Anti-falling early warning system and method based on multi-dimensional monitoring

Abstract

The invention provides an anti-falling early warning system and method based on multi-dimensional monitoring, and the system comprises a sensor configuration module which is used for carrying out the integrated deployment of different monitoring types of sensors on a wearable device, and carrying out the real-time monitoring of a walking state of a user according to an integrated deployment result, and obtaining multi-type sensing data; the tumble probability prediction module is used for constructing a tumble risk model, performing association analysis on the multi-category sensing data based on the tumble risk model, and predicting the tumble probability of the user; and the early warning notification module is used for carrying out condition triggering on an early warning mechanism based on the falling probability and carrying out falling early warning response on the user based on a condition triggering result. Therefore, corresponding early warning response can be carried out in time when the hidden danger of falling exists, so that abnormal conditions can be found in advance, and the accuracy of anti-falling early warning is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing and status monitoring, and in particular to a fall prevention warning system and method based on multi-dimensional monitoring. Background Art

[0002] At present, seriously ill or elderly patients often face the risk of falling due to their declining physical functions, poor balance, unclear consciousness and diseases. Falling may not only cause physical injuries, but also cause serious complications, affecting the quality of life of patients. Therefore, fall prevention warning is particularly important;

[0003] However, existing fall warning systems mainly rely on single sensor data, such as accelerometers or pressure sensors, or simply rely on regular patrols by personnel to detect fall events. However, the single sensor method has problems such as high false alarm rate, inability to fully consider the patient's various status information and low prediction accuracy. In addition, it is difficult for regular patrols to achieve real-time monitoring, resulting in failure to detect abnormalities in a timely manner.

[0004] Therefore, in order to overcome the above-mentioned defects, the present invention provides a fall prevention warning system and method based on multi-dimensional monitoring. Summary of the invention

[0005] The present invention provides an anti-fall warning system for critically ill or elderly patients based on multi-sensor data, which is used to integrate and deploy sensors of different monitoring categories on wearable devices, thereby facilitating comprehensive and effective monitoring of the user's walking state through sensors of different monitoring categories, thereby achieving effective acquisition of multi-category sensor data, and secondly, constructing a fall risk model, and performing correlation analysis on the obtained multi-category sensor data through the fall risk model, so as to achieve accurate and reliable prediction of the user's fall probability, providing a reliable reference basis for anti-fall warning, and finally, conditionally triggering the warning mechanism according to the obtained fall probability, so as to facilitate timely corresponding warning response when there is a risk of falling, thereby facilitating early detection of abnormal conditions and improving the accuracy of anti-fall warning.

[0006] The present invention provides a fall prevention warning system based on multi-dimensional monitoring, comprising:

[0007] The sensor configuration module is used to integrate and deploy sensors of different monitoring categories on wearable devices, and monitor the user's walking status in real time based on the integrated deployment results to obtain multi-category sensor data;

[0008] The fall probability prediction module is used to build a fall risk model and perform correlation analysis on multi-category sensor data based on the fall risk model to predict the user's fall probability;

[0009] The warning notification module is used to conditionally trigger the warning mechanism based on the probability of falling, and respond to the user's fall warning based on the conditional triggering result.

[0010] Preferably, in a fall prevention warning system based on multi-dimensional monitoring, sensors of different monitoring categories include: acceleration sensors, gyroscope sensors, pressure sensors and environmental sensors.

[0011] Preferably, a fall prevention warning system based on multi-dimensional monitoring, a sensor configuration module, comprises:

[0012] A sensor configuration unit, used to retrieve fall prevention prediction projects and project requirements based on the knowledge system, and determine the corresponding applicable sensor categories based on the fall prevention prediction projects, and at the same time, determine the project execution parameters of each applicable sensor category based on the project requirements;

[0013] Configure parameters of corresponding applicable sensors based on project execution parameters;

[0014] The integrated deployment unit is used to extract the structural features of sensors of different monitoring categories after parameter configuration, and deploy sensors of different monitoring categories on available wearable devices based on the structural features. At the same time, communication grafting of sensors of different monitoring categories is performed based on the original communication routes of available wearable devices to complete the deployment of sensors of different monitoring categories.

