data better, and the safety and reliability of the data are improved.

In a specific embodiment of the present application, the actual combat simulation countermeasure unit further includes a failure prediction module, where the failure prediction module is configured to: whether the simulated weapon has a fault is determined based on the sound effect of the simulated weapon. It should be appreciated that various failures may occur during use of the simulated weapon in training, such as problems with stock loosening, barrel wear, magazine jamming, etc. These faults may lead to reduced performance of the simulated weapon, affect the training effect, and may even cause injury to the operator. Therefore, by collecting and analyzing the sound effects of the simulated weapons, the faults can be found in time, and the simulated weapons can be overhauled or replaced, so that the safety and the effectiveness of training are ensured. Therefore, it is necessary to determine whether the simulated weapon has a fault based on the sound effect of the simulated weapon.

FIG. 5 is a schematic flow chart of a simulated weapon based on sound effects of the simulated weapon in a real-soldier real-load combat training system according to an embodiment of the present application. As shown in fig. 5, the determining whether the simulated weapon has a fault based on the sound effect of the simulated weapon includes:

s110, acquiring sound effect acoustic wave signals of the simulated weapon. It should be appreciated that sound is considered an important signal that may reflect the state and performance of the simulated weapon. During use of the simulated weapon in training, different sound signals, such as gunshot, magazine loading and unloading sound, gun working sound, etc., are emitted. These sound signals may reflect the operational status and performance of the simulated weapon, and may change accordingly if the simulated weapon fails or performance is degraded. By acquiring acoustic sound signals of the simulated weapon, fault diagnosis and performance evaluation can be performed on the simulated weapon. By analyzing and processing the sound signals, useful characteristic information, such as frequency, amplitude, duration, etc., of the sound can be extracted, which can be used to determine if the simulated weapon is malfunctioning or has degraded performance. It is therefore necessary to acquire sound signals simulating the sound effects of a weapon.

In a specific embodiment of the application, the acoustic sound signals of the simulated weapon are acquired by an acoustic sensor.

In another specific embodiment of the application, a microphone or other sound collection device is used to collect sound effect acoustic signals of the simulated weapon. Specifically, using a microphone or other sound collection device to collect sound effect acoustic signals of the simulated weapon includes: a suitable microphone or other sound collection device is selected and connected to a computer or other device. A microphone or other sound collection device is placed in proximity to the simulated weapon to ensure that it can clearly collect the sound signals emitted by the simulated weapon. And opening sound collection software to start sound collection. During the acquisition process, the simulated weapon may be subjected to different operations, such as firing, changing clips, etc., in order to acquire different sound signals. After the collection is completed, the collected sound signals are stored as audio files, such as wav format files. Through the steps, the sound effect acoustic wave signal of the simulated weapon can be obtained. In the subsequent processing, the audio file may be analyzed and processed to extract useful characteristic information for simulating fault diagnosis and performance assessment of the weapon.

S120, sliding window sampling based on the sampling window is carried out on the sound effect sound wave signals so as to obtain a plurality of sound effect sampling windows. It should be understood that in actual processing of the sound signal, the sound signal is often continuous and has a long time, which may lead to problems of large calculation amount, low efficiency, inaccurate result, etc. if the whole signal is directly processed. Therefore, the sliding window sampling method is adopted to divide the long-time sound effect signal into a plurality of short-time sampling windows, so that the problems can be effectively solved. Specifically, sliding window sampling based on sampling windows is performed on the sound effect acoustic signals, and the long-time sound effect signals can be divided into a plurality of shorter sampling windows, so that subsequent processing and analysis are facilitated. By sampling the sound effect acoustic signals with sliding windows based on sampling windows, a plurality of sound effect sampling windows can be obtained, and each sampling window contains a sound effect signal within a short period of time. These sampling windows may be used for subsequent feature extraction and analysis, such as extracting characteristic information of frequency, amplitude, duration, etc. of sound for determining whether a simulated weapon is malfunctioning or degraded. Meanwhile, by adopting a sliding window sampling method, the time sequence information of the sound effect signal can be reserved, so that the state and the performance of the simulated weapon can be reflected more accurately.

