Intelligent fire alarm system based on off-line voice recognition

Abstract

The invention relates to an intelligent fire-fighting alarm system based on off-line voice recognition. The intelligent fire-fighting alarm system comprises a fire-fighting alarm device, an alarm response device and a monitoring host, the fire-fighting alarm device comprises alarm panels arranged at doorways of all rooms in a monitoring area range, a micro-control unit A with a voice recognition function is arranged in each alarm panel, and a wireless communication module A is connected to each micro-control unit A; the alarm response device comprises response panels installed on the sides of all fire-fighting facilities, a micro-control unit B is arranged in each response panel, and each micro-control unit B is connected with a wireless communication module B and an alarm module. The invention has the advantages that: 1, the response time is greatly shortened, and the fire behavior can be effectively controlled before the soft fire spreads into the big fire, namely at the initial stage of the fire behavior; 2, the judgment of the fire behavior is more accurate in a mode of recognizing keywords through voice; 3, a voice response reminding mode can guide people to find fire-fighting equipment at the first time and inform people of the exact position of a fire behavior;

Description

Technical Field

[0001] The present invention relates to the field of fire protection, and in particular to an intelligent fire alarm system based on off-line speech recognition. Background Art

[0002] When a fire breaks out in a public place, people often panic and become confused, which not only prevents them from taking emergency measures quickly and effectively, but also prolongs the time it takes to put out the fire. When people try to find firefighting equipment, they often feel confused and disoriented, further complicating the fire response. This problem is particularly prominent in mixed situations, as it exacerbates the chaos at the scene and may delay the fire rescue process. In an emergency, the lack of clear guidance will further complicate the fire response and may cause a small fire to spread into a large fire, exposing lives and property to a higher risk.

[0003] Traditional fire alarm systems have many deficiencies when facing sudden fires, which restricts their effectiveness in emergency situations:

[0004] ① Response lag: Traditional fire detection systems mainly rely on static sensors placed in fixed positions. These sensors first need to detect signs of fire before they can transmit signals to the controller and eventually trigger the corresponding response program. However, when a small fire source occurs, the key fire detection indicators such as temperature and smoke released may not meet the sensor's recognition standards, and the system often cannot respond to the alarm immediately. Only when the fire source expands, such as spreading and spreading into a large fire, can the above indicators meet the sensor's response standards.

[0005] ② Button trigger panel: Traditional manual alarm devices (such as emergency call buttons or manual alarms) can provide accurate alarm instructions through manual operation. Users must manually trigger the alarm, and in a fire emergency, the panic and confusion on the scene may make it impossible to quickly find the relevant alarm panel, delaying the emergency response.

[0006] ③ High false positive rate: Traditional fire protection systems usually use static, fixed-position sensors to detect signs of fire. These sensors may be interfered with by environmental factors, leading to false positives. For example, some systems may trigger alarms when there is an actual fire, mistaking normal activities or environmental changes for fire alarms. Such false alarms not only cause unnecessary emergency responses, but may also cause waste of resources and users' doubts about the reliability of the system.

[0007] ④ Limited fire-fighting facilities: Even if fire extinguishers are used to put out the fire on site immediately, there is still a limited number of fire extinguishers, which makes it impossible to completely extinguish small fires, and people nearby do not know the detailed location of the fire source and fire extinguishers.

[0008] Overall, traditional fire protection systems have a series of problems, such as delayed perception and response speed, difficulty in finding relevant alarm panels, high misjudgment rate of automatic detection, and difficulty in detecting small fires.

[0009] At present, there are also solutions that use AI machine vision to monitor fires on the market. Although machine vision fire monitoring systems can achieve active detection of fire sources to a certain extent, there are still a series of problems:

[0010] ① High cost: Fire monitoring systems based on machine vision usually require expensive hardware and program equipment, including high-resolution cameras, professional image processing units, etc., resulting in high system construction and maintenance costs.

[0011] ② Difficulty in large-scale coverage: Achieving comprehensive and effective machine vision monitoring coverage in large buildings or wide areas faces huge challenges. A large number of monitoring devices need to be deployed, which increases the complexity and difficulty of the system.

[0012] ③ Harsh environmental requirements: The machine vision system has high requirements for environmental conditions and is greatly affected by factors such as light and climate. In complex and changeable outdoor environments or extreme climatic conditions, the stability and reliability of the system may be affected.

[0013] ④ High misjudgment rate: When facing complex scenes, the machine vision system may make misjudgments, such as misjudging non-fire light or dynamic objects as signs of fire, which may lead to false alarms and reduce the reliability of the system.