[0015] Preferably, a fall prevention warning system based on multi-dimensional monitoring, a sensor configuration module, comprises:

[0016] An active trigger monitoring unit is used to monitor the continuous swinging motion of the user's legs in real time based on the deployment result, and when the continuous swinging motion of the legs is detected, different monitoring category sensors deployed on the wearable device are started in parallel;

[0017] A monitoring unit, for monitoring the motion parameters of each monitored part of the user's body in real time when the user is walking based on the parallel start result, and obtaining multiple categories of sensor data based on the real-time monitoring result;

[0018] The transmission unit is used to sequentially transmit the obtained multi-category sensor data to the main control terminal based on the communication route.

[0019] Preferably, a fall prevention warning system based on multi-dimensional monitoring, a transmission unit, comprises:

[0020] A receiving subunit, used for receiving the transmitted multi-category sensor data based on the master control terminal, and classifying and caching the received multi-category sensor data;

[0021] The preprocessing subunit is used to:

[0022] Determine the working characteristics of each type of sensor, and determine the data standard state and data deviation focus points of each type of sensor data based on the working characteristics;

[0023] Based on the data standard status and data offset focus, the classification cache results are cleaned and standardized in turn to obtain the final multi-category sensor data.

[0024] Preferably, a fall prevention warning system based on multi-dimensional monitoring, a fall probability prediction module, comprises:

[0025] Data Access Unit, used to:

[0026] Access and log in to the cloud server based on big data, and retrieve daily activity data and fall event data of a target number of different objects from the cloud server based on the access and login results;

[0027] Based on the data structure of daily activity data, the activity postures of different subjects before falling are determined. At the same time, the fall event data is divided into stages based on the stage characteristics of the fall stage, and a stage fall data set is obtained based on the stage division result;

[0028] Perform state traversal on the fall data set at each stage to obtain the state representation of each fall stage, and associate the activity postures of different objects before falling with the state representations of the fall stage at the same time based on the spatiotemporal attributes to obtain the fall behavior characteristics of different objects under different activity postures;

[0029] Model building unit for:

[0030] Initialize the parameters of the neural network based on the model building requirements, and perform iterative cycle training on the neural network after parameter initialization based on the fall behavior characteristics of different objects in different activity postures;

[0031] At the same time, the model loss function is determined based on the model construction requirements, and the parameters of the iterative cycle training process of the neural network are evaluated based on the model loss function. When the model loss condition is met, the fall risk model is obtained;

[0032] Probabilistic prediction unit, used to:

[0033] Based on the fall risk model, a first analysis is performed on the multiple categories of sensor data to obtain a first fall probability under each category of sensor data, and a mutual restriction relationship and an influence weight between the multiple categories of sensor data are determined based on the fall prediction service;

[0034] The first fall probability is integrated based on the mutual limiting relationship and the influence weight to obtain a second fall probability, and a prediction result of the fall probability of the user is obtained based on the second fall probability.

[0035] Preferably, a fall prevention warning system based on multi-dimensional monitoring, a model building unit, comprises:

[0036] A data partitioning subunit is used to perform set partitioning on the fall behavior characteristics of different objects in different activity postures, extract a test data set for the fall risk model, and perform M performance tests on the fall risk model based on the test data set;

[0037] Model update subunit, used to:

[0038] When the number of qualified results in the M performance tests is less than a preset threshold, determining model update data for the fall risk model based on the test results and the model loss function;

[0039] The parameters of the fall risk model are updated based on the model update data.

[0040] Preferably, a fall prevention warning system based on multi-dimensional monitoring, a warning notification module, comprises:

[0041] A judgment unit, used to compare the obtained fall probability with the safety probability, and judge that the user is safe when the fall probability is less than the safety probability, otherwise, judge that the user has a fall risk;

[0042] A conditional trigger unit, used to classify the probability of falling based on the risk classification strategy when there is a risk of falling, and trigger the early warning mechanism based on the risk classification result;

[0043] The response unit is used to determine the fall warning dimensions for the user based on the trigger results and risk classification results, and to initiate the corresponding warning plan according to the fall warning dimensions to respond to the fall warning. The fall warning dimensions include the reminder warning of the device itself, the warning notification of the monitoring terminal, and the warning notification of the medical terminal.

[0044] Preferably, a fall prevention warning system based on multi-dimensional monitoring, the response unit comprises:

[0045] The result acquisition subunit is used to obtain the warning response and fall warning dimension of the user during the monitoring period, and determine the time node corresponding to each warning response and fall warning dimension;

[0046] The recording subunit is used to construct an early warning record report, and to record each early warning response and fall warning dimension based on the development sequence of the time nodes, and generate an early warning record log;

[0047] The storage subunit is used to store the warning record log.

