Elevator car jumping intelligent voice alarm device and recognition early warning method

Abstract

The invention relates to the technical field of elevator safety monitoring and intelligent early warning, and particularly discloses an elevator car jumping intelligent voice alarm device which comprises a car body, a voice alarm device and a voice alarm device. Vibration sensors are installed at the bottom end and the side wall face of the inner wall face of the lift car body correspondingly, a human body infrared sensor is installed on the top end face of the inner wall of the lift car body, and a full-color prompting lamp is installed on the edge of the human body infrared sensor. According to the elevator car jumping intelligent voice alarm device and the recognition early warning method, the vibration sensor, the human body infrared sensor and the sound sensors uniformly distributed at the corners of the inner side wall are arranged on the car body, so that a multi-dimensional data acquisition structure is formed; through cooperation of filtering noise reduction processing of the signal acquisition module and a built-in convolutional neural network model of the analysis layer processing module, cross validation and accurate analysis of vibration, sound and human body motion data are realized, so that the false alarm rate is reduced, and rapid early warning response is realized after dangerous behaviors occur.

Description

Technical Field

[0001] This invention relates to the field of elevator safety monitoring and intelligent early warning technology, specifically to an intelligent voice alarm device for elevator car movement and a method for identification and early warning. Background Technology

[0002] Currently, elevator cars are generally equipped with alarm devices based on vibration sensing components. These devices are routine auxiliary equipment in the elevator operation safety assurance system. They mainly rely on a single vibration sensor to monitor abnormal vibrations in the car and provide basic alarm notification functions. They have been widely used in the daily safety operation and maintenance of various elevators. However, existing alarm devices have some defects in terms of structure and functional configuration during actual use, such as: Elevator alarm devices rely solely on a single vibration sensor structure, making it difficult to effectively distinguish between normal elevator vibrations and abnormal human-induced jumping behavior. This increases the false alarm rate and overwhelms management personnel. Furthermore, the device's functional structure is limited to passive alarms, lacking supporting structures for active behavior recognition and real-time intervention. It cannot promptly stop dangerous behaviors when they occur, only triggering an alarm after an accident. Moreover, its generic structure is not customized for special groups such as intellectually disabled or deaf children, or for the cognitive characteristics of ordinary children, making it difficult to effectively convey safety warnings. In addition, the device lacks a remote data interaction structure with property management, maintenance, and regulatory departments, only triggering alarms locally within the car or control cabinet, resulting in low emergency response efficiency.

[0003] To address the aforementioned issues, there is an urgent need for innovative designs based on existing elevator car alarm devices. Summary of the Invention

[0004] The purpose of this invention is to provide an intelligent voice alarm device and identification and warning method for elevator car jumping, in order to solve the problems mentioned in the background art. Elevator alarm devices rely solely on a single vibration sensing structure, making it difficult to effectively distinguish between normal elevator vibration and abnormal human jumping behavior, resulting in an increased false alarm rate. At the same time, the device's functional structure is limited to passive alarm, lacking a supporting structure for active behavior recognition and real-time intervention, making it unable to stop dangerous behavior in time. Furthermore, it adopts a general structure and is not customized for special groups such as intellectually disabled and deaf children, as well as ordinary children, making it difficult to effectively convey safety prompts. In addition, the device does not build a remote data interaction structure with property management, maintenance, and regulatory departments, and can only alarm locally in the car or control cabinet, resulting in low emergency response efficiency.

[0005] To achieve the above objectives, the present invention provides the following technical solution: an intelligent voice alarm device for elevator car vibration, comprising a car body, which is the main body of the elevator car; Vibration sensors are installed on the bottom and side walls of the inner wall of the car body, and a human infrared sensor is installed on the top surface of the inner wall of the car body. A full-color indicator light is installed on the edge of the human infrared sensor. A signal acquisition module, a signal communication module, an analysis layer processing module, and an instruction processing module are installed on the inner wall of the car body. The signal acquisition module is electrically connected to the analysis layer processing module. The analysis layer processing module is electrically connected to the instruction processing module and the signal communication module. The instruction processing module is electrically connected to the full-color indicator light, the active buzzer, and the interactive microphone.

[0006] By adopting the above technical solution, the rational layout of multiple components and electrical connection structure enable multi-dimensional data acquisition and rapid command transmission, achieving proactive early warning and real-time intervention, and improving the overall operational reliability of the device.

