Overwater autonomous rescue device and system

Abstract

The invention relates to a water autonomous rescue device, which floats in a water area, continuously senses an environment message of the water area, performs sound identification on the environment message to identify a help-seeking message from the environment message, calculates the direction of a drowning person who sends the help-seeking message through a time difference algorithm, and transmits the direction of the drowning person to the water area. The water autonomous rescue device automatically advances towards the direction of the emission source, so that the drowning person can climb on the water autonomous rescue device to be carried to a safe position; according to the water autonomous rescue system, the water autonomous rescue device is further connected to the monitoring device, monitoring personnel can check the state of the water autonomous rescue device through the monitoring device, and when a drowning accident occurs, the monitoring personnel can go to carry out subsequent rescue according to the position of the water autonomous rescue device.

Description

Technical Field

[0001] The present invention relates to a rescue device and a system, in particular to an autonomous rescue device floating on the water surface and a system comprising the device. Background Art

[0002] In summer, a large number of people go to the beach to play in the water to relieve the heat, but there are frequent reports of drowning incidents caused by playing in the water. The main reasons for the above-mentioned regrettable drowning accidents are usually: when the drowning accident occurs, the drowning person is alone or no one is around to notice him, and he misses the golden opportunity to be rescued, or when the drowning accident occurs, the drowning person is immediately discovered, but the discoverer cannot immediately rescue him, for example: the discoverer lacks a rescue object (such as a lifebuoy) to provide the drowning person with, or the discoverer is not a professional lifeguard. When the discoverer spends time looking for the rescue object or notifying the lifeguard, the golden opportunity for the drowning person to be rescued is gradually passing. Therefore, how to increase the probability of successful rescue of drowning people is a topic that needs to be paid great attention to in today's society. Summary of the invention

[0003] When a drowning accident occurs, the golden opportunity for rescue is usually missed due to various reasons as mentioned in the background technology, resulting in many drowning victims losing their lives without being successfully rescued. In view of this, the present invention proposes an autonomous water rescue device and system to increase the probability of successful rescue of drowning victims.

[0004] The present invention provides an autonomous water rescue device, comprising:

[0005] A life buoy, floating in a water area;

[0006] a sound acquisition module, disposed on the lifesaving buoy, the sound acquisition module sensing an environmental sound signal of the water area;

[0007] a driving module, disposed on the lifesaving buoy; and

[0008] A central processing module is arranged on the life-saving float and electrically connected to the sound capture module and the driving module. The central processing module receives the environmental sound and executes program data of a sound recognition model to recognize a distress audio from the environmental sound. The central processing module calculates the direction of a transmission source according to the distress audio to control the driving module to drive the life-saving float to move toward the transmission source.

[0009] In the above-mentioned autonomous water rescue device, the central processing module includes an audio separation unit and a feature extraction unit. The audio separation unit receives the ambient audio and independently separates a plurality of independent audios from the ambient audio through independent components analysis (ICA) and blind source separation (BSS). The feature extraction unit receives the plurality of independent audios and performs feature extraction on the plurality of original audios through the Mel-Frequency Cepstral Coefficients (MFCC) algorithm to output a plurality of feature information.

[0010] As described above, in the autonomous water rescue device, the central processing module includes a sound recognition unit, which receives the multiple feature information and executes the program data of the sound recognition model to classify the multiple feature information to confirm whether the multiple feature information includes the feature information of the distress audio.

[0011] In the above-mentioned autonomous water rescue device, the central processing module calculates the direction of the transmitting source of the distress signal through a time difference algorithm (TDOA); and the sound recognition model is trained through a convolutional neural network algorithm (CNN).

[0012] In the above-mentioned autonomous water rescue device, the sound capture module includes four sound receiving units, and the central processing module calculates a time difference between the distress audio being sent to any two of the four sound receiving units, and then calculates an angle between the two sound receiving units and the transmitting source through the time difference, wherein the direction of the transmitting source is the average value of six non-repetitive angles obtained by the central processing module after six calculations.