[0048] The present invention provides a fall prevention warning method based on multi-dimensional monitoring, comprising:

[0049] Step 1: Integrate and deploy sensors of different monitoring categories on wearable devices, and monitor the user's walking status in real time based on the integrated deployment results to obtain multi-category sensor data;

[0050] Step 2: Build a fall risk model, and perform correlation analysis on multi-category sensor data based on the fall risk model to predict the user's fall probability;

[0051] Step 3: Conditionally trigger the warning mechanism based on the fall probability, and respond to the user's fall warning based on the conditional triggering result.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] By integrating and deploying sensors of different monitoring categories on wearable devices, it is convenient to comprehensively and effectively monitor the user's walking status through sensors of different monitoring categories, thereby realizing the effective acquisition of multi-category sensor data. Secondly, a fall risk model is constructed, and the obtained multi-category sensor data is correlated and analyzed through the fall risk model to achieve accurate and reliable prediction of the user's fall probability, providing a reliable reference for anti-fall warning. Finally, the warning mechanism is conditionally triggered according to the obtained fall probability, so that when there is a risk of falling, the corresponding warning response can be made in time, thereby facilitating the early detection of abnormal situations and improving the accuracy of anti-fall warning.

[0054] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures specifically pointed out in this application document.

[0055] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0057] Figure 1 It is a structural diagram of a fall prevention warning system based on multi-dimensional monitoring in an embodiment of the present invention;

[0058] Figure 2 It is a structural diagram of a fall probability prediction module in a fall prevention warning system based on multi-dimensional monitoring in an embodiment of the present invention;

[0059] Figure 3 The present invention is a flowchart of a fall prevention warning method based on multi-dimensional monitoring in an embodiment of the present invention. DETAILED DESCRIPTION

[0060] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0061] Embodiment 1:

[0062] This embodiment provides a fall prevention warning system based on multi-dimensional monitoring, such as Figure 1 As shown, including:

[0063] The sensor configuration module is used to integrate and deploy sensors of different monitoring categories on wearable devices, and monitor the user's walking status in real time based on the integrated deployment results to obtain multi-category sensor data;

[0064] The fall probability prediction module is used to build a fall risk model and perform correlation analysis on multi-category sensor data based on the fall risk model to predict the user's fall probability;

[0065] The warning notification module is used to conditionally trigger the warning mechanism based on the probability of falling, and respond to the user's fall warning based on the conditional triggering result.

[0066] In this embodiment, sensors of different monitoring categories include acceleration sensors, gyroscope sensors, pressure sensors and environmental sensors.

[0067] In this embodiment, integrated deployment refers to adapting and installing sensors of different monitoring categories in the wearable device, thereby facilitating real-time monitoring of the user's walking status.

[0068] In this embodiment, the walking state refers to the inclination angle of the user's body, the walking speed, and the pressure values ​​applied to the ground by different parts of the toes.

[0069] In this embodiment, the multi-category sensor data refers to corresponding data obtained after monitoring the user's walking state through different types of monitoring sensors.

[0070] In this embodiment, the fall probability refers to the possibility of the user falling.

[0071] In this embodiment, the early warning mechanism refers to a plan or early warning measure for warning a user of a fall.

[0072] In this embodiment, the conditional trigger means that when the probability of falling meets the set conditions, the corresponding warning operation is performed, and when the set conditions are not met, no warning operation is performed. For example, the set condition is that a warning is performed when the probability of falling exceeds 40%.

[0073] The working principle and beneficial effects of the above technical solution are: by integrating and deploying sensors of different monitoring categories on wearable devices, it is convenient to comprehensively and effectively monitor the user's walking status through sensors of different monitoring categories, thereby realizing the effective acquisition of multi-category sensor data; secondly, a fall risk model is constructed, and the obtained multi-category sensor data is correlated and analyzed through the fall risk model to achieve accurate and reliable prediction of the user's fall probability, providing a reliable reference basis for anti-fall warning; finally, the warning mechanism is conditionally triggered according to the obtained fall probability, so as to facilitate timely corresponding warning response when there is a risk of falling, thereby facilitating the early detection of abnormal situations and improving the accuracy of anti-fall warning.