[0007] Preferably, sound sensors are evenly distributed at the corners of the inner sidewall of the car body.

[0008] By adopting the above technical solution, the sound sensors are evenly distributed, enabling the collection of sound signals in the car without blind spots, supplementing vibration and human movement data, and improving the comprehensiveness and accuracy of abnormal behavior recognition.

[0009] Preferably, the signal acquisition module is installed in the electrical installation box on the top of the car body, and is used to acquire detection signals from the vibration sensor, the human infrared sensor, and the sound sensor.

[0010] By adopting the above technical solution, the signal acquisition module is centrally installed and adapted to multi-sensor signal acquisition, ensuring stable acquisition and preliminary processing of raw data, providing reliable data support for subsequent accurate analysis, and reducing the impact of environmental interference.

[0011] Preferably, the signal communication module adopts a combination structure of RS-485 bus module and Ethernet module.

[0012] The above technical solution, with its dual communication module combination structure, balances stable internal command transmission with external remote data interaction, breaking the limitations of local alarms, enabling real-time uploading of abnormal information, and strengthening multi-departmental collaborative response.

[0013] Preferably, the full-color indicator light and the active buzzer are mounted on the outer surface of the analysis layer processing module.

[0014] By adopting the above technical solution, the active buzzer is installed close to the analysis layer processing module to ensure that the alarm sound is clearly transmitted and is quickly triggered in conjunction with the warning command, thereby enhancing the timeliness and effectiveness of the sound warning.

[0015] Preferably, the analysis layer processing module and the instruction processing module are integrated and installed on the side wall of the car body.

[0016] By adopting the above technical solution, the two core modules are integrated and installed, optimizing the installation space inside the car, shortening the signal transmission path, improving the command response speed, and facilitating later maintenance and repair.

[0017] Preferably, the analysis layer processing module has a built-in convolutional neural network model, which is used to extract and analyze features from the collected vibration, human activity and sound information, and to distinguish between normal elevator vibration and abnormal human jumping behavior.

[0018] By adopting the above technical solution and incorporating a processing module with a built-in CNN model, in-depth analysis of vibration, human activity, and sound information can be achieved, accurately distinguishing between normal vibration and abnormal jumping, and significantly reducing the false alarm rate.

[0019] Preferably, the interactive microphone is installed next to the control panel of the car body, and the interactive microphone is electrically connected to the command processing module to realize voice interactive warning.

[0020] Using the above technical solution, the interactive microphone is installed nearby and electrically connected to the command module to ensure clear voice prompts, realize customized voice warnings and remote voice guidance, and adapt to the needs of different groups of people.

[0021] A method for recognizing and warning of elevator car movement, applied to the aforementioned intelligent voice alarm device for elevator car movement, includes the following steps: Step 1: Data acquisition by the sensing layer: Vibration sensors collect the vibration frequency and amplitude information of the car body, human infrared sensors collect the number of people and the amplitude of their movements in the car, and sound sensors collect the sound intensity and frequency information in the car. Each sensor synchronously transmits the raw information it collects to the signal acquisition module. Step 2: Data preprocessing in the analysis layer: The signal acquisition module processes the raw information through filtering and noise reduction circuits, removes environmental interference, extracts effective feature data, and transmits it to the analysis layer processing module; Step 3: The interaction layer performs behavior recognition and risk classification: The analysis layer processing module analyzes the effective feature data through the built-in CNN model to determine whether it is abnormal human jumping behavior. Then, combined with the vibration amplitude, sound intensity and the intensity of human movement, it classifies the risk into three levels: general warning, severe warning and emergency alarm, and generates corresponding control signals. Step 4: The communication layer executes the warning: The analysis layer processing module transmits the control signal to the instruction processing module, which then controls the full-color indicator light, the active buzzer, and the interactive microphone to activate the corresponding warning mode. Meanwhile, the analysis layer processing module uploads abnormal event information to the property management, maintenance and supervision departments in real time through the signal communication module; If an emergency alarm is triggered, a stop operation command is also sent to the elevator control system.

[0022] By adopting the above technical solutions and closed-loop early warning methods, we can achieve full-chain control from data collection to emergency response, proactively identify risks and respond in a tiered manner, and stop elevators to avoid danger in emergencies, thereby reducing accidents caused by dangerous behaviors.