[0013] As described above, in the autonomous water rescue device, the central processing module is further electrically connected to a positioning module, and the positioning module is arranged on the life-saving float and transmits a positioning information to the central processing module according to the position of the life-saving float, and the positioning module includes an electronic compass, so that the positioning information includes the coordinate information and the azimuth information.

[0014] In the above-mentioned autonomous water rescue device, the driving module is two motors arranged at the bottom of the life-saving float, and the rotation speeds of the two motors can be controlled separately. When the central processing module calculates the direction of the emission source, the central processing module combines the emission source direction and the azimuth information to control the driving module to drive the life-saving float to move toward the emission source direction.

[0015] In the above-mentioned autonomous water rescue device, the central processing module stores coordinate information of a safe position, and the central processing module is electrically connected to an input module, the input module is arranged on the life-saving float, and can generate an input signal to the central processing module according to a user's operation. When the central processing module receives the input signal, the central processing module controls the driving module to drive the life-saving float toward the direction of the safe position, wherein the user's position is the position of the transmitting source.

[0016] In the above-mentioned autonomous water rescue device, the central processing module is further electrically connected to a photography module, and the photography module is arranged on the life-saving float and is used to capture underwater scenes of the water area to output a real-time image to the central processing module. The central processing module identifies a target object that sends out the distress signal in the real-time image, and controls the driving module to drive the life-saving float to move toward the target object.

[0017] The present invention provides another autonomous water rescue system, comprising:

[0018] At least one autonomous water rescue device as described above, wherein the at least one autonomous water rescue device sends an operation status message to the outside;

[0019] a server, connected to the at least one autonomous water rescue device for receiving and storing the operation status information of the at least one autonomous water rescue device; and

[0020] A monitoring device is connected to the server for reading and displaying the operation status information of the at least one autonomous water rescue device.

[0021] The autonomous water rescue device of the present invention floats in a water area, and continuously senses an environmental sound of the water area, and performs sound recognition on the environmental sound to determine whether there is a distress sound in the environmental sound. When the distress sound is included, the direction of a transmitter (a drowning person) that sends the distress sound is calculated based on the distress sound, and the autonomous water rescue device will automatically move in the direction of the transmitter. For the drowning person, he or she can cling to the lifesaving buoy and float on the water area waiting for rescue, so that the drowning person can be rescued within the golden time for rescue, thereby increasing the probability of the drowning person being successfully rescued later. The autonomous water rescue system of the present invention connects the autonomous water rescue device to a monitoring device through a server. The monitoring personnel can continuously check the operating status of the autonomous water rescue device through the monitoring device to determine whether a drowning accident has occurred. When a drowning accident occurs, the autonomous water rescue device immediately goes to the drowning person to seize the golden time for rescue, and the monitoring personnel can learn the location of the autonomous water rescue device through the monitoring device and go there for subsequent rescue. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 : Circuit block diagram of the autonomous water rescue device of the present invention.

[0023] Figure 2 : Schematic diagram of the appearance of the autonomous water rescue device of the present invention.

[0024] Figure 3 : A circuit block diagram of an embodiment of the autonomous water rescue device of the present invention.

[0025] Figure 4 : Flowchart of calculating Mel-frequency cepstral coefficients by the feature extraction unit in the autonomous water rescue device of the present invention.

[0026] Figure 5 : Flowchart of the calculation of the direction of the transmitting source by the azimuth calculation unit in the autonomous water rescue device of the present invention.

[0027] Figure 6 :Circuit block diagram of the autonomous water rescue system of the present invention.

[0028] Figure 7 :Circuit block diagram of the autonomous water rescue system of the present invention, wherein the central processing module includes a wireless communication unit.

[0029] Figure 8 : Schematic diagram of the monitoring interface displayed by the monitoring device in the autonomous water rescue system of the present invention.

[0030] Fig. 9 : Schematic diagram of the monitoring interface displayed by the monitoring device in the autonomous water rescue system of the present invention.

[0031] Fig.10 : Schematic diagram of the autonomous water rescue system of the present invention executing the formation navigation mode. DETAILED DESCRIPTION

[0032] The following is a detailed description of the technical means used by the present invention to achieve the intended purpose of the invention with reference to the drawings and preferred embodiments of the present invention.