[0074] Embodiment 2:

[0075] On the basis of Example 1, this embodiment provides an anti-fall warning system based on multi-dimensional monitoring, and sensors of different monitoring categories include: acceleration sensors, gyroscope sensors, pressure sensors and environmental sensors.

[0076] Embodiment 3:

[0077] Based on Example 1, this embodiment provides an anti-fall warning system based on multi-dimensional monitoring, and a sensor configuration module, including:

[0078] A sensor configuration unit, used to retrieve fall prevention prediction projects and project requirements based on the knowledge system, and determine the corresponding applicable sensor categories based on the fall prevention prediction projects, and at the same time, determine the project execution parameters of each applicable sensor category based on the project requirements;

[0079] Configure parameters of corresponding applicable sensors based on project execution parameters;

[0080] The integrated deployment unit is used to extract the structural features of sensors of different monitoring categories after parameter configuration, and deploy sensors of different monitoring categories on available wearable devices based on the structural features. At the same time, communication grafting of sensors of different monitoring categories is performed based on the original communication routes of available wearable devices to complete the deployment of sensors of different monitoring categories.

[0081] In this embodiment, the knowledge system is known in advance, including recording data information related to fall prevention, for example, items or parameter requirements related to fall prevention.

[0082] In this embodiment, the fall prevention prediction project is a service type that needs to be predicted when performing fall prevention prediction, for example, it may be monitoring the walking speed of the user and the current inclination degree of the body.

[0083] In this embodiment, the project requirement refers to the conditions that the fall prevention prediction project needs to meet when monitoring, such as the accuracy of monitoring.

[0084] In this embodiment, the project execution parameters refer to the requirements that each applicable sensor category needs to meet when working, for example, parameters such as refresh frequency and working power.

[0085] In this embodiment, the structural feature refers to the structural form of sensors of different monitoring categories, so that it is convenient to deploy sensors of different monitoring categories on available wearable devices according to the structural features.

[0086] In this embodiment, the original communication route refers to the communication method or communication link that can be used by the wearable device itself.

[0087] In this embodiment, communication grafting refers to configuring the communication links of sensors of different monitoring categories using the original communication routes of the available wearable devices, with the purpose of ensuring that sensors of different monitoring categories can transmit the monitored data.

[0088] The working principle and beneficial effects of the above technical solution are: by determining the anti-fall prediction project and project requirements, the applicable sensor categories and the project execution parameters of each applicable sensor category are accurately and effectively determined; secondly, the corresponding applicable sensors are configured according to the determined project execution parameters; at the same time, the structural characteristics of sensors of different monitoring categories are determined, and the configured sensors are deployed on available wearable devices according to the structural characteristics; finally, the communication reliability of sensors of different monitoring categories is ensured by communicating with the original communication routes of the available wearable devices, and the collected walking data is effectively transmitted, thereby providing reliable guarantee for anti-fall warning.

[0089] Embodiment 4:

[0090] Based on Example 1, this embodiment provides an anti-fall warning system based on multi-dimensional monitoring, and a sensor configuration module, including:

[0091] An active trigger monitoring unit is used to monitor the continuous swinging motion of the user's legs in real time based on the deployment result, and when the continuous swinging motion of the legs is detected, different monitoring category sensors deployed on the wearable device are started in parallel;

[0092] A monitoring unit, for monitoring the motion parameters of each monitored part of the user's body in real time when the user is walking based on the parallel start result, and obtaining multiple categories of sensor data based on the real-time monitoring result;

[0093] The transmission unit is used to sequentially transmit the obtained multi-category sensor data to the main control terminal based on the communication route.

[0094] In this embodiment, the motion parameter refers to the swing amplitude or the tilt angle of each monitored body part when the user walks, so as to determine the walking state data of the user.

[0095] The working principle and beneficial effects of the above technical solution are: through the deployment of sensors of different monitoring categories, the user's walking movements are monitored in real time, and when walking movements are detected, the motion parameters of the monitored parts of the body are determined, thereby effectively determining multiple categories of sensor data. Finally, the obtained multiple categories of sensor data are transmitted to the main control terminal in sequence, providing reliable data for anti-fall warning.

[0096] Embodiment 5:

[0097] Based on Example 4, this embodiment provides an anti-fall warning system based on multi-dimensional monitoring, and the transmission unit includes:

[0098] A receiving subunit, used for receiving the transmitted multi-category sensor data based on the master control terminal, and classifying and caching the received multi-category sensor data;

[0099] The preprocessing subunit is used to:

[0100] Determine the working characteristics of each type of sensor, and determine the data standard state and data deviation focus points of each type of sensor data based on the working characteristics;

[0101] Based on the data standard status and data offset focus, the classification cache results are cleaned and standardized in turn to obtain the final multi-category sensor data.

