Event information pushing system and method based on intelligent riding helmet

Abstract

The invention relates to the technical field of intelligent helmets, and particularly discloses an event information pushing system and method based on an intelligent riding helmet, the intelligent helmet is used for monitoring environment state data and comfort condition data during riding of a rider in real time, and a time sequence analysis technology based on deep learning is introduced to carry out time sequence modeling analysis on the environment state data and the comfort condition data; the method comprises the following steps of: extracting time sequence characteristic representation of an environment state and a human body state according to an environment condition characteristic and a human body state characteristic, and further performing interactive response analysis based on a bidirectional attention mechanism on the environment state characteristic and the human body state characteristic to mine potential influence of an environment condition on the human body state in a riding process of a rider and adaptive change of the human body state on the environment condition; therefore, the riding risk index under the current condition is intelligently estimated, and early warning notification of the corresponding risk level is carried out based on the riding risk index. Through the mode, the riding risk can be evaluated more accurately, so that the accuracy and timeliness of early warning notification are improved, and the riding safety of a rider is ensured.

Description

Technical Field

[0001] The present application relates to the technical field of smart helmets, and more specifically, to an event information push system and method based on a smart cycling helmet. Background Art

[0002] With the rapid development of smart wearable devices, smart helmets, as an important tool to enhance the riding experience and ensure riding safety, have gradually integrated more advanced sensors and technologies to provide more comprehensive safety protection and personalized services. For example, the invention patent with publication number CN117579691A proposes a method for pushing event information based on smart helmets, which collects the environmental and physiological data of riders in real time by placing sensors such as light, humidity, wind speed, and heart rate and respiratory rate monitoring equipment inside the smart helmet, and uses these data to evaluate the safety and comfort of the current ride, calculate the riding risk index, and issue different levels of early warning notifications according to the riding risk index, such as continuous vibration or voice prompts, to help riders respond to potential risks in a timely manner.

[0003] However, in the prior art, the calculation of the cycling risk index mainly relies on a static data analysis method, that is, a simple weighted calculation is performed on the environmental and physiological data at a single point in time. In fact, during the riding process, the impact of environmental conditions on the rider's physical state changes dynamically, and the rider's physical state will also adjust its ability to adapt to the environment as the environmental conditions change. The prior art lacks in-depth analysis of the dynamic interaction between human body state and environmental parameters, which may lead to inaccurate assessment of the cycling risk index, especially in sudden environmental changes or complex riding scenarios.

[0004] Therefore, an optimized event information push system and method based on a smart cycling helmet is expected. Summary of the invention

[0005] In order to solve the above technical problems, the present application is proposed. The embodiment of the present application provides an event information push system and method based on a smart cycling helmet, which uses sensors deployed inside the smart helmet to monitor the relevant environmental state data information and relevant comfort situation data information when the rider is riding, and introduces a time series analysis technology based on deep learning to perform time series modeling analysis on the relevant environmental state data information and the relevant comfort situation data information, so as to extract the time series feature representation of the environmental state and the human body state, and then, by performing an interactive response analysis of the environmental state characteristics and the human body state characteristics based on a two-way attention mechanism, the potential impact of environmental conditions on the human body state during the rider's riding process and the adaptive changes of the human body state to the environmental conditions are excavated, so as to intelligently estimate the riding risk index under the current situation, and based on this, make early warning notifications of corresponding risk levels. In this way, a more accurate and dynamic assessment of riding risks can be achieved to improve the accuracy and timeliness of early warning notifications and ensure the riding safety of riders.

[0006] According to one aspect of the present application, a method for pushing event information based on a smart cycling helmet is provided, which includes: using several groups of sensors deployed inside the smart helmet to monitor relevant environmental status data information and relevant comfort situation data information when the rider is riding; based on the relevant environmental status data information and the relevant comfort situation information, calculating the cycling risk index; based on a preset evaluation threshold of the cycling risk index, to obtain a level event push warning.

[0007] In the above-mentioned event information push method based on the smart cycling helmet, the cycling risk index is calculated based on the relevant environmental state data information and the relevant comfort situation information, including: performing multivariate time series joint feature extraction on the relevant environmental state data information and the relevant comfort situation information respectively to obtain an environmental parameter time series multivariate joint representation vector and a human body state time series multivariate joint representation vector; performing interactive response analysis based on two-way attention on the environmental parameter time series multivariate joint representation vector and the human body state time series multivariate joint representation vector to obtain a human body state-environmental parameter time series fine-grained interactive response coding vector; based on the human body state-environmental parameter time series fine-grained interactive response coding vector, determining the cycling risk index.

[0008] According to another aspect of the present application, an event information push system based on a smart cycling helmet is provided, which includes: a multivariate time series feature extraction module, which is used to perform multivariate time series joint feature extraction on relevant environmental state data information and relevant comfort condition information respectively to obtain an environmental parameter time series multivariate joint representation vector and a human body state time series multivariate joint representation vector; an interactive response analysis module, which is used to perform interactive response analysis based on two-way attention on the environmental parameter time series multivariate joint representation vector and the human body state time series multivariate joint representation vector to obtain a human body state-environmental parameter time series fine-grained interactive response coding vector; a cycling risk assessment module, which is used to determine a cycling risk index based on the human body state-environmental parameter time series fine-grained interactive response coding vector.