[0033] The present invention is an autonomous water rescue device and system, please refer to Figure 1 and Figure 2 First, the autonomous water rescue device 1 is described. The device includes a life-saving float 10, a sound capture module 20, a driving module 30 and a central processing module 40, wherein the sound capture module 20 and the driving module 30 are electrically connected to the central processing module 40 respectively.

[0034] The life-saving float 10 can float in a water area, and the water area can be, for example, a lake, a bathing beach, etc. The life-saving float 10 can be a buoy, a floating board, a life buoy, etc. that can float on water. Taking the life-saving float 10 as a buoy as an example, when a drowning person is found in the water area, the life-saving float 10 can be moved to the vicinity of the drowning person so that the drowning person can grab it and float on the water surface without sinking. How to find the drowning person and how the life-saving float 10 is moved to the vicinity of the drowning person will be further explained later.

[0035] The sound capture module 20 is disposed on the life-saving float 10 to sense an environmental sound S1 of the water area and output the environmental sound to the central processing module 40. Specifically, the sound capture module 20 is an array microphone, and preferably, when the life-saving float 10 floats in the water area, the sound capture module 20 is exposed on the water surface and can receive the environmental sound from various angles. For example, the array microphone may include four sound receiving units, and the array microphone receives sound through the four sound receiving units and can perform preliminary audio processing (for example: noise suppression) on the environmental sound.

[0036] The driving module 30 is disposed on the lifesaving buoy 10 and can be started or shut down according to the control of the central processing module 40. For example, Figure 2As shown, the driving module 30 is two motors arranged at the bottom of the life-saving float 10, and the transmission shafts of the two motors are respectively connected to a fan blade, and preferably, when the life-saving float 10 floats in the water area, the two motors are located below the water surface, and when each of the motors is started, it can respectively drive the fan blades connected thereto to rotate to drive the life-saving float 10 to move, wherein the rotation speeds of the two motors can be controlled separately, for example, when the rotation speeds of the two motors are the same, the life-saving float 10 will move in a straight line, and when the rotation speeds of the two motors are different, the life-saving float 10 will turn left or right.

[0037] The central processing module 40 is disposed on the life-saving float 10. As mentioned above, taking the life-saving float 10 as a floating ball as an example, the central processing module 40 is disposed in a accommodating space inside the floating ball, and when the central processing module 40 receives the environmental audio signal S1 from the sound capture module 20, the central processing module 40 executes program data of a sound recognition model to identify a distress audio signal from the environmental audio signal, and the central processing module 40 calculates a direction of a transmission source of the distress audio signal based on the distress audio signal, and controls the driving module 30 to start, thereby driving the life-saving float 10 (the autonomous water rescue device 1) to move toward the transmission source direction, wherein the central processing module 40 can be an embedded system (Embedded system), such as Nvidia's Jetson Nano module.

[0038] For details on how the central processing module 40 performs sound recognition, please refer to Figure 3 The central processing module 40 further includes an audio separation unit 41 (blind source separation, BSS), a feature capture unit 42, a sound recognition unit 43, an azimuth calculation unit 44 and a storage unit 45, wherein the storage unit 45 can be a storage device such as a hard disk and a memory disposed on the life-saving buoy 10, or can be a network attached storage device (Network Attached Storage, NAS) not disposed on the life-saving buoy 10. As mentioned above, if the central processing module 40 is a Jetson Nano module, then the storage unit 45 is a memory in the Jetson Nano module.

[0039] The storage unit 45 stores the operation data of the audio separation unit 41, the feature extraction unit 42, the sound recognition unit 43 and the azimuth calculation unit 44. The aforementioned units 41-44 have the ability to read and write the storage unit 45. For example, the storage unit 45 stores the program data of the sound recognition model, wherein the sound recognition model is trained in advance using a deep learning algorithm, such as a convolutional neural network algorithm (CNN), a recurrent neural network algorithm (RNN) and other deep learning algorithms. The present invention mainly uses a convolutional neural network algorithm (CNN) to train the sound recognition model, and stores the trained sound recognition model in the storage unit 45 for the sound recognition unit 43 to read and use. The training process of the sound recognition model is common knowledge in the field and will not be described here.