And S130, respectively passing the sound effect sampling windows through a sound effect feature extractor based on a convolutional neural network model to obtain waveform feature vectors of the sound sampling windows. It should be appreciated that, given that sound signals are typically high-dimensional, contain a large amount of redundant information, direct use of the original signal for analysis and processing can result in computationally intensive, inefficient problems. By using a convolutional neural network model, the sound features within each sampling window can be automatically learned, and the most representative feature vectors can be extracted. These feature vectors may be used for subsequent signal analysis and processing, such as classification, clustering, dimension reduction, etc. The method can improve the accuracy and efficiency of feature extraction, thereby improving the accuracy and efficiency of fault detection and performance evaluation.

Optionally, in one embodiment of the present application, passing the plurality of acoustic sampling windows through an acoustic feature extractor based on a convolutional neural network model to obtain a plurality of acoustic sampling window waveform feature vectors, respectively, includes: processing the plurality of sound effect sampling windows by using the sound effect feature extractor based on the convolutional neural network model according to the following convolutional coding formula to obtain waveform feature vectors of the plurality of sound sampling windows;

wherein, the convolution coding formula is:

f _i ＝GAP{Sigmoid(N _i ×f _i-1 +B _i )}

wherein f _i-1 Input to the i-th layer sound effect feature extractor, f _i For the output of the i-th layer sound effect feature extractor, N _i A filter which is an i-layer sound effect feature extractor, and B _i For the bias matrix of the i-th layer sound effect feature extractor, sigmoid represents a nonlinear activation function, and GAP represents performing a global feature pooling operation on each feature matrix of the feature map.

And S140, passing the plurality of sound sampling window waveform characteristic vectors through a sequence encoder based on an RNN model to obtain classification characteristic vectors. It should be appreciated that, given that the sound signal has a time sequence, i.e. there may be a correlation between adjacent sound sampling windows, the acoustic effect feature extractor based on the convolutional neural network model cannot capture such a correlation. Thus, by using a RNN model-based sequence encoder, multiple sound sampling window waveform feature vectors can be subjected to time sequential feature extraction with global to obtain classification feature vectors. Specifically, the RNN model-based sequence encoder may take as input each sound sample window waveform feature vector, and process it through multiple RNN units to obtain a fixed length vector, which may be considered as a feature representation of the entire sequence. This feature representation may be used for classification, regression, etc. tasks. In addition, the RNN model has memory capability, and historical information can be considered when processing the sequence data, so that the characteristics of the sequence data can be better captured. Thus, passing a plurality of sound sampling window waveform feature vectors through an RNN model-based sequence encoder to obtain classification feature vectors may improve the accuracy of classification or regression tasks.

Optionally, in one embodiment of the present application, passing the plurality of sound sampling window waveform feature vectors through an RNN model-based sequence encoder to obtain the classification feature vector includes: processing the plurality of sound sampling window waveform feature vectors using the RNN model-based sequence encoder in a sequence encoding formula to obtain the classification feature vectors;

wherein the sequence coding formula is as follows

s _t ＝f(Ux _t +Ws _t-1 )

O _t ＝g(Vs _t )

Wherein x is _t Representing the input at the current time, s _t-1 Represents the output of the hidden layer at the previous moment, U represents the weight of the input sample at the current moment, W represents the weight of the hidden layer at the previous moment, s _t Representing the input of the hidden layer at the current moment, V representing the sample weight of the output, f and g being the activation functions, O _t Indicating the output of the hidden layer at the current time.

Fig. 6 is a schematic flowchart of feature extraction of the plurality of sound effect sampling windows in the real-soldier practical combat training system to obtain the classified feature vector according to the embodiment of the present application. In another embodiment of the present application, as shown in fig. 6, the determining whether the simulated weapon has a fault based on the sound effect of the simulated weapon includes: and carrying out feature extraction on the plurality of sound effect sampling windows to obtain classified feature vectors. The feature extraction of the plurality of sound effect sampling windows to obtain classified feature vectors comprises the following steps:

s210, preprocessing the sound effect sampling windows to obtain preprocessed sound effect sampling windows, wherein the preprocessing comprises operations of noise reduction, filtering, gain and the like. It will be appreciated that in order to improve signal-to-noise ratio and signal quality, thereby better preserving useful signal information, subsequent feature extraction and fault detection tasks are facilitated. In practical applications, audio signals often contain noise and noise interference components, which affect the quality and reliability of the signal. Thus, the preprocessing operation is performed on the original audio signal, so that the interference components can be removed or reduced, the signal-to-noise ratio and the signal quality can be improved, and useful signal information can be better reserved. Common preprocessing operations include noise reduction, filtering, gain, and the like. Noise reduction can remove noise components and improve the signal to noise ratio; the filtering can remove unnecessary frequency components and improve the signal quality; the gain may adjust the amplitude of the signal to better accommodate subsequent processing operations.