[0023] Preferably, the warning prompts in the steps include light warnings from full-color indicator lights, buzzer alarms from active buzzers, and voice prompts played through an interactive microphone that are adapted to the cognitive characteristics of intellectually disabled, deaf, and ordinary children.

[0024] By adopting the above technical solution, a multimodal early warning prompt combination is adopted, which is adapted to the cognitive characteristics of intellectually disabled, deaf and mute children and ordinary children, solving the problem of difficult safety prompts and ensuring that all types of people can perceive the early warning information.

[0025] Compared with the prior art, the beneficial effects of the present invention are: the intelligent voice alarm device and identification and warning method for elevator car tremors are as follows: 1. This device forms a multi-dimensional data acquisition structure by equipping the car body with vibration sensors, human infrared sensors, and sound sensors evenly distributed in the corners of the inner wall. Combined with the filtering and noise reduction processing of the signal acquisition module and the convolutional neural network model built into the analysis layer processing module, it achieves cross-validation and accurate analysis of vibration, sound, and human motion data, thereby reducing the false alarm rate. At the same time, the electrical connection structure between the analysis layer processing module and the instruction processing module can quickly drive the full-color indicator light, active buzzer, and interactive microphone to start the warning after abnormal behavior is detected, realizing active real-time intervention. The warning response can be completed within 2 seconds after the occurrence of dangerous behavior, effectively preventing the dangerous behavior from escalating. 2. This device adopts a combination of full-color indicator lights, multi-frequency adjustable active buzzers, and interactive microphones installed next to the operation panel. Through customized child-friendly voice prompts, three-color light tiered warnings, and other multi-modal warning methods, it is suitable for the cognitive habits of ordinary children, and helps to reduce the difficulty for children with intellectual disabilities and deafness to receive single prompts. It ensures that safety prompts can be accurately delivered to various user groups, and improves the effective delivery rate of safety prompts in special scenarios. 3. This device establishes a remote data interaction channel with property management, maintenance, and regulatory departments through a signal communication module and an analysis layer processing module. This enables abnormal event information to be uploaded to the external management platform in real time. At the same time, the linkage structure between the analysis layer processing module and the elevator control system can directly send a stop operation command when an emergency alarm is triggered. Combined with the remote voice access function of the interactive microphone, emergency guidance and rapid response can be achieved, shortening the emergency response time and reducing the risk of safety accidents. Attached Figure Description

[0026] Figure 1 This is a schematic diagram of the installation structure of the device of the present invention; Figure 2 This is a schematic diagram of the communication response device of the present invention; Figure 3 This is a system architecture diagram of the present invention; Figure 4 This is the electrical schematic diagram of the present invention; Figure 5 This is a schematic diagram of network communication according to the present invention.

[0027] In the diagram: 1. Car body; 2. Vibration sensor; 3. Human infrared sensor; 4. Sound sensor; 5. Signal acquisition module; 6. Signal communication module; 7. Full-color indicator light; 8. Analysis layer processing module; 9. Active buzzer; 10. Command processing module; 11. Interactive microphone. Detailed Implementation