[0009] Compared with the prior art, the event information push system and method based on the smart cycling helmet provided by the present application utilizes sensors deployed inside the smart helmet to monitor the relevant environmental state data information and relevant comfort situation data information when the rider is riding, and introduces the time series analysis technology based on deep learning to perform time series modeling analysis on the relevant environmental state data information and relevant comfort situation data information, so as to extract the time series feature representation of the environmental state and the human body state, and then, by performing interactive response analysis on the environmental state characteristics and the human body state characteristics based on the two-way attention mechanism, the potential impact of environmental conditions on the human body state during the rider's riding process and the adaptive changes of the human body state to the environmental conditions are excavated, so as to intelligently estimate the riding risk index under the current situation, and based on this, make early warning notifications of the corresponding risk levels. In this way, a more accurate and dynamic assessment of riding risks can be achieved to improve the accuracy and timeliness of early warning notifications and ensure the riding safety of riders. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0011] Figure 1 The present invention is a flowchart of a method for pushing event information based on a smart cycling helmet according to an embodiment of the present application.

[0012] Figure 2 This is a data flow diagram of the event information push method based on the smart cycling helmet according to an embodiment of the present application.

[0013] Figure 3This is a flowchart of sub-step S1 of the event information push method based on a smart cycling helmet according to an embodiment of the present application.

[0014] Figure 4 This is a flowchart of sub-step S2 of the event information push method based on a smart cycling helmet according to an embodiment of the present application.

[0015] Figure 5 This is a flowchart of sub-step S21 of the event information push method based on a smart cycling helmet according to an embodiment of the present application.

[0016] Figure 6 This is a flowchart of sub-step S23 of the event information push method based on a smart cycling helmet according to an embodiment of the present application.

[0017] Figure 7 This is a block diagram of an event information push system based on a smart cycling helmet according to an embodiment of the present application. DETAILED DESCRIPTION

[0018] As shown in this application and claims, unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "include" and "comprise" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0019] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules can be used and run on the user terminal and / or server. The modules are only illustrative, and different aspects of the system and method can use different modules.

[0020] Flowcharts are used in the present application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps may be processed in reverse order or simultaneously as required. Meanwhile, other operations may also be added to these processes, or a certain step or several steps of operations may be removed from these processes.

[0021] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0022] It is worth noting that in this application, all actions to obtain data are carried out in compliance with the relevant data protection laws and policies of the country where the data is located, and with the authorization given by the owner of the corresponding device.

[0023] As mentioned in the background technology above, patent CN117579691A proposes an event information push method based on a smart helmet, which includes: using several groups of sensors deployed inside the smart helmet to monitor relevant environmental status data information and relevant comfort situation data information when the rider is riding; based on the relevant environmental status data information and the relevant comfort situation information, calculating the riding risk index; based on the preset evaluation threshold of the riding risk index, to obtain a level event push warning.

[0024] The above-mentioned event information push method based on smart helmets collects the environmental and physiological data of the rider in real time while riding by placing sensors such as light, humidity, wind speed, and heart rate and respiratory rate monitoring equipment inside the smart helmet, and uses these data to evaluate the safety and comfort of the current ride, calculate the riding risk index, and issue different levels of early warning notifications according to the riding risk index, such as continuous vibration or voice prompts, to help riders respond to potential risks in a timely manner. However, the existing technology mainly relies on simple weighted calculations of environmental and physiological data at a single point in time when calculating the riding risk index, ignoring the dynamic interactive relationship between the environment and the human body state, resulting in inaccurate evaluation, especially in sudden environmental changes or complex scenarios. In response to this technical problem, this application proposes an optimized event information push method based on a smart cycling helmet, which uses sensors deployed inside the smart helmet to monitor the relevant environmental state data information and relevant comfort data information when the rider is riding, and introduces a time series analysis technology based on deep learning to perform time series modeling analysis on the relevant environmental state data information and relevant comfort data information, so as to extract the time series feature representation of the environmental state and the human body state, and then, through the interactive response analysis of the environmental state characteristics and the human body state characteristics based on the two-way attention mechanism, the potential impact of environmental conditions on the human body state during the rider's riding process and the adaptive changes of the human body state to the environmental conditions are excavated, so as to intelligently estimate the riding risk index under the current situation, and based on this, make early warning notifications of the corresponding risk levels. In this way, a more accurate and dynamic assessment of riding risks can be achieved to improve the accuracy and timeliness of early warning notifications and ensure the riding safety of riders.

[0025] Specifically, considering that single-point environmental status data and comfort status data cannot fully reflect the actual status of the rider during riding, in the technical solution of the present application, firstly, several groups of sensors deployed inside the smart helmet are used to collect the time queue of the air humidity value, the time queue of the particle density and the time queue of the wind speed value when the rider is riding, to construct relevant environmental status data information. At the same time, by collecting the time queue of the rider's heart rate value and the time queue of the breathing frequency, relevant comfort status data information is constructed, thereby forming a comprehensive monitoring of the human body status and environmental parameters, providing a more comprehensive data basis for subsequent riding risk analysis.