S220, dividing the preprocessed sound effect sampling window into a plurality of frames to obtain a plurality of preprocessed sound effect sampling frames, wherein each frame comprises a plurality of sampling points. It should be appreciated that the sound signal is sliced in time to facilitate frequency analysis of the sound signal. The sound signal is a time domain signal containing information in both time and frequency dimensions. The sound signal is divided into a plurality of frames, the time dimension can be discretized, and the continuous signal is divided into a plurality of discrete time slices, so that frequency analysis and processing are convenient. Each frame contains several sampling points because the frequency of the sound signal is continuously varying, requiring sampling of the sound signal to convert the continuous signal into a discrete signal. Each sample point contains amplitude information of the sound signal at a certain point in time, a plurality of sample points form a frame of the sound signal, and a plurality of frames form a time domain representation of the sound signal.

And S230, windowing the plurality of preprocessed sound effect sampling frames to obtain windowed sound effect sampling frames. It should be appreciated that to reduce abrupt changes in the time domain of the signal, the adjacent sampling frames are made more smoothly continuous, thereby helping to improve the frequency domain resolution of the signal and reduce spectral leakage. In the processing of sound signals, commonly used window functions are hamming windows, hanning windows, blackman windows, etc. The windowed sound effect sampling frame can better reflect the frequency domain characteristics of the signals, so that more accurate characteristic vectors are provided, and subsequent fault detection and diagnosis are facilitated.

S240, carrying out Fourier transform on the windowed sound effect sampling frame to obtain a transformed sound effect sampling frame. It should be appreciated that the fourier transform may convert a signal in the time domain to a signal in the frequency domain, thereby better reflecting the frequency domain characteristics of the signal. In sound signal processing, the windowed sound effect sampling frame can be converted into a spectrogram through Fourier transformation, and the spectrogram reflects the strength and distribution conditions of signals on different frequencies and is an important basis for feature extraction and fault detection. By further processing the transformed sound effect sampling frame, such as filtering, dimension reduction and the like, more useful feature vectors can be extracted, so that fault diagnosis and performance evaluation of the simulated weapon are realized.

S250, calculating the energy spectrum density of the sound effect sampling frames after the change, namely summing the square of the frequency domain signals of each frame. It should be understood that energy spectral density refers to the average energy per unit frequency band and is an important indicator reflecting the signal strength distribution. In the sound signal processing, by calculating the energy spectrum density of the sound effect sampling frame after transformation, the energy distribution situation of the signal can be more accurately described, and thus, more useful feature vectors can be extracted. The energy spectral density is calculated by squaring and summing the frequency domain signals of each frame, which can be implemented by a Fast Fourier Transform (FFT). By calculating the energy spectral density, the main frequency components in the signal can be identified, thereby realizing fault diagnosis and performance evaluation of the simulated weapon.

S260, the energy spectrum density is passed through a set of Mel filter banks, and the linear interval on the frequency axis is converted into Mel scale, so as to obtain Mel spectrum. It should be appreciated that mel frequency is a non-linear frequency scale based on the auditory characteristics of the human ear that better reflects the perception of sound frequency by the human auditory system. In the sound signal processing, the linear interval on the frequency axis is converted into the Mel scale by passing the energy spectrum density through a group of Mel filter groups, so that the perception process of human ears on sound can be more accurately simulated, and more useful feature vectors can be extracted. In the mel filter bank, the low-frequency area filter is denser, and the high-frequency area filter is sparser, so that the characteristic that the sensitivity of human ears to low-frequency sounds is higher and the sensitivity to high-frequency sounds is lower can be reflected better. By converting the energy spectrum density into the mel spectrum, the frequency distribution situation of the sound signal can be described more accurately, thereby realizing fault diagnosis and performance evaluation of the simulated weapon.

