Old people accident prevention comprehensive alarm system based on multi-mode artificial intelligence model

Abstract

The invention discloses an old people accident prevention comprehensive alarm system based on a multi-modal artificial intelligence model. The system comprises a data acquisition module, a multi-modal fusion module, an artificial intelligence analysis module, an alarm module and a background management module. The data acquisition module is used for acquiring multi-modal data such as video image data, wearable equipment data and environment data of old people; the multi-modal fusion module carries out fusion processing on the data to generate a multi-modal feature vector; the artificial intelligence analysis module analyzes the feature vectors based on a multi-modal artificial intelligence model, and outputs the probability and confidence of occurrence of various accident risks; when the alarm condition is met, the alarm module is started and sends alarm information to related personnel; the background management module carries out all-around management on the system. Various modal data are fused, the artificial intelligence technology is utilized to realize early warning and timely alarm of various accident risks of the elderly, related personnel can be prompted to take measures in time, the occurrence probability of accidents is reduced, the life safety of the elderly is guaranteed, and the system is suitable for various old-age care institutions and old-age living places. The method has important safety protection value.

Description

[0001] manual Technical Field

[0002] The present invention relates to the technical field of health and safety management for the elderly, and in particular to a comprehensive alarm system for accident prevention for the elderly based on a multimodal artificial intelligence model, which aims to comprehensively prevent and promptly alert the elderly of various accidents that they may encounter, including but not limited to falling from a building, getting lost, choking, and falling. Background Art

[0003] With the increasing aging of society, the safety of the elderly is receiving increasing attention. Due to factors such as declining physical function and cognitive abilities, the elderly face a variety of accidental risks, such as falling, getting lost, choking, and falling. Traditional preventive measures often rely on manual care and simple physical protective devices, which have many limitations, such as high costs and the risk of negligence, and the possibility of accidental movement of physical protective devices. In recent years, although systems have emerged that utilize single-modal data (such as video images or wearable device data) for specific accident prevention, these systems suffer from the following shortcomings: First, their monitoring range is limited, with each system only able to monitor for a specific type of accident; second, they are easily affected by environmental factors or device performance, resulting in false alarms and missed alarms; and third, they are unable to provide comprehensive and integrated assessment and early warning of multiple accident risks for the elderly. Therefore, there is an urgent need for a comprehensive alarm system that can integrate multimodal data and address multiple accident risks to improve the safety of the elderly. Summary of the Invention

[0004] The purpose of the present invention is to provide a comprehensive alarm system for accident prevention for the elderly based on a multimodal artificial intelligence model. By fusing data from multiple modalities and using artificial intelligence technology, early warning and timely alarm of various accident risks for the elderly can be achieved.

[0005] The comprehensive alarm system for preventing accidents among the elderly of the present invention comprises the following modules:

[0006] Data acquisition module: used to collect multiple modal data of the elderly, including but not limited to video image data, wearable device data (such as location data, posture data, heart rate data, swallowing action data and other physiological and motion data) and environmental data (such as indoor window and balcony status data, weather data, indoor noise data, smoke concentration data, temperature data, humidity data, etc.). Video image data can be collected through cameras installed in nursing homes or elderly activity areas to observe the elderly’s behavior, movements, posture, expression and other information as well as the environment they are in; wearable device data is monitored and transmitted in real time by devices worn by the elderly to capture the elderly’s real-time location, posture changes, heart rate and other physiological and motion characteristics; environmental data is obtained through corresponding sensors to help determine whether there are multiple accident risk factors in the elderly’s environment.

[0007] Multimodal Fusion Module: This module fuses collected data from different modalities. First, preprocessing the data, such as cleaning, normalization, and time synchronization, is performed to eliminate discrepancies and noise. Deep learning and other technologies are then used to organically integrate behavioral movements, posture, and facial expressions from video images, position, posture, and physiological characteristics from wearable devices, and various environmental factors from environmental data. This generates a multimodal feature vector that comprehensively reflects the elderly person's condition and various accident risk profiles.