[0026] In the specific implementation, it is first necessary to select the appropriate sensor type according to the environmental variables and human physiological reactions that may be encountered during riding. For the acquisition of environmental status data information, particle density is an important factor affecting the respiratory health of cyclists, so a laser scattering particle sensor can be used to monitor the particle density in the air; the humidity sensor is used to detect the air humidity to assess whether the riding environment is suitable for long-term outdoor activities; the wind speed sensor can be installed at an appropriate position on the outside of the helmet to help predict possible climate changes by measuring wind speed changes. For comfort data information, the heart rate belt is traditional but highly accurate, but considering the wearing comfort, non-contact heart rate monitoring technology can be used, such as optical blood oxygen saturation monitoring (PPG); the breathing rate can be obtained through a chest expansion sensor or a sound-based breathing monitoring system.

[0027] It should be understood that a reasonable sensor layout not only helps to reduce signal interference, but also improves the effectiveness of data collection. For example, environmental sensors should be placed in a position where they can directly sense the external environment, such as near the top of the helmet or the side vents, so that they can directly sense changes in particle density, humidity and wind speed. Sensors related to human body status should be as close to the surface of the head skin as possible to ensure that the most realistic physiological signals such as heart rate and breathing rate can be captured, but at the same time, the comfort of the wearer should also be considered to avoid causing unnecessary pressure or discomfort.

[0028] Next, consider the need to connect these scattered components to form a complete monitoring network. In terms of physical connection, all sensors are required to be firmly attached to the helmet structure and will not move even during strenuous exercise. This usually involves customized bracket design, the application of high-strength adhesives, and precision machining of sensor installation positions. Electrical connection refers to establishing a communication link from each sensor to the central processing unit (CPU), including the digital conversion of analog signals, the layout of signal lines, and the design of power supply lines. To this end, sufficient wiring space must be reserved inside the helmet, and technical means such as flexible circuit boards must be used to reduce the impact of wires on the wearer. In addition, the power supply solution must be considered to ensure that each sensor can work stably without data loss due to insufficient power.

[0029] After completing the design of the hardware part, software support is also indispensable. Each sensor requires a matching driver to initialize and control operations, and requires an accurate calibration algorithm to correct errors caused by manufacturing deviations or environmental factors. In order to ensure the consistency and reliability of data collection, it is also necessary to build an efficient data synchronization mechanism so that different types of sensors can sample simultaneously within the specified time interval, thereby generating a coherent and valuable data sequence.

[0030] Figure 1 The present invention is a flowchart of a method for pushing event information based on a smart cycling helmet according to an embodiment of the present application. Figure 2 FIG. 1 is a data flow diagram of a method for pushing event information based on a smart cycling helmet according to an embodiment of the present application. Figure 1 and Figure 2 As shown, the event information push method based on the smart cycling helmet includes the steps of: S1, performing multivariate time series joint feature extraction on the relevant environmental state data information and the relevant comfort condition information to obtain an environmental parameter time series multivariate joint representation vector and a human body state time series multivariate joint representation vector; S2, performing interactive response analysis based on two-way attention on the environmental parameter time series multivariate joint representation vector and the human body state time series multivariate joint representation vector to obtain a human body state-environmental parameter time series fine-grained interactive response coding vector; S3, determining the cycling risk index based on the human body state-environmental parameter time series fine-grained interactive response coding vector.

[0031] In the above-mentioned event information push method based on smart cycling helmets, the step S1 performs multivariate time series joint feature extraction on the relevant environmental state data information and the relevant comfort condition information respectively to obtain an environmental parameter time series multivariate joint representation vector and a human body state time series multivariate joint representation vector. In a specific example of the present application, the relevant environmental state data information includes a time queue of air humidity values, a time queue of particle density, and a time queue of wind speed values; the relevant comfort condition information includes a time queue of heart rate values ​​and a time queue of respiratory frequency. Among them, Figure 3 FIG. 1 is a flowchart of sub-step S1 of the method for pushing event information based on a smart cycling helmet according to an embodiment of the present application. Figure 3 As shown, the step S1 includes the steps of: S11, respectively extracting the time series features of the time queue of the air humidity value, the time queue of the particle density, the time queue of the wind speed value, the time queue of the heart rate value and the time queue of the respiratory frequency to obtain the air humidity time series implicit feature coding vector, the particle density time series implicit feature coding vector, the wind speed time series implicit feature coding vector, the heart rate time series implicit feature coding vector and the respiratory frequency time series implicit feature coding vector; S12, combining the air humidity time series implicit feature coding vector, the particle density time series implicit feature coding vector and the wind speed time series implicit feature coding vector to obtain the environmental parameter time series multivariate joint representation vector; S13, combining the heart rate time series implicit feature coding vector and the respiratory frequency time series implicit feature coding vector to obtain the human body state time series multivariate joint representation vector.