S270, carrying out logarithmic operation on the Mel frequency spectrum, and converting the Mel frequency spectrum into logarithmic scale to obtain logarithmic Mel frequency spectrum coefficients. It should be understood that the logarithmic mel-frequency spectrum coefficient is a set of eigenvectors obtained by converting the mel-frequency spectrum into a logarithmic scale through a logarithmic operation. The logarithmic scale may better describe the dynamic range of the sound signal, making the feature vector more stable and reliable. In the processing of the sound signal, the mel spectrum is converted into a logarithmic scale by performing a logarithmic operation on the mel spectrum, thereby obtaining a logarithmic mel spectrum coefficient. This reduces the differences between the individual components in the feature vector, making the feature vector more comparable. The logarithmic mel-frequency spectrum coefficient is one of characteristic vectors commonly used in sound signal processing, and can be used for tasks such as voice recognition, audio classification, music information retrieval and the like. In simulated weapon fault detection, logarithmic mel frequency spectrum coefficients can also be used to extract features to enable fault diagnosis and performance assessment of simulated weapons.

S280, performing discrete cosine transform on the logarithmic Mel frequency spectrum coefficient to obtain a coefficient after the sound effect sampling frame after the change. It should be appreciated that logarithmic mel-frequency spectral coefficients (Log Mel Spectrogram Coefficients) are a common method of representing sound features that can convert sound signals into a set of digital features that can be processed by machine learning algorithms. However, the logarithmic mel-frequency spectrum coefficient itself is a frequency domain feature, and there is much redundant information. To reduce this redundant information and improve the expressive power of the features, the logarithmic mel-frequency spectral coefficients may be discrete cosine transformed (Discrete Cosine Transform, DCT) and converted to another space for representation. In this new space, the DCT can compress information of a high frequency part to a low frequency part, thereby reducing redundant information and improving expressive power of features. Therefore, after DCT conversion is carried out on the logarithmic Mel frequency spectrum coefficient, the characteristics of the sound signal can be reflected better, and the accuracy of fault detection is improved.

S290, taking the coefficient after the sound effect sampling frame after the change as a feature vector to obtain a classification feature vector.

And S150, the classification feature vector passes through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the simulated weapon has a fault. It should be understood that after the classification feature vector is obtained, it needs to be input into a classifier to classify, so as to obtain a classification result of whether the simulated weapon has a fault. A classifier is a machine learning algorithm that can map input feature vectors into different classes. In the acoustic signal processing, various classifiers such as a support vector machine (Support Vector Machine, SVM), decision Tree (Decision Tree), random Forest (Random Forest), and the like may be used. The classification feature vector is passed through the classifier to obtain a classification result, so that fault detection and performance evaluation of the simulated weapon can be effectively performed.

Optionally, in one embodiment of the present application, passing the classification feature vector through a classifier to obtain a classification result includes: processing the classification feature vector with the following classification formula to obtain the classification result;

wherein, the classification formula is:

O＝softmax{(W _c ,B _c )|X}

wherein X represents the classification feature vector, W _c Is a weight matrix, B _c Representing the bias vector, softmax representing the normalized exponential function, and O representing the classification result.

It will be apparent to those skilled in the art that the elements or steps of the application described above may be implemented in a general purpose computing device, they may be concentrated on a single computing device, or they may alternatively be implemented in program code executable by a computer device, such that they may be stored in a storage device for execution by the computing device, or they may be separately fabricated into individual integrated circuit modules, or a plurality of modules or steps in them may be fabricated into a single integrated circuit module. Thus, the present application is not limited to any specific combination of hardware and software.

While the application has been described in detail in connection with specific preferred embodiments thereof, it is not to be construed as limited thereto, but rather as a result of a simple deduction or substitution by a person having ordinary skill in the art without departing from the spirit of the application, which is to be construed as falling within the scope of the application defined by the appended claims.

Claims (9)

1. A real-soldier real-world countermeasure training system, comprising:

the individual suit comprises a helmet and training clothes;

the simulated weapon comprises 92/92G simulated terminals, 95/95-1/03 simulated terminals, 95-B class light machine gun simulated terminals, 88 sniper/high-precision sniper simulated terminals and 35 durene simulated terminals;

a data transmission station;

a comprehensive display control system;

auxiliary equipment, wherein the auxiliary equipment comprises a gun calibrating instrument and a conditioning gun;

and the actual combat simulation countermeasure unit is connected with the individual harness, the simulated weapon, the data transmission radio station, the comprehensive display control system and the auxiliary equipment.