[0028] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0029] Please see Figures 1-5 The present invention provides a technical solution: an intelligent voice alarm device and identification and warning method for elevator car vibration, including a car body 1, a vibration sensor 2, a human infrared sensor 3, a sound sensor 4, a signal acquisition module 5, a signal communication module 6, a full-color indicator light 7, an analysis layer processing module 8, an active buzzer 9, an instruction processing module 10, and an interactive microphone 11. In this embodiment, the scenario selected is Chengjia School Special Education School in Jiading District, Shanghai. The school's existing elevator is an old model that has been in service for more than 8 years and is not equipped with a safety protection system for special groups. Considering that the school's students are mainly intellectually disabled, these students have weak safety awareness and frequently engage in dangerous behaviors such as jumping and colliding in the elevator car, which can easily cause safety hazards such as the elevator wire rope jumping out of the groove and the guide rail deformation. Therefore, it is necessary to install the intelligent voice alarm device of this invention on the elevator to realize real-time monitoring, accurate identification and active protection of students' abnormal jumping behavior. The components corresponding to the perception layer, analysis layer, communication layer and interaction layer are installed in the elevator car, and the various parts are connected by circuitry. Sensory layer component installation: Three vibration sensors 2 are installed and fixed to the center of the bottom of the inner wall of the car body 1 and the bottom of the left and right side walls respectively with M4 bolts. The sensor output ends are connected to the ANALOGIN1, ANALOGIN2, and ANALOGIN3 interfaces of the signal acquisition module 5 through twisted shielded wires. The shielding layer is grounded to reduce electromagnetic interference. The human infrared sensor 3 is fixed to the center of the top surface of the inner wall of the car body 1 by a buckle. The lens is tilted downward at 15°, and the detection range completely covers the inside of the car. The sensor pins A and D are signal terminals, which are connected to the ANALOGIN8 interface and GND terminal of the signal acquisition module 5, respectively. B and C are power terminals, which are connected to a 3.3V DC power supply. There are four sound sensors 4 in total. They are attached to the four corners of the inner wall of the car body 1 with strong traceless adhesive, and are evenly distributed without dead corners. The output ends are connected to the ANALOGIN4-ANALOGIN7 interface of the signal acquisition module 5 respectively. Pins B and C are connected to 5V power supply, and A and D are signal output ends. Installation of the analysis layer and instruction processing module: The signal acquisition module 5 is installed in the electrical mounting box on the top of the car body 1. The power interface is connected to the 3.3V-5V DC power supply of the elevator control cabinet, and the output end is electrically connected to the input end of the analysis layer processing module 8 through the SPI data bus. The analysis layer processing module 8 and the instruction processing module 10 are integrated and installed in the integrated mounting box on the side wall of the car body 1. The two are bidirectionally connected through internal wiring. The mounting box is reserved with heat dissipation holes to adapt to the temperature and humidity environment inside the elevator car. Communication layer installation: The core component of the communication layer is the signal communication module 6, which is integrated in the top electrical installation box. It adopts a combination structure of RS-485 bus module and Ethernet module. The RS-485 bus is wired along the inner wall of the elevator shaft, and its two ends are respectively connected to the communication interface of the analysis layer processing module 8 and the control interface of each device in the interaction layer to ensure the stable transmission of control commands. The Ethernet module is integrated in the top control cabinet and connects to the school's local area network via a network cable. It establishes a TCP / IP connection with the property management center server through a router, enabling remote uploading of abnormal data and receiving of remote control commands. Interaction layer component installation: The full-color indicator light 7 is installed at the top center of the car body 1, next to the human infrared sensor 3, and is fixed by a snap-on method to ensure that people inside the car can clearly observe the light status. Its control interface is connected to the RS-485 bus interface of the signal communication module 6 through a wire. The active buzzer 9 and the interactive microphone 11 are installed side by side on the side wall of the car body 1 and fixed to the reserved installation position on the inner wall by bolts. The control interfaces of both are connected to the RS-485 bus interface of the signal communication module 6 to receive the control signals of the instruction processing module 10. The interactive microphone 11 has both voice output and remote voice reception functions. The power supply link consists of: the elevator control cabinet 3.3V-5V DC power supply guide signal acquisition module 5, the guide analysis layer processing module 8, and the instruction processing module 10. Each module is equipped with an independent fuse to prevent short circuit damage. The signal from the sensing layer is transmitted to the signal acquisition module 5. The signal terminals of the vibration sensor 2, human infrared sensor 3, and sound sensor 4 are all connected to the ANALOGIN interface of the signal acquisition module 5 to achieve synchronous acquisition of raw data. The signals from the analysis layer are transmitted to the instruction processing module 10. The control output terminal of the analysis layer processing module 8 is connected to the input terminal of the instruction processing module 10 via a wire to transmit the risk classification control signal.

[0030] The signal from the instruction processing module 10 is transmitted to the interaction layer. The output of the instruction processing module 10 is connected to the control interfaces of the full-color indicator light 7, the active buzzer 9, and the interactive microphone 11 to realize multimodal warning drive.

[0031] The analysis layer transmits signals to the external platform. The analysis layer processing module 8 connects to the network through the Ethernet interface of the signal communication module 6 and establishes a data interaction link with the property, maintenance, and regulatory department terminals. Furthermore, the analysis layer processing module 8 establishes a linkage interface with the elevator control system through the signal communication module 6, using dry contact signal connection. In the event of an emergency alarm, it outputs a high-level signal to trigger the elevator to stop running command.