[0032] Specifically, in a specific example of the present application, the step S11 includes: performing time series encoding based on a time series convolutional network on the time queue of the air humidity value, the time queue of the particle density, the time queue of the wind speed value, the time queue of the heart rate value, and the time queue of the respiratory frequency to obtain the time series implicit feature encoding vector of the air humidity, the time series implicit feature encoding vector of the particle density, the time series implicit feature encoding vector of the wind speed, the time series implicit feature encoding vector of the heart rate, and the time series implicit feature encoding vector of the respiratory frequency. It should be understood that since the environmental state data and comfort data of the rider during riding are time series, that is, these data are constantly changing over time, therefore, in order to deeply understand the changing trends of various environmental parameters and physiological indicators during riding and their impact on riding risks, the present application further uses a time series convolutional network to perform time series feature modeling on the time queue of the air humidity value, the time queue of the particle density, the time queue of the wind speed value, the time queue of the heart rate value, and the time queue of the respiratory frequency. Those skilled in the art should know that a temporal convolutional network (TCN) is a neural network architecture specifically used to process time series data. It combines the advantages of convolutional neural networks (CNNs), such as local perception, weight sharing, and multi-layer feature extraction, while solving the training difficulties and gradient vanishing problems that traditional recurrent neural networks (RNNs) and their variants (such as LSTM and GRU) may encounter when processing long sequence data. Specifically, the temporal convolutional network is based on the basic principle of convolution operations, and performs point-by-point convolution operations on time series data by sliding the convolution kernel in the time dimension to capture the temporal correlation dependencies in the data. At the same time, the temporal convolutional network adopts a causal convolution structure to ensure that when extracting features at the current time point, it only relies on data from the past and current moments, which complies with the causal principle of time series data and ensures the rationality of time series modeling. Based on this, the present application uses a time series convolutional network to perform time series encoding on the time series data of the air humidity value, particle density, wind speed value, heart rate value and respiratory frequency. It can utilize the powerful time series feature extraction capability of the time series convolutional network to effectively capture the dynamic change trend of various parameters during riding, thereby obtaining the time series implicit feature coding vector of air humidity, the time series implicit feature coding vector of particle density, the time series implicit feature coding vector of wind speed, the time series implicit feature coding vector of heart rate and the time series implicit feature coding vector of respiratory frequency, providing strong data support for subsequent riding risk assessment.

[0033] Specifically, the step S12 combines the air humidity time series implicit feature coding vector, the particle density time series implicit feature coding vector and the wind speed time series implicit feature coding vector to obtain the environmental parameter time series multivariate joint representation vector. It should be understood that the present application takes into account that air humidity, particle density and wind speed each reflect a specific aspect of the environment, but in actual riding scenarios, these environmental factors interact and influence each other, and jointly affect the riding environment and the rider's state. Therefore, in order to comprehensively consider the combined effects of various environmental factors on the rider's physical state, the present application combines the air humidity time series implicit feature coding vector, the particle density time series implicit feature coding vector and the wind speed time series implicit feature coding vector to perform feature splicing and combination, so as to fuse the environmental factor feature information of different dimensions and form an environmental parameter time series multivariate joint representation vector, thereby providing a more comprehensive and accurate description of the environmental state for subsequent riding risk assessment.

[0034] Specifically, in step S13, the heart rate timing implicit feature coding vector and the respiratory frequency timing implicit feature coding vector are feature combined to obtain the human body state timing multivariate joint representation vector. It should be understood that heart rate and respiratory rate are important indicators that reflect the physiological state of the human body during cycling. There is a close physiological relationship between the two, which together reflect the adaptation of the human body to cycling movements and environmental changes. Therefore, the present application also constructs a human body state timing multivariate joint representation vector by feature splicing and combining the heart rate timing implicit feature coding vector and the respiratory frequency timing implicit feature coding vector, which is conducive to a more in-depth analysis of the dynamic changes of the human body state during cycling and the interactive relationship with the environment.

[0035] In the above-mentioned event information push method based on smart cycling helmets, the step S2 performs interactive response analysis based on two-way attention on the environmental parameter time series multivariable joint representation vector and the human body state time series multivariable joint representation vector to obtain a human body state-environmental parameter time series fine-grained interactive response encoding vector. It should be understood that the present application takes into account the complex two-way interactive relationship between human body state and environmental parameters, that is, changes in environmental conditions will have an impact on human body state, and at the same time, changes in human body state will also react on the perception and adaptation to the environment. Therefore, in order to deeply explore the complex interaction pattern between human body state and environmental parameters, the present application proposes an interactive response analysis method based on two-way attention, which performs two-way correlation interaction analysis and attention weight calculation on the human body state time series multivariable joint representation vector and the environmental parameter time series multivariable joint representation vector to focus on the important interactive response relationship between human body state and environmental conditions, thereby accurately capturing the mutual influence and adaptive changes between the two during the riding process. Among them, Figure 4 FIG. 1 is a flowchart of sub-step S2 of the method for pushing event information based on a smart cycling helmet according to an embodiment of the present application. Figure 4 As shown, the step S2 includes the steps of: S21, constructing a bidirectional attention field between the environmental parameter time series multivariable joint representation vector and the human body state time series multivariable joint representation vector to obtain a human body state-environment parameter time series interaction positive attention score field and a human body state-environment parameter time series interaction reverse attention score field; S22, performing attention balance modulation on the human body state-environment parameter time series interaction positive attention score field and the human body state-environment parameter time series interaction reverse attention score field to obtain a human body state-environment parameter time series interaction positive and negative bidirectional attention balance field; S23, based on the human body state-environment parameter time series interaction positive and negative bidirectional attention balance field, performing feature mapping interaction response analysis on the environmental parameter time series multivariable joint representation vector and the human body state time series multivariable joint representation vector to obtain the human body state-environment parameter time series fine-grained interaction response encoding vector.