[0014] ⑤ When a small fire source is obscured, it cannot be effectively identified and detected: Traditional fire detection facilities that rely on temperature and smoke sensors are usually unable to effectively identify and detect small fires or when the fire source is obscured.

[0015] In general, the existing machine vision-based solutions in traditional firefighting systems have a series of challenges and shortcomings in terms of cost, coverage, environmental adaptability and misjudgment rate. Summary of the invention

[0016] The purpose of the present invention is to provide an intelligent fire alarm system based on offline speech recognition to solve the problems of high cost, low coverage, poor environmental adaptability and high misjudgment rate in traditional fire alarm systems.

[0017] The object of the present invention is achieved through the following technical solutions: an intelligent fire alarm system based on offline voice recognition, which includes a fire alarm device, an alarm response device and a monitoring host;

[0018] The fire alarm device comprises an alarm panel arranged at the door of each room within the monitoring area, wherein a micro control unit A with a voice recognition function is arranged in the alarm panel, and a wireless communication module A is connected to the micro control unit A;

[0019] The alarm response device includes a response panel installed beside each fire-fighting facility, in which a microcontroller unit B is provided, and the microcontroller unit B is connected to a wireless communication module B and an alarm module;

[0020] The wireless communication module A and the wireless communication module B are connected through the Internet of Things server. After the micro control unit A recognizes the voice alarm and confirms it, it sends a confirmation alarm message to the Internet of Things server through the wireless communication module A. The Internet of Things server issues a command to the wireless communication module B located within the set range of the wireless communication module A, and controls the alarm module to respond to the alarm through the micro control unit B corresponding to the wireless communication module B.

[0021] Furthermore, the monitoring host includes a host shell, a micro control unit C is arranged in the host shell, and the micro control unit C is connected to a wireless communication module C; the wireless communication module C is respectively connected to the wireless communication module A and the wireless communication module B through the Internet of Things server, and after the micro control unit A recognizes the voice alarm, it sends alarm information to the Internet of Things server through the wireless communication module A, and the wireless communication module C controls the alarm module at a specific location through the Internet of Things server to respond to the alarm.

[0022] Furthermore, the alarm panel is also provided with a voice module (microphone and speaker) A, which is connected to the micro control unit A. After the micro control unit A recognizes the voice alarm, it guides the voice secondary confirmation through the voice module A and sends the suspected alarm information to the Internet of Things server. After the micro control unit A receives the secondary confirmation information, it sends the confirmation alarm information to the Internet of Things server through the wireless communication module A.

[0023] Furthermore, the alarm panel is also provided with a voice module (microphone and speaker) A and a fire alarm button. Both the voice module A and the fire alarm button are connected to the micro control unit A. After the micro control unit A recognizes the voice alarm, it guides the personnel to press the fire alarm button for secondary confirmation through the voice module A and sends suspected alarm information to the Internet of Things server. After receiving the secondary confirmation information, the micro control unit A sends confirmation alarm information to the Internet of Things server through the wireless communication module A.

[0024] Furthermore, the voice module A supports real-time two-way voice intercom (real-time VoLTE voice communication) through the Internet of Things server.

[0025] Furthermore, the voice module A is provided with a recording unit, which records the on-site sound information in real time and stores it in the micro-control unit A. When a suspected alarm or a confirmed alarm is triggered, the alarm panel will upload a recording file of a certain length before the triggering for judging the on-site situation and as auxiliary evidence.

[0026] Furthermore, the voice module A has a speaker, supports receiving response instructions from the Internet of Things server, and can respond to voice broadcast alarm information.

[0027] Furthermore, the alarm panel is also provided with red and blue flashing lights and a digital tube for displaying the device number.

[0028] Furthermore, the alarm module on the alarm response device includes a speaker and red and blue flashing lights.

[0029] Furthermore, the alarm response device supports issuing response instructions directly through the Internet of Things server, and can directly control the response alarm after receiving suspected alarm information (but not confirmed for the second time) and the management personnel reviews the situation and finds it is true.

[0030] Furthermore, the wireless communication module A, the wireless communication module B, and the wireless communication module C are 4G communication modules or Wi-Fi communication modules.