[0040] The audio separation unit 41 is electrically connected to the sound capture module 20 to receive the ambient audio S1 transmitted by the sound capture module 20, and reconstructs the ambient audio S1 to output multiple independent audios S2. Specifically, the ambient audio S1 is formed by the superposition of multiple original audios, and some of the original audios may be fragmented and incomplete. The audio separation unit 41 can calculate the complete content of the original audio according to different scenarios and reconstruct it, and then separate the multiple original audios independently through the independent component analysis principle (ICA) and blind sources separation (BSS) technology to output each independent audio S2.

[0041] The feature extraction unit 42 is electrically connected to the audio separation unit 41 to receive the plurality of independent audio signals S2, and extract features from each of the independent audio signals S2 to output a plurality of feature information S3 of each of the independent audio signals S2. For example, the feature extraction unit 42 may divide each of the independent audio signals S2 into a plurality of analysis segments through a Mel-Frequency Cepstral Coefficients (MFCC) algorithm, and obtain a Mel-Frequency Cepstral of each of the independent audio signals S2 through the algorithm. Each of the Mel-Frequency Cepstral contains a plurality of feature information S3 of each of the independent audio signals S2. The calculation method of the Mel-Frequency Cepstral Coefficients algorithm is as follows: Figure 4 As shown, the following steps are included:

[0042] Step P10: Pre-emphasize each of the independent audio signals S2, that is, calculate each of the independent audio signals S2 through a high-pass filter formula to remove the low-frequency signal in each of the independent signals S2, so as to improve the recognition of the signals in other frequency bands in each of the independent signals S2. The high-pass filter formula is as follows:

[0043] Y(n)=x(n)-ax(n-1)

[0044] Wherein, Y(n) represents an output signal, x(n) is an independent signal S2, and a is a constant between 0.9 and 1.0.

[0045] Step P11: The pre-emphasized processing result is framed and windowed (Cosine window) because the pre-emphasized processing result is still a continuous signal. In order to avoid the signal length being too long and difficult to analyze, frame division is performed. For example, the time length of each frame of audio can be 20 to 30 (microseconds), and windowing is performed to reduce signal leakage caused by frame division. For example, a Hamming window function is used for windowing. The formula is as follows:

[0046]

[0047] Wherein, n=0, 1, 2, 3...N-1, W(n) represents each frame of audio after windowing processing, Y(n) represents the output signal after the preprocessing, and N represents the number of sampling frames.

[0048] Step P12: Perform fast Fourier transform on each frame of audio after windowing to convert each frame of audio into a frequency domain spatial signal, the formula of which is as follows:

[0049]

[0050] Wherein, X(k) represents each of the frequency domain spatial signals, W(n) represents each frame of audio signal after the windowing process, N represents the number of sampling frames, and k is an integer between 0 and N.

[0051] Step P13: Calculate a plurality of logarithmic energies using a Mel filter bank and logarithmic operation, wherein the Mel filter bank includes a plurality of Mel filters, and each of the Mel filters is composed of a plurality of triangular bandpass filters, and the plurality of triangular bandpass filters are evenly distributed on a Mel frequency, and the calculation formula of the logarithmic operation is as follows:

[0052]

[0053] Wherein, L(m) is the logarithmic energy of each, X(k) is the frequency domain spatial signal of each, H m(k) is a frequency response of the Mel filter bank, m is the number of Mel filters, and f is the frequency.

[0054] Step P14: Perform discrete cosine transformation on each of the logarithmic energies to obtain a plurality of Mel-frequency cepstral coefficients, which are calculated as follows:

[0055]

[0056] Where n = 0, 1, 2, 3…L, C m represents each of the Mel-frequency cepstral coefficients, L(m) is each of the logarithmic energies, N is the number of triangular filters, L is an order number, usually an integer between 12 and 16. For example, here L=13 means that the first 13 Mel-frequency cepstral coefficients are captured, and each of the Mel-frequency cepstral coefficients is the multiple feature information S3.