[0102] In this embodiment, the classified cache refers to storing sensor data of different categories separately.

[0103] In this embodiment, the working characteristics refer to the types of data collected by sensors of different categories and the values ​​of the data defined by each category of sensors under normal circumstances.

[0104] In this embodiment, the data standard state refers to the value range and corresponding data structure of the data corresponding to each category of sensors.

[0105] In this embodiment, the data deviation focus point refers to the data anomaly that needs to be paid special attention to for each category of sensor data, for example, it may be the data value aspect or the data distortion aspect.

[0106] The working principle and beneficial effect of the above technical solution are: by classifying and caching the received multi-category sensor data, and determining the data standard state and data offset focus points of each category of sensor data according to the working characteristics of each category of sensor, the multi-category sensor data in the classified cache are cleaned and standardized according to the determined data standard state and data offset focus points, thereby ensuring the accuracy and reliability of the multi-category sensor data finally obtained.

[0107] Embodiment 6:

[0108] Based on Example 1, this embodiment provides an anti-fall warning system based on multi-dimensional monitoring, such as Figure 2 As shown, the fall probability prediction module includes:

[0109] Data Access Unit, used to:

[0110] Access and log in to the cloud server based on big data, and retrieve daily activity data and fall event data of a target number of different objects from the cloud server based on the access and login results;

[0111] Based on the data structure of daily activity data, the activity postures of different subjects before falling are determined. At the same time, the fall event data is divided into stages based on the stage characteristics of the fall stage, and a stage fall data set is obtained based on the stage division result;

[0112] Perform state traversal on the fall data set at each stage to obtain the state representation of each fall stage, and associate the activity postures of different objects before falling with the state representations of the fall stage at the same time based on the spatiotemporal attributes to obtain the fall behavior characteristics of different objects under different activity postures;

[0113] Model building unit for:

[0114] Initialize the parameters of the neural network based on the model building requirements, and perform iterative cycle training on the neural network after parameter initialization based on the fall behavior characteristics of different objects in different activity postures;

[0115] At the same time, the model loss function is determined based on the model construction requirements, and the parameters of the iterative cycle training process of the neural network are evaluated based on the model loss function. When the model loss condition is met, the fall risk model is obtained;

[0116] Probabilistic prediction unit, used to:

[0117] Based on the fall risk model, a first analysis is performed on the multiple categories of sensor data to obtain a first fall probability under each category of sensor data, and a mutual restriction relationship and an influence weight between the multiple categories of sensor data are determined based on the fall prediction service;

[0118] The first fall probability is integrated based on the mutual limiting relationship and the influence weight to obtain a second fall probability, and a prediction result of the fall probability of the user is obtained based on the second fall probability.

[0119] In this embodiment, the target quantity is set in advance and is used to represent the quantity requirement when retrieving daily activity data and fall event data of different objects, and can be adjusted.

[0120] In this embodiment, the daily activity data refers to data such as the walking speed and walking environment of different subjects every day.

[0121] In this embodiment, the data structure refers to the types of data included in the daily activity data and the association relationship between the data.

[0122] In this embodiment, the activity posture refers to information such as the corresponding action postures of different objects before they fall.

[0123] In this embodiment, the falling stage refers to the stages when different objects fall, including the stages before falling, during falling, and after iteration.

[0124] In this embodiment, the stage feature refers to the characteristics presented in each fall stage, including the corresponding postures and corresponding data information of different objects in different fall stages.

[0125] In this embodiment, the stage fall data set refers to all fall data corresponding to each stage obtained after the fall event data are divided accordingly according to the stage characteristics.

[0126] In this embodiment, the state representation refers to the data type and corresponding value conditions contained in the fall data set at each stage.

[0127] In this embodiment, the spatiotemporal attribute refers to the same environment space and the same time node.

[0128] In this embodiment, the falling behavior characteristics refer to the specific situations corresponding to the falling of different objects in different activity postures, including the specific situation of falling while walking and the specific situation of falling while standing still.

[0129] In this embodiment, the model building requirements are known in advance and are used to characterize the conditions or requirements that need to be met when building the model, such as the value of the accuracy that needs to be achieved or the value of the computing speed that needs to be achieved.