[0036] Figure 5 FIG. 1 is a flowchart of sub-step S21 of the method for pushing event information based on a smart cycling helmet according to an embodiment of the present application. Figure 5 As shown, the step S21 includes the steps of: S211, performing homography projection transformation on the environmental parameter time series multivariable joint representation vector and the human body state time series multivariable joint representation vector to obtain the environmental parameter time series multivariable joint feature homography projection coding vector and the human body state time series multivariable joint feature homography projection coding vector; S212, calculating the forward attention score field of the environmental parameter time series multivariable joint feature homography projection coding vector relative to the human body state time series multivariable joint feature homography projection coding vector to obtain the human body state-environmental parameter time series interaction forward attention score field; S213, calculating the reverse attention score field of the human body state time series multivariable joint feature homography projection coding vector relative to the environmental parameter time series multivariable joint feature homography projection coding vector to obtain the human body state-environmental parameter time series interaction reverse attention score field.

[0037] More specifically, the step S211 is expressed as follows:

[0038]

[0039]

[0040] in, represents the time series multivariate joint representation vector of the environmental parameters, represents the human body state time series multivariate joint representation vector, and are different homography projection matrices, and They respectively represent the homography projection coding vector of the time series multivariate joint feature of the environmental parameters and the homography projection coding vector of the time series multivariate joint feature of the human body state.

[0041] Specifically, this application takes into account the differences in the characteristic dimensions of the environmental parameter time series multivariable joint representation vector and the human body state time series multivariable joint representation vector. In order to achieve effective alignment and fusion of the two in interactive analysis, this application first performs a homography projection transformation on the environmental parameter time series multivariable joint representation vector and the human body state time series multivariable joint representation vector to map the two into a common feature space to keep the characteristic dimensions consistent. In this way, in the subsequent interactive analysis stage, the two can be analyzed based on the same characteristic dimensions and are comparable.

[0042] More specifically, the step S212 is expressed by the formula:

[0043]

[0044] in, represents the transpose of a vector, represents the matrix multiplication operation, is the attention score scaling factor, Represents the positive attention score field of temporal interaction between human body state and environmental parameters.

[0045] That is, the forward attention score field is calculated for the environmental parameter time series multivariate joint representation vector and the human body state time series multivariate joint representation vector after homography projection transformation, so as to show the importance of each part of the environmental parameter time series information when the environmental parameter characteristics and the human body state characteristics are interactively analyzed, and the forward attention score field is generated accordingly, which reflects the weight distribution of the impact of environmental parameters on the human body state, so that in the subsequent time series interactive analysis link, this weight distribution can be used to accurately focus on the key feature parts in the environmental parameter time series multivariate joint characteristics that have an important impact on the human body state, thereby improving the analysis efficiency and accuracy.

[0046] More specifically, the step S213 is expressed by the formula:

[0047]

[0048] in, Represents the reverse attention score field of temporal interaction between human body state and environmental parameters.

[0049] That is, referring to the calculation method of the forward attention score field, the reverse attention score field is calculated from the perspective of human body state information. In the calculation process, the human body state feature parts that are closely related to the changes in environmental parameters are reversely identified. Based on the calculation results, the corresponding attention weight distribution is generated, that is, the reverse attention score field.

[0050] Specifically, the step S22 is expressed by the formula:

[0051]

[0052] in, Indicates cascade, Represents the positive and negative bidirectional attention balance field of human body state-environmental parameter temporal interaction, Representation 3 3 convolution operations.

[0053] Specifically, in order to ensure that the correlation between environmental conditions and human body states can be fully captured in the final temporal interaction response analysis, the present application further constructs a positive and negative bidirectional attention balance field for temporal interaction between human body states and environmental parameters, and comprehensively considers the positive and negative attention information to ensure that in the final temporal interaction encoding, it will neither be overly biased towards one side nor ignore the important information of any side, so as to better balance the temporal correlation interaction relationship between environmental conditions and human body states.

[0054] Figure 6 FIG. 2 is a flowchart of sub-step S23 of the method for pushing event information based on a smart cycling helmet according to an embodiment of the present application. Figure 6 As shown, the step S23 includes the steps of: S231, mapping the environmental parameter time series multivariable joint feature homography projection coding vector and the human body state time series multivariable joint feature homography projection coding vector to the human body state-environmental parameter time series interactive positive and negative bidirectional attention balance field respectively to obtain the environmental parameter time series multivariable joint feature homography projection attention modulation coding vector and the human body state time series multivariable joint feature homography projection attention modulation coding vector; S232, calculating the position point division between the environmental parameter time series multivariable joint feature homography projection attention modulation coding vector and the human body state time series multivariable joint feature homography projection attention modulation coding vector to obtain the human body state-environmental parameter time series fine-grained interaction response coding vector.