[0008] The AI Analysis Module builds a variety of accident risk identification algorithms based on a multimodal AI model. Using fused multimodal feature vectors as input, the model learns and trains on a large amount of labeled and unlabeled elderly behavioral data (including normal behavior data and data on behaviors exhibiting various accident risks). This enables the model to accurately identify various possible elderly behavior patterns (such as nearing windows or balcony edges, lingering near doorways, abnormal gait, choking, and other characteristic combinations) and outputs the probability and confidence level of each accident risk. The Alarm Module: When the AI Analysis Module determines that the probability of any accident risk reaches the corresponding preset alarm threshold and the confidence level meets the preset requirements, the alarm module is immediately activated. Alarms include local audible and visual alarms to alert nursing home staff and those in the area. Simultaneously, an alarm message is sent via the network to designated guardians, family members, or emergency response agencies. This message may include the elderly person's identity, accident risk type, time and location of occurrence, as well as relevant on-site images or video clips and wearable device data screenshots, enabling rescue personnel to quickly obtain accurate information and take appropriate measures. The Backend Management Module: This module is responsible for configuring, managing, monitoring, and maintaining the entire system. It can set parameters for data acquisition equipment (such as camera shooting parameters and wearable device data collection frequency), adjust multimodal fusion strategies, optimize and update AI models, and set alarm thresholds. Furthermore, it can query, compile, and analyze historical accident risk data and alarm records, providing data support for the safety management of nursing homes and the care arrangements for the elderly.

[0009] The beneficial effects of the present invention are:

[0010] It can integrate multi-modal data on the elderly and environmental data, overcome the shortcomings of single-modal detection, improve the accuracy and reliability of detection of multiple accident risks, and effectively reduce the occurrence of false alarms and missed alarms.

[0011] With the help of advanced artificial intelligence models, early warning of various accident risks for the elderly can be achieved, which helps to detect and take measures in time before danger occurs, reduce the probability of accidents, and protect the lives of the elderly.

[0012] A timely and comprehensive alarm mechanism can ensure that relevant personnel can obtain accurate information in the shortest time possible, quickly carry out rescue and response work, and minimize the damage that may be caused by accidents.

[0013] The system has good scalability and maintainability, and can be flexibly configured and optimized and upgraded according to different scenarios and needs, adapting to the safety protection requirements of various nursing homes and living environments. DETAILED DESCRIPTION

[0014] The following is a detailed description of the comprehensive alarm system for preventing accidents among the elderly according to the present invention.

[0015] The comprehensive alarm system for preventing accidents among the elderly of the present invention is mainly composed of a data acquisition module, a multimodal fusion module, an artificial intelligence analysis module, an alarm module and a background management module.

[0016] In the data acquisition module, cameras are installed in the elderly's rooms, living rooms, corridors, dining rooms, windows, and balconies in the nursing home to ensure comprehensive coverage of key locations, capture real-time video image data of the elderly, and clearly record their behavior, movements, postures, changes in expression, and the environment they are in. At the same time, the wearable devices worn by the elderly are integrated with multiple sensors such as positioning chips, posture sensors, heart rate sensors, and swallowing sensors. They can monitor the elderly's location coordinates (achieved through indoor positioning technologies such as Wi-Fi positioning, Bluetooth positioning, or ultra-wideband positioning), body posture (such as body tilt angle, limb posture, etc.), heart rate, and other physiological and motion data in real time, and transmit the data to the system via wireless networks (such as Bluetooth or Wi-Fi). In addition, window switch sensors and balcony door sensors are installed on the corresponding windows and balcony doors to monitor their opening and closing status; weather sensors are arranged outdoors to collect weather data (such as wind speed, rainfall, etc.); indoor noise sensors, smoke sensors, temperature sensors, humidity sensors, etc. are distributed in different locations indoors to collect environmental data such as environmental noise, smoke concentration, temperature, humidity, etc. These environmental data are collected to the central processing unit of the data acquisition module through wired or wireless communication.