[0032] The data collected by the device in the elevator car is processed by the signal acquisition module 5 through the RC filter circuit to remove the mechanical noise of elevator operation and environmental interference noise. The sampling frequency is set to 100Hz, and the vibration signal and sound signal are continuously sampled. The original data is denoised by the mean filtering algorithm, and effective feature parameters are extracted, such as the peak value of vibration amplitude, the effective value of sound intensity, and the frequency of human body movements. The data format is converted to 16-bit binary and then transmitted to the analysis layer processing module 8. The processed data enters the CNN model built into the analysis layer processing module 8. This model is trained based on 5,000 sets of diverse sample data, which include: 1,500 sets of elevator vibration during normal operation, 1,000 sets of students jumping slightly, 1,000 sets of students jumping violently, 800 sets of elevator vibration due to malfunction, and 700 sets of environmental interference. The model's input layer has 3 channels: vibration, sound, and human movement. The hidden layer contains 2 convolutional layers and 1 pooling layer. The output layer is the abnormal behavior judgment result. The parameters are optimized using the gradient descent algorithm. The vibration amplitude threshold of 0.3g and the sound intensity threshold of 85dB are set as the abnormal judgment benchmarks. The model's recognition accuracy is ≥95%. The risk classification rules are configured as follows: General warning: Vibration amplitude 0.3g-0.5g and sound intensity 85dB-95dB, human infrared sensor 3 did not detect violent movement amplitude ≤30cm, judged as slight abnormal jumping behavior; Severe warning: Vibration amplitude of 0.5g-0.8g and sound intensity of 95dB-105dB, human infrared sensor 3 detects obvious violent movement amplitude of 30cm-50cm, which is judged as moderate abnormal jumping behavior; Emergency Alarm: If the vibration amplitude is ≥0.8g or the sound intensity is ≥105dB, and the human infrared sensor 3 detects a violent movement amplitude ≥50cm, it is determined to be a severe abnormal jumping behavior, which may cause elevator malfunction.

[0033] The implementation process of the above alarm identification and early warning method is as follows: Step 1: Data Acquisition at the Perception Layer Vibration sensor 2 collects the vibration frequency and amplitude information of the car body 1 in real time; Human infrared sensor 3 collects information on the number of people and the range of their movements in the car, and sound sensor 4 collects information on the sound intensity and frequency in the car. Each sensor collects data synchronously every 10ms and transmits it to the signal acquisition module 5 in real time through shielded wires. Step 2: Data Preprocessing in the Analysis Layer The signal acquisition module 5 filters and reduces noise in the received raw data to remove normal vibration interference such as elevator start-up and braking, as well as environmental background noise. It extracts effective feature data such as vibration amplitude peak, sound intensity effective value, and human movement frequency. The processing delay is ≤50ms. Then, the feature data is transmitted to the analysis layer processing module 8.

[0034] Step 3: Behavior Identification and Risk Classification The analysis layer processing module 8 analyzes the effective feature data using a built-in CNN model and combines the results of multi-sensor cross-validation to determine whether it is a human-induced abnormal jumping behavior. When the general early warning rules are met, a level one control signal is generated; When the severe warning rules are met, a secondary control signal is generated; When the emergency alarm rules are met, a level three control signal is generated.

[0035] Risk classification determination delay ≤300ms, ensuring real-time response.

[0036] Step 4: Execution of communication layer early warning Internal warning execution: The analysis layer processing module 8 transmits the control signal to the instruction processing module 10, and the instruction processing module 10 drives the interaction layer device to start the corresponding warning mode: General warning: The full-color indicator light 7 lights up and the yellow light stays on. The active buzzer 9 emits a low-frequency 500Hz intermittent buzzing sound, which repeats for 1 second and then pauses for 1 second. The interactive microphone 11 plays a child-friendly voice prompt: "Please keep quiet and avoid jumping in the car~", which is suitable for the cognitive characteristics of young students.

[0037] Serious warning: The full-color indicator light 7 switches to orange light and flashes once per second. The active buzzer 9 emits a continuous 1000Hz medium-frequency buzzer. The interactive microphone 11 plays a clear warning voice: "Your behavior has affected elevator safety. Please stop jumping immediately!" At the same time, a sign language animation signal is output to adapt to deaf and mute students. The elevator car needs to be equipped with a display screen.