[0055] More specifically, the step S231 is expressed by the formula:

[0056]

[0057]

[0058] in, and They respectively represent the homography projection attention modulation coding vector of the temporal multivariate joint feature of the environmental parameters and the homography projection attention modulation coding vector of the temporal multivariate joint feature of the human body state.

[0059] That is, the time series multivariate joint representation vector of the environmental parameters and the time series multivariate joint representation vector of the human body state after homography projection transformation are respectively projected into the positive and negative bidirectional attention balance field of the human body state-environmental parameter time series interaction. Based on the attention weight distribution of the positive and negative bidirectional attention balance field, the feature representations of the two are adjusted so that the feature representations with significant correlation will be strengthened in the final time series interaction coding; the features with weak correlation will be weakened accordingly, so as to more effectively focus on the important correlation information between environmental conditions and human body states, and obtain the time series multivariate joint feature homography projection attention modulation coding vector of the environmental parameter and the time series multivariate joint feature homography projection attention modulation coding vector of the human body state.

[0060] More specifically, the step S232 is expressed by the formula:

[0061]

[0062] in, A coding vector representing the temporal fine-grained interaction response between the human body state and the environmental parameters.

[0063] That is, by using position-by-position point division operations, the environmental parameter time series multivariate joint feature homography projection attention modulation coding vector and the human body state time series multivariate joint feature homography projection attention modulation coding vector are subjected to element-by-element temporal interaction coding to capture the fine-grained interactive response pattern of the human body state to environmental conditions in the time dimension, and obtain the human body state-environmental parameter time series fine-grained interactive response coding vector. In this way, it helps to more accurately quantify the dynamic impact of environmental conditions on cyclists during riding, and provide a more accurate and comprehensive data basis for riding risk assessment.

[0064] In the above-mentioned event information push method based on the smart cycling helmet, the step S3 determines the cycling risk index based on the human body state-environmental parameter time series fine-grained interactive response coding vector. In a specific example of the present application, the step S3 includes: inputting the human body state-environmental parameter time series fine-grained interactive response coding vector into a cycling risk index estimator based on a feedforward neural network model to obtain the cycling risk index. It should be understood that the feedforward neural network model is a neural network model that forward transmits and processes information through a multi-layer neuron structure. It learns the sensitivity and adaptation mechanism of the rider's human body state to changes in environmental conditions by feature extraction and pattern recognition of the human body state-environmental parameter time series fine-grained interactive response coding vector, and estimates the cycling risk index based on this to quantitatively reflect the rider's cycling safety under specific environmental conditions, so as to provide the rider with timely risk warnings based on the cycling risk index, thereby effectively preventing the occurrence of cycling accidents and improving cycling safety.

[0065] In a preferred example of the present application, inputting the human body state-environment parameter time series fine-grained interactive response encoding vector into a cycling risk index estimator based on a feedforward neural network model to obtain a cycling risk index includes:

[0066] First, based on the time series fine-grained interaction response encoding vector of the human body state and environmental parameters, the Distance and Distance matrix for obtaining the temporal fine-grained interaction response of human state and environmental parameters and human body state-environment parameter temporal fine-grained interaction response two distance matrices ;

[0067] Secondly, the weighted sum of the human body state-environmental parameter temporal fine-grained interactive response 1 distance matrix and the human body state-environmental parameter temporal fine-grained interactive response 2 distance matrix is ​​calculated to obtain the human body state-environmental parameter temporal fine-grained interactive response joint distance matrix, which is expressed as follows:

[0068]

[0069] in, and Represent different weight parameters respectively. It means point multiplication by position. represents matrix addition, Represents the joint distance matrix of the temporal fine-grained interaction response between human body state and environmental parameters;

[0070] Next, the eigenvalues ​​of the joint distance matrix of the human body state-environment parameter temporal fine-grained interactive response are determined. , and compose the joint distance eigenvector of the human body state-environment parameter temporal fine-grained interactive response ;

[0071] Then, the human body state-environment parameter temporal fine-grained interactive response encoding vector as a row vector is matrix-multiplied with the human body state-environment parameter temporal fine-grained interactive response-distance matrix to obtain a human body state-environment parameter temporal fine-grained interactive response-distance query vector, which is expressed as follows:

[0072]

[0073] in, Represents the human body state-environment parameter temporal fine-grained interactive response encoding vector, represents matrix multiplication, Represents the human body state-environment parameter temporal fine-grained interactive response-distance query vector;

[0074] Next, the human body state-environment parameter temporal fine-grained interactive response two distance matrices are matrix-multiplied with the autocorrelation matrix of the human body state-environment parameter temporal fine-grained interactive response encoding vector to obtain the human body state-environment parameter temporal fine-grained interactive response two distance correlation matrix, which is expressed as follows:

[0075]

[0076] in, represents the transpose of a vector, Represents the human body state-environment parameter temporal fine-grained interactive response two-distance correlation matrix;