[0057] In addition, more of the plurality of feature information may be calculated by a difference frequency cepstrum calculation formula in step P15, and the calculation formula is as follows:

[0058]

[0059] Where, ΔC m (t) is the change of each Mel-frequency cepstral coefficient with respect to time, τ is the time change, M is a constant, usually between 2 and 3, when M is equal to 2, 26 Mel-frequency cepstral coefficients will be generated, and when M is equal to 3, 39 Mel-frequency cepstral coefficients will be generated.

[0060] It should be noted that the sound recognition unit 43 can only receive two-dimensional information (image), and the environmental audio S1 is one-dimensional information, so the multiple feature information S3 is calculated by the above-mentioned Mel-Frequency Cepstral Coefficients (MFCC) algorithm and input into the sound recognition unit 43.

[0061] When the sound recognition unit 43 receives the multiple feature information S3, the sound recognition unit 43 executes the program data of the sound recognition model pre-trained in the storage unit 45, and classifies the multiple feature information S3 to confirm whether the multiple feature information S3 contains feature information of a distress audio. If the multiple feature information S3 contains the feature information of the distress audio, it means that the ambient sound S1 contains the distress audio. If the multiple feature information S3 does not contain the feature information of the distress audio, it means that the ambient sound S1 does not contain the distress audio.

[0062] As mentioned above, for example, if the sound recognition model is trained by a convolutional neural network (CNN) algorithm, the process of the sound recognition model classifying the plurality of feature information S3 is as follows: by giving a convolution kernel (filter) on each of the feature information S3 (image), each of the convolution kernels continuously moves and scans in each of the feature information S3 to obtain a feature map, which is calculated as follows:

[0063]

[0064] Among them, y is the feature map (expressed in the form of matrix values), W ij is each convolution kernel, usually a k*k matrix, C m is each of the feature information (each of the Mel-frequency cepstral coefficients mentioned above), b is a deviation value, S is an activation function, for example, it can be an S-type function (Sigmoid function), a rectified linear unit function (Rectified linear unit, ReLU), etc.; first, each of the feature information is highlighted by the activation function, and then each of the feature information is compressed through at least one pooling layer to retain important information, at least one fully connected layer converts the content output by the at least one pooling layer into multiple one-dimensional feature vectors, performs a 1×1 convolution operation, and classifies the calculation results to obtain the feature map.

[0065] The sound capture module 20 continuously transmits the environmental audio S1 of the water area to the central processing module 40, so that the audio separation unit 41, the feature capture unit 42 and the sound recognition unit 43 perform operations on the environmental audio S1 to continuously determine whether the environmental audio S1 contains the distress audio. When the environmental audio S1 contains the distress audio, the azimuth calculation unit 44 calculates the direction of a transmission source that sends the distress audio through a time difference algorithm (TDOA).

[0066] Specifically, the sound capture module 20 includes four sound receiving units as described above, and the azimuth angle calculation unit 44 calculates the time difference between the distress audio being sent to any two of the four sound receiving units through the time difference algorithm, and then calculates an angle between the two sound receiving units and the transmitting source through the time difference. Since the sound capture module 20 has a total of four sound receiving units, there are six possible combinations of selecting any two of them. Six angles can be obtained. Figure 5 , the time difference algorithm may include the following calculation steps:

[0067] Step P20: Calculate the cross-correlation (CC) of two audio frequencies input into two radio units. Specifically, the two radio units are a first radio unit and a second radio unit. The transmitting source transmits the distress signal. The first radio unit receives a first audio frequency, and the second radio unit receives a second audio frequency. The following formula is used to calculate the cross-correlation of the first audio frequency and the second audio frequency. The formula is as follows:

[0068]

[0069] Among them, R i (λ) is the correlation between the two audio signals, N is the number of audio frame samples, m1 and m2 are the two audio signals, i is the audio frame number, and λ is the number of displacement sampling points (the number of sampling points that differ between the starting points of two adjacent audio frames); by continuously changing the value of λ to find R i The maximum value of (λ) means finding the maximum value of the cross-correlation between the two audio signals.

[0070] Step P21: Calculate a time difference (TDOA) between the two audio frequencies and the two receiving units. The calculation method is as follows:

[0071] TDOA=λ / f

[0072] Wherein, TDOA is the time difference, λ is the λ when the cross-correlation of the two audio signals is maximum, and f is the sampling frequency.