[0031] Compared with the prior art, the advantages of the present invention are:

[0032] 1. The response time is greatly reduced, and the fire can be effectively controlled in the early stage before a small fire spreads into a large fire;

[0033] 2. Make fire situation judgment more accurate by using voice recognition of keywords;

[0034] 3. Voice response reminders can guide people to find fire-fighting facilities in the first place and inform people of the exact location of the fire;

[0035] 4. When a suspected alarm or confirmed alarm is triggered, a recording file of a certain length of time during the triggering period will be uploaded for management personnel to review the on-site situation and as subsequent auxiliary evidence;

[0036] 5. Support real-time two-way voice intercom to more accurately review the on-site situation;

[0037] 6. Low cost, with higher cost performance;

[0038] 7. Easy to cover and maintain;

[0039] 8. Extended functions can be used for emergency situations in multiple occasions. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 The present invention is a schematic diagram of the module composition of an intelligent fire alarm system based on offline speech recognition.

[0041] Figure 2 It is a working principle diagram of the intelligent fire alarm system based on offline speech recognition.

[0042] Figure 3 It is a schematic diagram of the split state of the response panel.

[0043] Figure 4 This is a schematic diagram of the alarm panel's split state.

[0044] Explanation of the numbers: 1 Response panel, 2 Alarm panel. DETAILED DESCRIPTION

[0045] The present invention is described in detail below with reference to the accompanying drawings and embodiments:

[0046] like Figure 1-4 Shown is a schematic diagram of an embodiment of an intelligent fire alarm system based on offline speech recognition provided by the present invention.

[0047] An intelligent fire alarm system based on offline voice recognition, comprising a fire alarm device, an alarm response device and a monitoring host;

[0048] The fire alarm device comprises an alarm panel 2 arranged at the door of each room within the monitoring area, wherein a micro control unit A with a voice recognition function is arranged in the alarm panel 2, and a wireless communication module A is connected to the micro control unit A;

[0049] The alarm panel 2 can be installed in the 86-type electrical box, and its structure includes a middle support and a top cover. A main PCB circuit board is configured inside, on which a core module 32-bit MCU (microcontroller unit A) with voice recognition function and a wireless communication module A responsible for wireless communication are integrated. Various wiring terminals are provided in the edge area to facilitate users to install and remove components more conveniently.

[0050] A fire alarm button is provided in the middle of the front side of the alarm panel 2, red and blue flashing lights are provided on both sides of the fire alarm button, and a digital tube for displaying the equipment number is provided on the upper side of the fire alarm button.

[0051] The alarm response device comprises a response panel 1 installed beside each fire-fighting facility, wherein a microcontroller unit B is provided in the response panel 1, and the microcontroller unit B is connected to a wireless communication module B and an alarm module;

[0052] The alarm module on the alarm response device includes a speaker and red and blue flashing lights.

[0053] The alarm response device supports issuing response instructions directly through the Internet of Things server, and can directly control the response alarm after receiving suspected alarm information (but not confirmed for the second time) and the management personnel reviews the situation to be true.

[0054] The response panel 1 can be installed in the 86 box. Its structure includes a middle layer structure and a top plate. The central area of ​​the top plate is designed with a sound hole to optimize the sound transmission effect of the speaker. In the upper right part of the top plate, the device is equipped with a microphone head and a status indicator. Red and blue flashing lights are set on the left and right sides of the top plate to provide obvious visual indications to match the sound signals.

[0055] The wireless communication module A and the wireless communication module B are connected through the Internet of Things server. After the micro control unit A recognizes the voice alarm and confirms it, it sends a confirmation alarm message to the Internet of Things server through the wireless communication module A. The Internet of Things server issues a command to the wireless communication module B located within the set range of the wireless communication module A, and controls the alarm module to respond to the alarm through the micro control unit B corresponding to the wireless communication module B.

[0056] The IoT server can also issue instructions to the fire alarm device within the set range of the wireless communication module A.

[0057] The monitoring host comprises a host housing, a micro control unit C is arranged in the host housing, and the micro control unit C is connected to a wireless communication module C;

[0058] The right side of the host housing is equipped with a power supply interface to ensure stable operation of the system. The left side of the host housing is equipped with an antenna for the wireless communication module for reliable communication with the IoT server. The central area of ​​the host housing is equipped with a large touch screen, providing users with an intuitive and flexible operation interface. In addition, in the normal working state of power on, the monitoring host displays a clear interface through the display screen, providing users with real-time and comprehensive system information.

[0059] The wireless communication module C is respectively connected to the wireless communication module A and the wireless communication module B through wireless communication with the Internet of Things server. After the micro control unit A recognizes the voice alarm, it sends alarm information to the monitoring host through the Internet of Things server. The wireless communication module C controls the alarm panel and response panel at a specific location to respond to the alarm through the wireless communication network with the Internet of Things server.