2. The real-soldier practical combat training system according to claim 1, characterized in that said real-soldier combat simulation combat unit comprises:

the personnel management module is used for managing personnel information and configuration;

the training management module is used for displaying and controlling the training state;

the achievement statistics module is used for counting battlefield data;

the data management module is used for managing the combat scheme and weapon information;

the system setting module is used for setting system parameters;

and the system pushing module is used for closing the system.

3. The real-soldier actual combat training system of claim 2, wherein the real-soldier combat simulation combat unit further comprises a training strategy module configured with a real-soldier combat simulation combat strategy comprising:

starting a training;

judging whether to use the existing training program;

responsive to determining to use the existing training program, opening an existing training program and clicking on the start to activate the device;

creating a training plan in response to not using the existing training plan, and clicking on the start to activate the device;

conducting pilot control, situation display and data analysis;

and (5) ending clicking.

4. The real-soldier packing countermeasure training system according to claim 3, wherein the real-soldier engagement simulation countermeasure unit further includes a history data management module for managing history data.

5. The real-soldier packing countermeasure training system of claim 4, wherein the real-soldier engagement simulation countermeasure unit further includes a failure prediction module for: whether the simulated weapon has a fault is determined based on the sound effect of the simulated weapon.

6. The real-soldier packing countermeasure training system of claim 5, wherein the failure prediction module is further configured to:

acquiring sound effect acoustic wave signals of the simulated weapon;

sliding window sampling based on a sampling window is carried out on the sound effect sound wave signals so as to obtain a plurality of sound effect sampling windows;

respectively passing the sound effect sampling windows through a sound effect feature extractor based on a convolutional neural network model to obtain waveform feature vectors of the sound sampling windows;

passing the plurality of sound sampling window waveform feature vectors through an RNN model-based sequence encoder to obtain classification feature vectors;

and passing the classification feature vector through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the simulated weapon has faults.

7. The real-world countermeasure training system of claim 6, wherein the passing the plurality of acoustic sampling windows through acoustic feature extractors based on convolutional neural network models, respectively, to obtain a plurality of acoustic sampling window waveform feature vectors comprises:

processing the plurality of sound effect sampling windows by using the sound effect feature extractor based on the convolutional neural network model according to the following convolutional coding formula to obtain waveform feature vectors of the plurality of sound sampling windows;

wherein, the convolution coding formula is:

f _i ＝GAP{Sigmoid(N _i ×f _i-1 +B _i )}

wherein f _i-1 Input to the i-th layer sound effect feature extractor, f _i For the output of the i-th layer sound effect feature extractor, N _i A filter which is an i-layer sound effect feature extractor, and B _i For the bias matrix of the i-th layer sound effect feature extractor, sigmoid represents a nonlinear activation function, and GAP represents performing a global feature pooling operation on each feature matrix of the feature map.

8. The real-world combat training system of claim 7, wherein said passing said plurality of sound sampling window waveform feature vectors through an RNN model based sequence encoder to obtain a classification feature vector, comprises: processing the plurality of sound sampling window waveform feature vectors using the RNN model-based sequence encoder in a sequence encoding formula to obtain the classification feature vectors;

wherein,the sequence coding formula is s _t ＝f(Ux _t +Ws _t-1 )

O _t ＝g(Vs _t )

Wherein x is _t Representing the input at the current time, s _t-1 Represents the output of the hidden layer at the previous moment, U represents the weight of the input sample at the current moment, W represents the weight of the hidden layer at the previous moment, s _t Representing the input of the hidden layer at the current moment, V representing the sample weight of the output, f and g being the activation functions, O _t Indicating the output of the hidden layer at the current time.

9. The real-world combat training system of claim 8, wherein said passing said classification feature vectors through a classifier to obtain a classification result includes: processing the classification feature vector with the following classification formula to obtain the classification result;

wherein, the classification formula is:

O＝softmax{(W _c ，B _c )|X}

wherein X represents the classification feature vector, W _c Is a weight matrix, B _c Representing the bias vector, softmax representing the normalized exponential function, and O representing the classification result.