[0073] Step P22: Calculate the angle between the two sound receiving units and the transmitting source by using the time difference, and the calculation method is as follows:

[0074]

[0075] Among them, m x1 and m x2 is the position of the two sound receiving units, v is the sound transmission speed, and TDOA is the time difference; after the six non-repetitive angles are calculated through the above-mentioned time difference algorithm, the average value of the six angles is taken as the emission source direction of the distress signal.

[0076] Please refer to Figure 3When the central processing module 40 calculates the direction of the emission source, it will output a control signal S4 to the driving module 30. The driving module 30 will be activated according to the control signal S4, and drive the life-saving float 10 (the autonomous water rescue device 1) to move toward the emission source. In one embodiment of the present invention, the central processing module 40 is also electrically connected to a positioning module 50. The positioning module 50 can be, for example, a positioning module supporting the global positioning system (GPS). The positioning module 50 is arranged on the life-saving float 10 and can transmit a positioning information S5 to the central processing module 40 according to the position of the life-saving float 10. The positioning module 50 includes an electronic compass, so the positioning information S5 includes the coordinate information and the azimuth information.

[0077] The coordinate information is used to indicate the current position of the life-saving float 10, and the azimuth information indicates the direction in which the life-saving float 10 is facing. When the central processing module 40 calculates the direction of the emission source, the central processing module 40 combines the emission source direction and the azimuth information to control the driving module 30 to drive the life-saving float 10 to move in a straight line, turn left or turn right to move toward the emission source.

[0078] In addition, the storage unit 45 stores the coordinate information of a safe location, the safe location is located around the water area, and Figure 3 As shown, in one embodiment of the present invention, the central processing module 40 is further connected to an input module 60, and the input module 60 is arranged on the life-saving float 10, and can generate an input signal S6 to the central processing module 40 according to the operation of a user (drowning person). When the central processing module 40 receives the input signal S6, the central processing module 40 controls the driving module 30 to drive the life-saving float 10 to move toward the safe position, wherein the position of the user (drowning person) is the position of the transmitting source; for example, the input module 6 0 is a button on the outside of the life-saving float 10. When the user (drowning person) clings to the life-saving float 10 and presses the button, the water autonomous rescue device 1 will take the drowning person to the safe place. Alternatively, the input module 60 is a weight sensor. When the user (drowning person) clings to the life-saving float 10, the weight sensed by the weight sensor will change and transmit the input signal S6 to the central processing module 40, so that the central processing module 40 controls the water autonomous rescue device 1 to take the user (drowning person) to the safe place.

[0079] Please refer to Figure 2 and Figure 3In one embodiment of the present invention, the central processing module 40 is further electrically connected to a camera module 31, and the camera module 31 is arranged on the life-saving buoy 10. The camera module 31 is used to shoot the underwater picture of the water area to output a real-time image S7 to the central processing module 40. The central processing module 40 identifies a target object that sends the distress signal in the real-time image S7, and controls the driving module 30 to drive the life-saving buoy 10 to move toward the target object. Specifically, when the drowning person (the target object) sinks under the water surface and the sound capture module 20 cannot When receiving the distress signal, the autonomous water rescue device 1 can also use the camera module 31 to take underwater pictures of the water area, so that the central processing module 40 can identify the target object in the real-time image S7 through image recognition technology, and control the driving module 30 to drive the life-saving float 10 to move toward the target object, thereby achieving a more accurate search and rescue effect. Among them, how the central processing module 40 identifies the target object from the real-time image S7 and calculates the distance to the target object through image recognition technology is common knowledge in the field of image processing technology and will not be described in detail here.