[0060] For example, the alarm module at a specific location is controlled by the wireless communication module B to respond to the alarm, and the voice module A is controlled by the wireless communication module A.

[0061] The alarm panel is also provided with a voice module A, which is connected to the micro control unit A. After the micro control unit A recognizes the voice alarm, it guides the voice secondary confirmation through the voice module A and sends the suspected alarm information to the Internet of Things server. After receiving the secondary confirmation information, the micro control unit A sends the confirmation alarm information to the Internet of Things server through the wireless communication module A.

[0062] The alarm panel is also provided with a voice module A and a fire alarm button, which are both connected to the micro control unit A. After the micro control unit A recognizes the voice alarm, it guides the personnel to press the fire alarm button for secondary confirmation through the voice module A and sends suspected alarm information to the Internet of Things server. After receiving the secondary confirmation information, the micro control unit A sends confirmation alarm information to the Internet of Things server through the wireless communication module A.

[0063] The voice module A records the on-site sound information in real time and stores it in the micro control unit A. When a suspected alarm or a confirmed alarm is triggered, the alarm panel will upload a recording file of a certain length before the trigger to judge the on-site situation and as auxiliary evidence. At the same time, the voice module A also supports real-time two-way voice intercom (real-time VoLTE voice communication) through the Internet of Things server.

[0064] The voice module A is equipped with a speaker, supports receiving response instructions from the Internet of Things server, and can respond to voice broadcast alarm information.

[0065] The fire alarm device, alarm response device and monitoring host are all equipped with a 32-bit MCU program burning port for writing voice recognition programs and wireless communication programs. Separate wiring terminals are convenient for program writing and burning, providing a highly flexible operation mode for system development, debugging and upgrading.

[0066] The working principle of the present invention is as follows:

[0067] 1. In case of sudden fire in a manned environment, an intelligent fire alarm system based on offline voice recognition is set up. People can quickly wake up the fire alarm devices installed in conspicuous locations such as entrances and exits. Each fire alarm device has a digital tube displaying the corresponding device number, which can also represent the location information.

[0068] Mode 1: When the fire alarm device recognizes keywords such as "fire" or "fire started", it will repeatedly broadcast "This is xx (equipment number or ID number), please come here as soon as possible and press the button to confirm the fire alarm again." When people press the fire alarm button, it is the second confirmation.

[0069] Mode 2: When the fire alarm device recognizes keywords such as "fire" or "fire started", it will repeatedly broadcast "Do you want to confirm the fire for the second time? If yes, please reply to confirm the fire." When people confirm the fire through voice reply, it is the second confirmation.

[0070] Mode 3: When people directly press the fire alarm button, it is a direct confirmation of the alarm.

[0071] When a suspected alarm is triggered, the alarm panel sends suspected alarm information to the IoT server through wireless communication module A (the alarm panel will upload a recording file of a certain length before the trigger to determine the on-site situation). After receiving the alarm information issued by the IoT server, the management personnel will immediately review the on-site situation based on the on-site recording file or real-time two-way intercom. If a misjudgment occurs, the triggering approximate words will be registered for subsequent model optimization; if there are multiple malicious triggering behaviors, the person will be held accountable and educated according to the law; if a fire does occur, the alarm response command can be manually issued directly through the IoT server.

[0072] After the second confirmation, the alarm panel sends a confirmation alarm message to the Internet of Things server through the wireless communication module A (the alarm panel will upload a recording file of a certain length of time before the trigger to judge the on-site situation and serve as subsequent auxiliary evidence). The alarm panel and response panel within the set range of the wireless communication module A send instructions to the wireless communication module through the Internet of Things server, and the microcontroller unit controls the alarm module to respond to the alarm.

[0073] At the same time, the IoT server sends a command to the wireless communication module C. After receiving the alarm information, the monitoring host will display similar feedback information such as "xxx suspected fire" and "xxx has a fire incident" on the touch screen, accompanied by voice response, so that the staff can understand the fire information at the first time. The staff can adjust the alarm panel and response panel more accurately according to the fire information to issue an alarm reminder. As a result, the various alarm panels and response panels around the fire source respond with voice broadcasts to promptly inform the people around the fire location and the location of the fire equipment, prompting people to bring firefighting equipment to extinguish the fire and prevent small fires from spreading into large fires.

[0074] When a fire is confirmed, the red and blue flashing lights on the fire alarm device and the alarm response device will begin to flash alternately to serve as a warning.