[0038] Emergency Alarm: The full-color indicator light 7 flashes red light at a frequency of 2 times per second, the active buzzer 9 emits a high-frequency, piercing 1500Hz buzzer, and the interactive microphone 11 plays a loop: "The elevator has experienced a serious malfunction. Please remain calm and wait for rescue by staff!" At the same time, the analysis layer processing module 8 sends a stop operation command to the elevator control system through the signal communication module 6. The elevator immediately decelerates to a stop, keeps the doors closed, and prevents the malfunction from escalating.

[0039] Remote linkage execution: The analysis layer processing module 8 uploads abnormal event information, including event occurrence time, car position, risk level, sensor raw data, and warning type, to the property management center server, maintenance unit monitoring platform, and regulatory department data terminal in real time through the signal communication module 6. The upload delay is ≤1 second. In case of emergency alarm, the property duty room's audible and visual alarm and the maintenance personnel's mobile APP push notification are triggered simultaneously to ensure rapid response. A 30-day continuous test was conducted in the elevators of pilot schools, simulating two typical scenarios each day: 100 instances of students jumping slightly and 100 instances of students jumping violently, for a total of 4,000 tests. During the test, a dedicated person recorded the type of device warning trigger, response time, and false alarms. At the same time, the elevator monitoring system was used to statistically analyze the actual occurrence rate of student jumping behavior and the safety status of elevator operation. The core performance metrics of this embodiment are: Recognition accuracy: In a total of 2000 tests on minor jumping scenarios, general warnings were effectively triggered 1942 times, with an accuracy rate of 97.1%. The system underwent 2,000 tests in violent jumping scenarios, effectively triggering emergency alarms 1,930 times with an accuracy rate of 96.5%; the overall recognition accuracy rate was 96.8%, with no missed events.

[0040] Response time: The average response time from the occurrence of abnormal behavior to the activation of the corresponding early warning mode by the device is 1.8 seconds, and the fastest response time is 1.2 seconds, which meets the requirements for real-time intervention.

[0041] False alarm rate: Only 32 false alarms occurred during the entire test period, with a false alarm rate of 0.8%. The main causes of false alarms were the violent vibrations at the moment of elevator start-up and external impacts on the car. This can be further reduced after algorithm optimization.

[0042] The practical application effect of this embodiment: Intervention effect: Before the test, the average daily occurrence of students jumping in the elevator at the school was about 200 times; after the installation of this device, the average daily occurrence of jumping dropped to 20 times, a reduction of 90%, and students' safety awareness was significantly improved. Safety benefits: Throughout the entire testing period, no elevator malfunctions or entrapments occurred due to students jumping. The elevator failure rate decreased by 80% compared to before the test, the number of elevator repairs decreased from 3 times per month to 0.5 times per month, and the annual maintenance cost decreased by an average of 30%. Scene adaptability: The customized early warning solution for children with intellectual disabilities and deafness has a 92% effective rate in conveying information through child-friendly voice, three-color lights, and sign language animation. The speed at which special students can recognize and respond to early warning information is 60% faster than that of ordinary early warning devices.

[0043] Working principle: When using the elevator car vibration intelligent voice alarm device and identification warning method, data collection is first performed. The vibration sensor 2, human infrared sensor 3, and sound sensor 4 in the sensing layer simultaneously collect vibration signals, abnormal sounds, number of people and movement amplitude data in the car body 1, and transmit them to the signal acquisition module 5 in the analysis layer in real time. Next, data processing is performed. The signal acquisition module 5 filters and reduces noise in the raw data to extract effective features. The analysis layer processing module 8 analyzes the feature data through a CNN model to distinguish between normal vibration and abnormal jumping behavior. Next, risk classification is performed. The analysis layer processing module 8 combines the vibration amplitude collected by the vibration sensor 2, the sound intensity collected by the sound sensor 4, and the intensity of the movement collected by the human infrared sensor 3 to classify the risk into three levels: general warning, severe warning, and emergency alarm, and generates corresponding control commands. Secondly, the signal communication module 6 transmits control commands to the instruction processing module 10 of the interaction layer via RS-485 bus, and uploads abnormal event data to the external management platform via Ethernet. Finally, real-time intervention is carried out. The instruction processing module 10 drives the full-color indicator light 7, the active buzzer 9, and the interactive microphone 11 in the interactive layer to start the corresponding warning mode. The interactive microphone 11 also supports remote voice access to realize emergency guidance.