[0080] See also Figure 6 The autonomous water rescue system of the present invention comprises at least one autonomous water rescue device 1, a server 70 and a monitoring device 80. The server 70 is communicatively connected to the autonomous water rescue device 1 and the monitoring device 80. For details, please refer to Figure 7 , Figure 7 Only one autonomous water rescue device 1 is drawn in order to facilitate the description of the connection relationship between the autonomous water rescue device 1, the server 70 and the monitoring device 80; in this system, the central processing module 40 also includes a wireless communication unit 46, and the central processing module 40 can communicate with the server 70 through the wireless communication unit 46. For example, the wireless communication unit 46 is a communication unit that supports the Ethernet communication protocol. The server 70 is a core of data access and management, integrating and managing different databases, and the server 70 has functions such as data access. The central processing module 40 transmits an operation status information of the autonomous water rescue device 1 to the server 70 through the wireless communication unit 46. The operation status information, for example, includes information such as the direction of the transmitting source and the azimuth information of the autonomous water rescue device 1 itself. The server 70 stores each information in the operation status information in different databases.

[0081] The monitoring device 80 can be communicatively connected to the server 70 to read the operating status of the autonomous water rescue device 1, and the monitoring device 80 can also control the autonomous water rescue device 1 through the server 70, for example: controlling the autonomous water rescue device to start sensing the environmental audio S1. For example, the monitoring device 80 can be an electronic device such as a mobile phone, a tablet, a computer, etc. For example, the monitoring device 80 is a computer. The monitoring device 80 includes a display. When the monitoring device 80 is connected to the server 70 and performs a monitoring function, the display will display the following information: Figure 8 A monitoring interface is shown, which displays an electronic map 81 and a start identification button 82.

[0082] When the monitoring personnel clicks the start recognition button 82, the sound capture module 20 will start to sense the environmental audio signal S1 and transmit it to the central processing module 40 for information calculation and judgment, and the central processing module 40 will receive the positioning information S5 transmitted by the positioning module 50. Fig. 9 As shown, in the above process, the operating status of the central processing module 40 will be transmitted to the server 70 through the wireless communication unit 46, and the monitoring interface will display the information received by the server 70 in real time. For example, the monitoring interface will display an icon 83 (icon) in the electronic map 81 to indicate the current position of the autonomous water rescue device 1, and a status bar 84 will be displayed on one side of the electronic map 81. The operating status of the autonomous water rescue device 1 is displayed in the status bar 84, such as: the current azimuth (i.e., the azimuth information of the autonomous water rescue device 1), the position of the drowning person (i.e., the direction of the transmitting source), the moving status of the lifesaving float 10, etc.

[0083] In addition, the autonomous water rescue system of the present invention can also execute a formation navigation mode. For example, see Fig.10The at least one autonomous water rescue device 1 includes a pilot rescue device 1A and at least one follower rescue device 1B. The pilot rescue device 1A and the at least one follower rescue device 1B have the same hardware structure and float together in a rescue water area W. When the pilot rescue device 1A calculates the direction of a transmitting source E (a drowning person) according to the received environmental sound, the pilot rescue device 1A transmits the direction information of the transmitting source E to the server 70, and the server 70 transmits the direction information of the transmitting source to the at least one follower rescue device 1B, so that the pilot rescue device 1A and the at least one follower rescue device 1B move toward the transmitting source in the form of a formation. The formation method of the pilot rescue device 1A and the at least one follower rescue device 1B is such as the formation flying method of unmanned aerial vehicle, for example: wedge formation, echelon formation, etc., which is not limited to this. How the formation flying of unmanned aerial vehicle moves forward is common knowledge in the technical field and will not be elaborated here.

[0084] The autonomous water rescue device of the present invention floats in a water area and continuously senses an environmental sound of the water area, and performs sound recognition on the environmental sound to determine whether there is a distress sound in the environmental sound. When the distress sound is included, the direction of a transmitting source (a drowning person) that sends the distress sound is calculated through a time difference algorithm, and the autonomous water rescue device will automatically move in the direction of the transmitting source. The drowning person can cling to the lifesaving buoy and float in the water area waiting for rescue. That is, when a drowning accident occurs, the autonomous water rescue device can immediately go to the drowning person, so that the drowning person can be rescued within the golden time for rescue, thereby increasing the probability of the drowning person being successfully rescued later. The autonomous water rescue system of the present invention connects the autonomous water rescue device to a monitoring device through a server. The monitoring personnel can continuously check the operating status of the autonomous water rescue device through the monitoring device to determine whether a drowning accident has occurred. When a drowning accident occurs, the autonomous water rescue device immediately goes to the drowning person to seize the golden time for rescue, and the monitoring personnel can learn the location of the autonomous water rescue device through the monitoring device and go there for subsequent rescue.