[0075] In an unmanned environment, if a fire breaks out suddenly, when the staff discovers the fire through monitoring patrol, they can immediately use the one-button fire alarm function of the monitoring host to send alarm instructions to various alarm response devices through wireless communication of the Internet of Things server, and then immediately adjust the alarm panels and response panels around the fire point, so that the alarm panels and response panels can remind people of the location of the fire source and the location of the fire-fighting facilities through voice response. This will encourage people to bring fire-fighting facilities to extinguish the fire as soon as possible, so that the fire source can be quickly controlled in the early stage.

[0076] Regarding the expansion of the fire alarm system and the emergency help alarm system, when a series of keywords mainly "help" are identified, the message "Are you sure you want to send a distress signal" will be broadcast repeatedly, and at the same time, a suspected distress message will be sent to the monitoring host. After the rescuer or manager makes a second confirmation, the surrounding alarm panels and response panels will repeatedly play "xxx (place) needs help, please go to support".

Claims

1. An intelligent fire alarm system based on offline speech recognition, characterized in that: It includes fire alarm device, alarm response device and monitoring host; The fire alarm device comprises an alarm panel (2) arranged at the door of each room within the monitoring area, wherein a micro control unit A having a voice recognition function is arranged in the alarm panel (2), and a wireless communication module A is connected to the micro control unit A; The alarm response device comprises a response panel (1) installed beside each fire-fighting facility, wherein a microcontroller unit B is provided in the response panel (1), and the microcontroller unit B is connected to a wireless communication module B and an alarm module; The wireless communication module A and the wireless communication module B are connected to each other through the Internet of Things server. After the micro control unit A recognizes the voice alarm and confirms it, it sends a confirmation alarm message to the Internet of Things server through the wireless communication module A. The Internet of Things server issues a command to the wireless communication module B located within the set range of the wireless communication module A, and controls the alarm module to respond to the alarm through the micro control unit B corresponding to the wireless communication module B.

2. The intelligent fire alarm system based on offline speech recognition according to claim 1 is characterized in that: The monitoring host comprises a host housing, a micro control unit C is arranged in the host housing, and the micro control unit C is connected to a wireless communication module C; The wireless communication module C is connected to the wireless communication module A and the wireless communication module B through the Internet of Things server respectively. After the micro control unit A recognizes the voice alarm, it sends alarm information to the Internet of Things server through the wireless communication module A. The wireless communication module C controls the alarm module at a specific location to respond to the alarm through the Internet of Things server.

3. The intelligent fire alarm system based on offline speech recognition according to claim 1 is characterized in that: The alarm panel (2) is also provided with a voice module A, which is connected to the micro control unit A. After the micro control unit A recognizes the voice alarm, it guides the voice module A to perform a secondary voice confirmation and sends suspected alarm information to the Internet of Things server. After receiving the secondary confirmation information, the micro control unit A sends confirmation alarm information to the Internet of Things server through the wireless communication module A.

4. The intelligent fire alarm system based on offline speech recognition according to claim 1 is characterized in that: The alarm panel (2) is also provided with a voice module A and a fire alarm button, both of which are connected to the micro control unit A. After the micro control unit A recognizes the voice alarm, the voice module A guides the personnel to press the fire alarm button for secondary confirmation and sends suspected alarm information to the Internet of Things server. After receiving the secondary confirmation information, the micro control unit A sends confirmation alarm information to the Internet of Things server through the wireless communication module A.

5. The intelligent fire alarm system based on offline speech recognition according to claim 3 or 4, characterized in that: The voice module A supports real-time two-way voice intercom via the Internet of Things server.

6. The intelligent fire alarm system based on offline speech recognition according to claim 3 or 4, characterized in that: The voice module A is provided with a recording unit, which records the on-site sound information in real time and stores it in the micro control unit A. When a suspected alarm or a confirmed alarm is triggered, the alarm panel (2) uploads a recording file of a certain length before the triggering.

7. The intelligent fire alarm system based on offline speech recognition according to claim 3 or 4, characterized in that: The alarm panel (2) is also provided with red and blue flashing lights and a digital tube for displaying the device number.

8. The intelligent fire alarm system based on offline speech recognition according to claim 1 is characterized in that: The alarm module on the alarm response device includes a speaker and red and blue flashing lights.

9. The intelligent fire alarm system based on offline speech recognition according to claim 2 is characterized in that: The wireless communication module A, the wireless communication module B and the wireless communication module C are 4G communication modules or Wi-Fi communication modules.

Images