[0077] Then, after matrix multiplication of the human body state-environmental parameter temporal fine-grained interactive response one distance query vector and the human body state-environmental parameter temporal fine-grained interactive response two distance association matrix, further multiplication with the human body state-environmental parameter temporal fine-grained interactive response joint distance eigenvector is performed to obtain an optimized human body state-environmental parameter temporal fine-grained interactive response encoding vector, which is expressed as follows:

[0078]

[0079] in, Represents the joint distance eigenvector of the fine-grained interaction response between human body state and environmental parameters in time series, Representing the optimized human body state-environment parameter temporal fine-grained interaction response encoding vector;

[0080] Finally, the optimized human body state-environmental parameter time series fine-grained interaction response encoding vector is input into a cycling risk index estimator based on a feedforward neural network model to obtain a cycling risk index.

[0081] Here, when the human body state temporal multivariate joint representation vector and the environmental parameter temporal multivariate joint representation vector represent the human body state multivariate temporal joint coding features and the environmental state multivariate temporal joint coding features respectively, when performing an interactive response based on the feature positive and negative attention fields, the insufficient correspondence of the positive and negative attention field reinforcement association caused by the difference in the temporal distribution of the parameter sources will cause the sparse association interactive response of the human body state-environmental parameter temporal fine-grained interactive response coding vector, thereby reducing the accuracy of the cycling risk index obtained by the cycling risk index estimator based on the feedforward neural network model due to the lack of decoding inference degree.

[0082] Therefore, the first distance matrix and the second distance matrix of the human body state-environment parameter time series fine-grained interaction response coding vector are used as the fine-grained metric association cluster representation of the human body state-environment parameter time series fine-grained interaction response coding vector, and the dynamic programming of the relationship between association clusters of different association clusters is performed on the human body state-environment parameter time series fine-grained interaction response coding vector and the self-association representation of the human body state-environment parameter time series fine-grained interaction response coding vector respectively to simulate the sparse activation based on neuron clusters of the association system, and the fine-grained predictable sparsity of the human body state-environment parameter time series fine-grained interaction response coding vector is coordinated with the intrinsic representation of the metric association cluster of the first distance matrix and the second distance matrix of the human body state-environment parameter time series fine-grained interaction response coding vector, so as to avoid the lack of decoding inference degree affected by insufficient association caused by sparsity, and improve the accuracy of the cycling risk index obtained by inputting the human body state-environment parameter time series fine-grained interaction response coding vector into the cycling risk index estimator based on the feedforward neural network model.

[0083] After obtaining the cycling risk index, we further push warnings based on the preset evaluation threshold of the cycling risk index to obtain level events. First, we need to define different levels of risk thresholds, which represent different levels from safe to dangerous, and are usually set based on historical data, expert opinions, and user feedback. Each threshold corresponds to a specific warning level, such as low risk, medium risk, and high risk. When the cycling risk index reaches or exceeds a preset threshold, the system will trigger a warning notification of the corresponding level. To this end, it is necessary to analyze in detail the risk factors that may appear in various cycling scenarios and quantify the degree of impact of these factors on cycling safety. For example, in severe weather conditions, factors such as particle density, humidity level, and wind speed may significantly increase the difficulty of cycling; and after long-term high-intensity exercise, changes in physiological indicators such as heart rate and respiratory rate also reflect the deepening of physical fatigue. By conducting in-depth research on these variables, the appropriate threshold range can be determined for each warning level.

[0084] For low-risk situations, it may only take a slight vibration or soft sound to remind people to pay attention to the current situation; for medium-risk situations, more obvious signals should be used, such as continuous vibration combined with voice prompts, to ensure that potential threats are noticed; as for high-risk situations, it is strongly recommended to stop riding immediately and find a safe area. At this time, the system will issue a continuous alarm until it is confirmed that the risk has been eliminated. In order to ensure the effectiveness of the early warning, all forms of notifications must undergo rigorous testing, including but not limited to sound volume, vibration intensity, voice clarity, etc., to ensure that they can be clearly perceived whether in a noisy environment or when wearing headphones. In addition, considering extreme situations (such as sudden natural disasters or traffic accidents), an emergency response plan is also needed to effectively convey emergency information when conventional means of communication fail.

[0085] At the same time, considering individual differences, it is also important to allow users to customize their warning preferences according to their own needs. For example, some users may prefer visual warnings rather than auditory warnings, or want to adjust the sensitivity between different levels of warnings. Therefore, flexible and adjustable options are provided in the settings interface, allowing users to freely choose the warning method that best suits their habits.

[0086] In addition, in order to ensure the effective transmission of early warning information, a stable and reliable communication platform needs to be established to send relevant information to designated devices, such as smartphones or other networked devices, immediately when the risk exceeds the standard. In this process, an encrypted transmission protocol is used to protect data security and prevent sensitive information leakage. At the same time, considering the network delay problem, the time interval from detection to notification is shortened as much as possible to make the early warning as immediate as possible. For areas without mobile network coverage, a part of the commonly used early warning content can be pre-cached in offline mode and then updated synchronously after the connection is restored. In order to enhance communication stability, a multimodal communication solution can be considered, that is, multiple connection methods such as Wi-Fi, Bluetooth, and cellular networks can be supported at the same time to cope with communication challenges in different environments. In addition, in the event of possible signal interference or loss, a redundant design should be provided to ensure that even if there is a problem with one channel, other backup channels can still maintain normal communication.