[0044] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An intelligent voice alarm device for elevator car vibration, including: The car body (1) is the main body of the elevator car; The features are as follows: vibration sensors (2) are installed on the bottom and side walls of the inner wall of the car body (1), and a human infrared sensor (3) is installed on the top surface of the inner wall of the car body (1), and a full-color indicator light (7) is installed on the edge of the human infrared sensor (3). A signal acquisition module (5), a signal communication module (6), an analysis layer processing module (8) and an instruction processing module (10) are installed on the inner wall of the car body (1). The signal acquisition module (5) is electrically connected to the analysis layer processing module (8), the analysis layer processing module (8) is electrically connected to the instruction processing module (10) and the signal communication module (6), and the instruction processing module (10) is electrically connected to the full-color indicator light (7), the active buzzer (9), and the interactive microphone (11).

2. The intelligent voice alarm device for elevator car tremors according to claim 1, characterized in that: Sound sensors (4) are evenly distributed at the corners of the inner wall of the car body (1).

3. The intelligent voice alarm device for elevator car tremors according to claim 1, characterized in that: The signal acquisition module (5) is installed in the electrical installation box on the top of the car body (1) and is used to acquire the detection signals of the vibration sensor (2), the human infrared sensor (3), and the sound sensor (4).

4. The intelligent voice alarm device for elevator car tremors according to claim 1, characterized in that: The signal communication module (6) adopts a combination structure of RS-485 bus module and Ethernet module.

5. The intelligent voice alarm device for elevator car tremors according to claim 1, characterized in that: The active buzzer (9) is mounted on the outer surface of the analysis layer processing module (8).

6. The intelligent voice alarm device for elevator car tremors according to claim 1, characterized in that: The analysis layer processing module (8) and the instruction processing module (10) are integrated and installed on the side wall of the car body (1).

7. The intelligent voice alarm device for elevator car tremors according to claim 1, characterized in that: The analysis layer processing module (8) has a built-in convolutional neural network model, which is used to extract and analyze the features of the collected vibration, human activity and sound information, and to distinguish between the normal operation vibration of the elevator and abnormal human jumping behavior.

8. The intelligent voice alarm device for elevator car tremors according to claim 1, characterized in that: The interactive microphone (11) is installed next to the operation panel of the car body (1), and the interactive microphone (11) is electrically connected to the instruction processing module (10) to realize voice interactive warning.

9. A method for recognizing and warning of elevator car movement, applied to the intelligent voice alarm device for elevator car movement as described in any one of claims 1-8, characterized in that, Includes the following steps: Step 1: The sensing layer collects data: the vibration sensor (2) collects the vibration frequency and amplitude information of the car body (1), the human infrared sensor (3) collects the number of people and the amplitude of their movements in the car, and the sound sensor (4) collects the sound intensity and frequency information in the car. Each sensor transmits the collected raw information synchronously to the signal acquisition module (5). Step 2: Data preprocessing in the analysis layer: The signal acquisition module (5) processes the original information through filtering and noise reduction circuits, removes environmental interference, extracts effective feature data, and transmits it to the analysis layer processing module (8). Step 3: Interaction layer performs behavior recognition and risk classification: The analysis layer processing module (8) analyzes the effective feature data through the built-in CNN model to determine whether it is a human abnormal jumping behavior. Then, combined with the vibration amplitude, sound intensity and intensity of human movement, it classifies the risk into three levels: general warning, severe warning and emergency alarm, and generates corresponding control signals. Step 4: The communication layer executes the warning: The analysis layer processing module (8) transmits the control signal to the instruction processing module (10), and the instruction processing module (10) controls the full-color indicator light (7), the active buzzer (9), and the interactive microphone (11) to start the corresponding warning mode; Meanwhile, the analysis layer processing module (8) uploads abnormal event information to the property, maintenance and supervision departments in real time through the signal communication module (6); If an emergency alarm is triggered, a stop operation command is also sent to the elevator control system.

10. The elevator car sway recognition and early warning method according to claim 9, characterized in that: The warning prompts mentioned in step (3) include light warnings from the full-color indicator light (7), buzzer alarms from the active buzzer (9), and voice prompts played by the interactive microphone (11) that are adapted to the cognitive characteristics of intellectually disabled, deaf and mute children and ordinary children.

Images