[0085] The above is only a preferred embodiment of the present invention, and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technician familiar with this profession can make some changes or modifications to equivalent embodiments of equivalent changes using the technical contents disclosed above without departing from the scope of the technical solution of the present invention. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention still fall within the scope of the technical solution of the present invention.

Claims

1. An autonomous water rescue device, characterized in that: Include: A life buoy, floating in a water area; a sound acquisition module, disposed on the lifesaving buoy, the sound acquisition module sensing an environmental sound signal of the water area; a driving module, disposed on the lifesaving buoy; and A central processing module is arranged on the life-saving float and electrically connected to the sound capture module and the driving module. The central processing module receives the environmental sound and executes program data of a sound recognition model to recognize a distress audio from the environmental sound. The central processing module calculates the direction of a transmission source according to the distress audio to control the driving module to drive the life-saving float to move toward the transmission source.

2. The autonomous water rescue device according to claim 1, characterized in that: The central processing module includes an audio separation unit and a feature extraction unit. The audio separation unit receives the ambient audio and independently separates multiple independent audios from the ambient audio through independent component analysis principle (ICA) and blind signal separation technology (BSS). The feature extraction unit receives the multiple independent audios and extracts features of the multiple original audios through Mel-frequency cepstral coefficient algorithm (MFCC) to output multiple feature information.

3. The autonomous water rescue device according to claim 2, characterized in that: The central processing module includes a sound recognition unit, which receives the multiple feature information and executes the program data of the sound recognition model to classify the multiple feature information to confirm whether the multiple feature information includes the feature information of the distress audio.

4. The autonomous water rescue device according to claim 1, characterized in that: The central processing module calculates the direction of the source of the distress signal through a time difference algorithm (TDOA); the sound recognition model is trained through a convolutional neural network algorithm (CNN).

5. The autonomous water rescue device according to claim 1, characterized in that: The sound capture module includes four sound receiving units, and the central processing module calculates a time difference between the distress audio being sent to any two of the four sound receiving units, and then calculates an angle between the two sound receiving units and the transmitting source based on the time difference, wherein the direction of the transmitting source is the average of six non-repetitive angles obtained by the central processing module after six calculations.

6. The autonomous water rescue device according to claim 1, characterized in that: The central processing module is further electrically connected to a positioning module, which is arranged on the life-saving float and transmits positioning information to the central processing module according to the position of the life-saving float. The positioning module includes an electronic compass so that the positioning information includes the coordinate information and the azimuth information.

7. The autonomous water rescue device according to claim 6, characterized in that: The driving module is two motors arranged at the bottom of the life-saving float, and the rotation speeds of the two motors can be controlled separately. When the central processing module calculates the direction of the emission source, the central processing module combines the emission source direction and the azimuth information to control the driving module to drive the life-saving float toward the emission source direction.

8. The autonomous water rescue device according to claim 1, characterized in that: The central processing module stores coordinate information of a safe position, and is electrically connected to an input module, which is disposed on the life-saving float and can generate an input signal to the central processing module according to an operation of a user. When the central processing module receives the input signal, the central processing module controls the driving module to drive the life-saving float toward the direction of the safe position, wherein the position of the user is the position of the transmitting source.

9. The autonomous water rescue device according to claim 1, characterized in that: The central processing module is further electrically connected to a camera module, which is disposed on the life-saving float and is used to capture underwater images of the water area to output a real-time image to the central processing module. The central processing module identifies a target object that sends out the distress signal in the real-time image and controls the driving module to drive the life-saving float to move toward the target object.

10. An autonomous water rescue system, characterized in that: Include: At least one autonomous water rescue device according to any one of claims 1 to 9, wherein the at least one autonomous water rescue device sends an operation status information to the outside; a server, connected to the at least one autonomous water rescue device for receiving and storing the operation status information of the at least one autonomous water rescue device; as well as A monitoring device is connected to the server for reading and displaying the operation status information of the at least one autonomous water rescue device.

Images