[0130] In this embodiment, the model loss function is a tool for measuring the effectiveness of a trained model, and can be used to measure the loss of data processing accuracy of the trained model.

[0131] In this embodiment, the model loss condition is set in advance.

[0132] In this embodiment, the first fall probability refers to the fall probability obtained after analyzing multiple categories of sensor data through the obtained fall risk model, which is the probability of falling under each category of sensor data considered separately.

[0133] In this embodiment, the mutual limiting relationship refers to the interaction relationship between different types of sensor data, such as the influence of walking speed on the pressure between the foot and the ground.

[0134] In this embodiment, the second fall probability refers to the result obtained by integrating the first fall probability, that is, the probability value that can characterize the overall fall of the user, and is the result obtained by comprehensively considering the relationship between various types of sensor data.

[0135] The working principle and beneficial effects of the above technical solution are: by obtaining daily activity data and fall event data of different objects from the cloud server, and analyzing the obtained daily activity data and fall event data, the fall behavior characteristics of different objects in different activity postures can be accurately and effectively determined; secondly, the neural network is trained according to the determined fall behavior characteristics under different activity postures, and the training process is monitored in real time to ensure the effectiveness and reliability of the final fall risk model; finally, the multi-category sensor data is analyzed through the obtained fall risk model, and finally the prediction result of the user's fall probability can be accurately and effectively determined, so as to facilitate emergency response to the user according to the predicted fall probability, thereby improving the timeliness and accuracy of anti-fall warning.

[0136] Embodiment 7:

[0137] Based on Example 6, this embodiment provides a fall prevention warning system based on multi-dimensional monitoring, and the model building unit includes:

[0138] A data partitioning subunit is used to perform set partitioning on the fall behavior characteristics of different objects in different activity postures, extract a test data set for the fall risk model, and perform M performance tests on the fall risk model based on the test data set;

[0139] Model update subunit, used to:

[0140] When the number of qualified results in the M performance tests is less than a preset threshold, determining model update data for the fall risk model based on the test results and the model loss function;

[0141] The parameters of the fall risk model are updated based on the model update data.

[0142] In this embodiment, the preset threshold is set in advance and is used to measure whether the fall risk model meets the minimum standard required, and can be adjusted.

[0143] In this embodiment, the model update data refers to data that needs to be referenced when updating or optimizing the fall risk model, including the specific structure and degree of optimization of the fall risk model.

[0144] The working principle and beneficial effects of the above technical solution are: by performing M performance tests on the obtained fall risk model, and judging the qualification of the fall risk model based on the results of M performance tests, at the same time, when the qualification requirements are not met, the model update data of the fall risk model is locked according to the test results and the model loss function, and then the parameters of the fall risk model are updated according to the model update data, thereby ensuring the reliability of the final fall risk model and providing a guarantee for anti-fall warning.

[0145] Embodiment 8:

[0146] Based on Example 1, this embodiment provides a fall prevention warning system based on multi-dimensional monitoring, and a warning notification module, including:

[0147] A judgment unit, used to compare the obtained fall probability with the safety probability, and judge that the user is safe when the fall probability is less than the safety probability, otherwise, judge that the user has a fall risk;

[0148] A conditional trigger unit, used to classify the probability of falling based on the risk classification strategy when there is a risk of falling, and trigger the early warning mechanism based on the risk classification result;

[0149] The response unit is used to determine the fall warning dimensions for the user based on the trigger results and risk classification results, and to initiate the corresponding warning plan according to the fall warning dimensions to respond to the fall warning. The fall warning dimensions include the reminder warning of the device itself, the warning notification of the monitoring terminal, and the warning notification of the medical terminal.

[0150] In this embodiment, the safety probability is set in advance and is used as a reference for measuring whether the user falls.

[0151] In this embodiment, the risk grading strategy is graded according to the obtained fall probability, and different fall probability values ​​correspond to different levels.

[0152] In this embodiment, the early warning mechanism refers to a strategy or plan for providing a fall prevention early warning to the user.

[0153] The working principle and beneficial effect of the above technical solution are: by comparing the obtained fall probability with the safety probability, when there is a fall risk, the user's risk is graded according to the determined fall probability, and the early warning mechanism and fall warning dimension are triggered according to the risk classification result, ultimately achieving accurate and effective early warning fall warning response.