[0017] After receiving various data types, the multimodal fusion module first performs preprocessing operations on the data. For example, video image data undergoes noise removal, resizing, and frame rate unification; wearable device data undergoes data format conversion, unit unification, and missing value imputation; and environmental data undergoes data smoothing. Then, using feature extraction techniques from deep learning algorithms, it extracts behavioral features (such as movement trajectory, climbing movements, and lingering at doorways), posture features (such as forward leaning and balance), and facial expressions (such as anxiety and tension) from the video images. This data is combined with the position, posture, and physiological data (such as abnormal heart rate fluctuations) collected by the wearable device and various environmental data monitored by environmental sensors (such as window opening, inclement weather, abnormal indoor noise, and excessive smoke concentration). Using a specific fusion strategy (such as feature fusion based on a deep neural network), these features are fused into a multimodal feature vector. This feature vector comprehensively reflects the various accident risk conditions faced by the elderly person, and is then fed into the artificial intelligence analysis module. The artificial intelligence analysis module pre-builds and trains a multi-model accident risk identification model based on a multimodal artificial intelligence model. During the model training phase, a large amount of multimodal data, encompassing scenarios of normal elderly behavior and scenarios involving various risky behaviors, is collected and annotated, and the annotated dataset is divided into training, validation, and test sets. The model is trained using the training set, and its parameters are adjusted through an optimization algorithm. This allows the model to learn the characteristic patterns of various modal data when various risky behaviors occur, as well as the correlations between them. This allows the model to accurately distinguish between normal behavior and various risky behaviors. The model is then validated and fine-tuned on the validation set to determine the optimal model structure and hyperparameter settings. During actual runtime, the fused multimodal feature vector is input into the trained model, which then calculates and analyzes the probability of each risky behavior occurring at the current moment, along with the corresponding confidence level.

[0018] When the probability of any accident risk output by the AI analysis module reaches the corresponding preset alarm threshold (such as 80% risk of falling, 75% risk of getting lost, 85% risk of choking, 70% risk of falling, etc.) and the confidence level reaches the preset requirement (such as 70%), the alarm module will immediately activate. The local sound and light alarm device will emit a strong sound and light signal in the area near the elderly person and the monitoring center of the nursing home, alerting on-site personnel and monitoring personnel to the risk of the elderly person. At the same time, the alarm module sends detailed alarm information to preset management personnel (such as nursing staff and security officers of nursing homes), family members, and corresponding emergency rescue agencies (such as falling risk alarms sent to fire departments, choking risk alarms sent to emergency centers, etc.) and other relevant personnel through network communication (such as 4G, 5G networks or wired networks, etc.), including the elderly’s identity information (name, room number, etc.), accident risk type, specific time of occurrence, location (accurate to the location in the room or the specific area in the nursing home), on-site video images (a long video clip can be transmitted in real time so that relevant personnel can view the specific situation of the elderly), screenshots of location and posture data recorded by wearable devices, and environmental data (such as whether the window is open, weather conditions, indoor noise, smoke concentration, etc.), so that relevant personnel can quickly grasp the dangerous situation of the elderly and respond in time, such as nursing staff immediately going to the scene to stop the elderly’s dangerous behavior, rescue personnel rushing to the scene for protection, etc.

[0019] The backend management module provides system administrators with a comprehensive interface for managing data acquisition devices, such as cameras, wearable devices, and environmental sensors. This allows administrators to configure parameters for camera capture parameters (resolution, frame rate, etc.), set the data collection frequency and upload cycle for wearable devices, and configure the sensitivity of window and balcony sensors. Furthermore, the algorithms and strategies used in multimodal fusion can be adjusted and optimized to meet the diverse accident risk monitoring needs of different nursing home environments and the behavioral characteristics of the elderly. The module also supports further training and optimization of AI models, such as incrementally training the models with new behavioral data, to continuously improve their recognition accuracy and generalization capabilities. Different alarm thresholds and confidence levels can be set in the backend management module, allowing flexible adjustment of alarm sensitivity based on actual needs and experience. The system also automatically records all alarm events and related data. Administrators can conveniently query historical alarm records through the backend management module and conduct statistical analysis on the frequency, time distribution, and location distribution of various accident risks. This provides strong data support for the security deployment of nursing homes and the care and monitoring of elderly residents, helping to formulate more scientific and reasonable safety protection strategies and measures.

[0020] To sum up, the comprehensive alarm system for accident prevention for the elderly based on a multimodal artificial intelligence model of the present invention can realize early warning and timely alarm of various accident risks for the elderly by integrating multi-modal elderly data and environmental data, and using advanced artificial intelligence technology for analysis and identification. It is of great significance to comprehensively improve the safety level of the elderly in nursing homes and prevent various types of accidents. It has broad application prospects and can be widely used in various nursing homes, senior apartments, and homes with elderly people.

Claims

1. A comprehensive alarm system for accident prevention for the elderly based on a multimodal artificial intelligence model, characterized in that: include: Data acquisition module, used to collect video image data, wearable device data and environmental data of the elderly; The multimodal fusion module is used to fuse the collected data of different modalities and generate a multimodal feature vector; The AI analysis module builds multiple accident risk identification algorithms based on multimodal AI models, takes the fused multimodal feature vector as input, and outputs the probability and confidence level of multiple accident risks. The alarm module activates the alarm mechanism and sends an alarm message to relevant personnel when the probability of any accident risk reaches the corresponding preset alarm threshold and the confidence level meets the preset requirements. The backend management module is used to configure, manage and monitor the system.

2. The comprehensive alarm system for accident prevention for the elderly based on a multimodal artificial intelligence model according to claim 1 is characterized in that: The data acquisition module includes: a camera installed in the elderly activity area of the nursing home, used to capture video image data of the elderly; a wearable device worn by the elderly, used to monitor the elderly's position data, posture data, heart rate data, swallowing movement data and other physiological and motion data in real time; environmental sensors, including window switch sensors, balcony door sensors, weather sensors, indoor noise sensors, smoke sensors, temperature sensors, humidity sensors, etc., which are used to monitor corresponding environmental factors and transmit the monitoring data to the central processing unit of the data acquisition module.

3. The comprehensive alarm system for accident prevention for the elderly based on a multimodal artificial intelligence model according to claim 1 is characterized in that: The method for the multimodal fusion module to perform data fusion processing includes: preprocessing various types of data, such as data cleaning, normalization and time synchronization; extracting the behavioral action characteristics, posture characteristics and expression characteristics of the elderly from video image data; extracting position characteristics, posture characteristics and physiological characteristics from wearable device data; extracting various environmental factor characteristics from environmental data; and using a deep learning algorithm to fuse the features of the above different modalities to generate a multimodal feature vector.

4. The comprehensive alarm system for accident prevention for the elderly based on a multimodal artificial intelligence model according to claim 1 is characterized in that: The multimodal artificial intelligence model of the artificial intelligence analysis module adopts a deep neural network architecture and is trained with a large amount of labeled and unlabeled elderly behavior data. It can accurately identify the behavioral patterns of the elderly who are at risk of various accidents.

5. The comprehensive alarm system for accident prevention for the elderly based on a multimodal artificial intelligence model according to claim 1 is characterized in that: The alarm modes of the alarm module include local sound and light alarm and network alarm. The local sound and light alarm is used to remind the staff and on-site personnel in the nursing home. The network alarm sends alarm information containing the elderly person's identity information, accident risk type, time and location of occurrence, as well as on-site images or video clips and wearable device data screenshots to the preset management personnel, family members or emergency rescue agencies through the communication network.

6. The comprehensive alarm system for accident prevention for the elderly based on a multimodal artificial intelligence model according to claim 1 is characterized in that: The background management module has the following functions: parameter setting of data acquisition equipment; adjustment of multimodal fusion algorithms and strategies; optimization and updating of artificial intelligence models; setting of alarm thresholds and confidence requirements; query, statistics and analysis of historical accident risk data and alarm records.