[0154] Embodiment 9:

[0155] Based on Example 8, this embodiment provides a fall prevention warning system based on multi-dimensional monitoring, and the response unit includes:

[0156] The result acquisition subunit is used to obtain the warning response and fall warning dimension of the user during the monitoring period, and determine the time node corresponding to each warning response and fall warning dimension;

[0157] The recording subunit is used to construct an early warning record report, and to record each early warning response and fall warning dimension based on the development sequence of the time nodes, and generate an early warning record log;

[0158] The storage subunit is used to store the warning record log.

[0159] In this embodiment, the time node refers to the specific time information corresponding to each warning response and fall warning dimension.

[0160] The working principle and beneficial effect of the above technical solution are: by determining the time node corresponding to each warning response and fall warning dimension, the corresponding warning response and fall warning dimensions are recorded in sequence according to the development order of the time nodes, thereby realizing accurate and effective determination of the warning record log, and finally, the obtained warning record log is stored to facilitate the user's anti-fall warning data to be traced.

[0161] Embodiment 10:

[0162] This embodiment provides a fall prevention warning method based on multi-dimensional monitoring, such as Figure 3 As shown, including:

[0163] Step 1: Integrate and deploy sensors of different monitoring categories on wearable devices, and monitor the user's walking status in real time based on the integrated deployment results to obtain multi-category sensor data;

[0164] Step 2: Build a fall risk model, and perform correlation analysis on multi-category sensor data based on the fall risk model to predict the user's fall probability;

[0165] Step 3: Conditionally trigger the warning mechanism based on the fall probability, and respond to the user's fall warning based on the conditional triggering result.

[0166] The working principle and beneficial effects of the above technical solution are: by integrating and deploying sensors of different monitoring categories on wearable devices, it is convenient to comprehensively and effectively monitor the user's walking status through sensors of different monitoring categories, thereby realizing the effective acquisition of multi-category sensor data; secondly, a fall risk model is constructed, and the obtained multi-category sensor data is correlated and analyzed through the fall risk model to achieve accurate and reliable prediction of the user's fall probability, providing a reliable reference basis for anti-fall warning; finally, the warning mechanism is conditionally triggered according to the obtained fall probability, so as to facilitate timely corresponding warning response when there is a risk of falling, thereby facilitating the early detection of abnormal situations and improving the accuracy of anti-fall warning.

[0167] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A fall prevention warning system based on multi-dimensional monitoring, characterized in that: include: The sensor configuration module is used to integrate and deploy sensors of different monitoring categories on wearable devices, and monitor the user's walking status in real time based on the integrated deployment results to obtain multi-category sensor data; The fall probability prediction module is used to build a fall risk model and perform correlation analysis on multi-category sensor data based on the fall risk model to predict the user's fall probability; The warning notification module is used to conditionally trigger the warning mechanism based on the probability of falling, and respond to the user's fall warning based on the conditional triggering result.

2. The anti-fall warning system based on multi-dimensional monitoring according to claim 1 is characterized in that: Different monitoring categories of sensors include: acceleration sensors, gyroscope sensors, pressure sensors and environmental sensors.

3. The anti-fall warning system based on multi-dimensional monitoring according to claim 1 is characterized in that: Sensor configuration module, including: A sensor configuration unit, used to retrieve fall prevention prediction projects and project requirements based on the knowledge system, and determine the corresponding applicable sensor categories based on the fall prevention prediction projects, and at the same time, determine the project execution parameters of each applicable sensor category based on the project requirements; Configure parameters of corresponding applicable sensors based on project execution parameters; The integrated deployment unit is used to extract the structural features of sensors of different monitoring categories after parameter configuration, and deploy sensors of different monitoring categories on available wearable devices based on the structural features. At the same time, communication grafting of sensors of different monitoring categories is performed based on the original communication routes of available wearable devices to complete the deployment of sensors of different monitoring categories.

4. The anti-fall warning system based on multi-dimensional monitoring according to claim 1 is characterized in that: Sensor configuration module, including: An active trigger monitoring unit is used to monitor the continuous swinging motion of the user's legs in real time based on the deployment result, and when the continuous swinging motion of the legs is detected, different monitoring category sensors deployed on the wearable device are started in parallel; A monitoring unit, for monitoring the motion parameters of each monitored part of the user's body in real time when the user is walking based on the parallel start result, and obtaining multiple categories of sensor data based on the real-time monitoring result; The transmission unit is used to sequentially transmit the obtained multi-category sensor data to the main control terminal based on the communication route.

5. The anti-fall warning system based on multi-dimensional monitoring according to claim 4 is characterized in that: Transmission unit, comprising: A receiving subunit, used for receiving the transmitted multi-category sensor data based on the master control terminal, and classifying and caching the received multi-category sensor data; The preprocessing subunit is used to: Determine the working characteristics of each type of sensor, and determine the data standard state and data deviation focus points of each type of sensor data based on the working characteristics; Based on the data standard status and data offset focus, the classification cache results are cleaned and standardized in turn to obtain the final multi-category sensor data.

6. The anti-fall warning system based on multi-dimensional monitoring according to claim 1, characterized in that: Fall probability prediction module, including: Data Access Unit, used to: Access and log in to the cloud server based on big data, and retrieve daily activity data and fall event data of a target number of different objects from the cloud server based on the access and login results; Based on the data structure of daily activity data, the activity postures of different subjects before falling are determined. At the same time, the fall event data is divided into stages based on the stage characteristics of the fall stage, and a stage fall data set is obtained based on the stage division result; Perform state traversal on the fall data set at each stage to obtain the state representation of each fall stage, and associate the activity postures of different objects before falling with the state representations of the fall stage at the same time based on the spatiotemporal attributes to obtain the fall behavior characteristics of different objects under different activity postures; Model building unit for: Initialize the parameters of the neural network based on the model building requirements, and perform iterative cycle training on the neural network after parameter initialization based on the fall behavior characteristics of different objects in different activity postures; At the same time, the model loss function is determined based on the model construction requirements, and the parameters of the iterative cycle training process of the neural network are evaluated based on the model loss function. When the model loss condition is met, the fall risk model is obtained; Probabilistic prediction unit, used to: Based on the fall risk model, a first analysis is performed on the multiple categories of sensor data to obtain a first fall probability under each category of sensor data, and a mutual restriction relationship and an influence weight between the multiple categories of sensor data are determined based on the fall prediction service; The first fall probability is integrated based on the mutual limiting relationship and the influence weight to obtain a second fall probability, and a prediction result of the fall probability of the user is obtained based on the second fall probability.

7. The anti-fall warning system based on multi-dimensional monitoring according to claim 6, characterized in that: Model building unit, including: A data partitioning subunit is used to perform set partitioning on the fall behavior characteristics of different objects in different activity postures, extract a test data set for the fall risk model, and perform M performance tests on the fall risk model based on the test data set; Model update subunit, used to: When the number of qualified results in the M performance tests is less than a preset threshold, determining model update data for the fall risk model based on the test results and the model loss function; The parameters of the fall risk model are updated based on the model update data.

8. The anti-fall warning system based on multi-dimensional monitoring according to claim 1, characterized in that: Early warning notification module, including: A judgment unit, used to compare the obtained fall probability with the safety probability, and judge that the user is safe when the fall probability is less than the safety probability, otherwise, judge that the user has a fall risk; A conditional trigger unit, used to classify the probability of falling based on the risk classification strategy when there is a risk of falling, and trigger the early warning mechanism based on the risk classification result; The response unit is used to determine the fall warning dimensions for the user based on the trigger results and risk classification results, and to initiate the corresponding warning plan according to the fall warning dimensions to respond to the fall warning. The fall warning dimensions include the reminder warning of the device itself, the warning notification of the monitoring terminal, and the warning notification of the medical terminal.

9. The anti-fall warning system based on multi-dimensional monitoring according to claim 8, characterized in that: Response unit, comprising: The result acquisition subunit is used to obtain the warning response and fall warning dimension of the user during the monitoring period, and determine the time node corresponding to each warning response and fall warning dimension; The recording subunit is used to construct an early warning record report, and to record each early warning response and fall warning dimension based on the development sequence of the time nodes, and generate an early warning record log; The storage subunit is used to store the warning record log.

10. A fall prevention warning method based on multi-dimensional monitoring, characterized in that: include: Step 1: Integrate and deploy sensors of different monitoring categories on wearable devices, and monitor the user's walking status in real time based on the integrated deployment results to obtain multi-category sensor data; Step 2: Build a fall risk model, and perform correlation analysis on multi-category sensor data based on the fall risk model to predict the user's fall probability; Step 3: Conditionally trigger the warning mechanism based on the fall probability, and respond to the user's fall warning based on the conditional triggering result.

Images