[0087] In summary, the event information push method based on the smart cycling helmet according to the embodiment of the present application is explained, which uses sensors deployed inside the smart helmet to monitor the relevant environmental state data information and relevant comfort situation data information when the rider is riding, and introduces the time series analysis technology based on deep learning to perform time series modeling analysis on the relevant environmental state data information and the relevant comfort situation data information, so as to extract the time series feature representation of the environmental state and the human body state, and then, by performing interactive response analysis on the environmental state characteristics and the human body state characteristics based on the two-way attention mechanism, the potential impact of environmental conditions on the human body state during the rider's riding process and the adaptive changes of the human body state to the environmental conditions are excavated, so as to intelligently estimate the riding risk index under the current situation, and based on this, make early warning notifications of corresponding risk levels. In this way, a more accurate and dynamic assessment of riding risks can be achieved to improve the accuracy and timeliness of early warning notifications and ensure the riding safety of riders.

[0088] Furthermore, an event information push system based on a smart cycling helmet is also provided.

[0089] Figure 7 FIG. 1 is a block diagram of an event information push system based on a smart cycling helmet according to an embodiment of the present application. Figure 7 As shown, according to the embodiment of the present application, the event information push system 100 based on the smart cycling helmet includes: a multivariate time series feature extraction module 110, which is used to perform multivariate time series joint feature extraction on relevant environmental state data information and relevant comfort condition information respectively to obtain an environmental parameter time series multivariate joint representation vector and a human body state time series multivariate joint representation vector; an interactive response analysis module 120, which is used to perform interactive response analysis based on two-way attention on the environmental parameter time series multivariate joint representation vector and the human body state time series multivariate joint representation vector to obtain a human body state-environmental parameter time series fine-grained interactive response coding vector; a cycling risk assessment module 130, which is used to determine a cycling risk index based on the human body state-environmental parameter time series fine-grained interactive response coding vector.

[0090] Here, those skilled in the art can understand that the specific operations of each module in the above-mentioned event information push system based on the smart cycling helmet have been referred to above. Figures 1 to 6 The event information push method based on the smart cycling helmet has been described in detail, and therefore, its repeated description will be omitted.

[0091] The basic principle of the present invention is described above in conjunction with specific embodiments. However, it should be pointed out that the advantages, strengths, effects, etc. mentioned in the present invention are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. must be possessed by each embodiment of the present invention. In addition, the specific details of the above embodiments are only for the purpose of illustration and facilitation of understanding, rather than limitation, and the above details do not limit the present invention to being implemented by adopting the above specific details.

[0092] In the above embodiments, the description of each embodiment has its own emphasis. For the parts that are not described or recorded in detail in a certain embodiment, please refer to the relevant description of other embodiments. In the several embodiments provided by the present invention, it should be understood that the disclosed system and method can be implemented in other ways. For example, the system embodiment described above is only schematic. For example, the unit division is only a logical function division, and there may be other division methods in actual implementation. The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the scheme of this embodiment.

[0093] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference to a figure in a claim should not be considered as limiting the claim to which it relates.

[0094] In addition, it is obvious that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. The multiple units stated in the system claims can also be implemented by one unit through software or hardware.

[0095] Finally, it should be noted that the above description has been given for the purpose of illustration and description. In addition, the above embodiments are only used to illustrate the technical solution of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for pushing event information based on a smart cycling helmet, comprising: Using several groups of sensors deployed inside the smart helmet to monitor the relevant environmental status data information and relevant comfort data information of the rider during riding; Calculating a riding risk index based on the relevant environmental status data information and the relevant comfort condition information; Based on the preset evaluation threshold of the riding risk index, a warning of a level event is obtained, wherein the riding risk index is calculated based on the relevant environmental status data information and the relevant comfort condition information, including: Performing multivariate time series joint feature extraction on the relevant environmental state data information and the relevant comfort condition information respectively to obtain an environmental parameter time series multivariate joint representation vector and a human body state time series multivariate joint representation vector; Performing a bidirectional attention-based interactive response analysis on the environmental parameter temporal multivariate joint representation vector and the human body state temporal multivariate joint representation vector to obtain a human body state-environmental parameter temporal fine-grained interactive response encoding vector; The cycling risk index is determined based on the human body state-environmental parameter time series fine-grained interactive response encoding vector.

2. The method for pushing event information based on a smart cycling helmet according to claim 1, characterized in that: The relevant environmental status data information includes a time queue of air humidity values, a time queue of particle density, and a time queue of wind speed values; the relevant comfort condition information includes a time queue of heart rate values ​​and a time queue of respiratory frequency.

3. The event information push method based on the smart cycling helmet according to claim 2 is characterized in that: The relevant environmental state data information and the relevant comfort condition information are respectively subjected to multivariate time series joint feature extraction to obtain an environmental parameter time series multivariate joint representation vector and a human body state time series multivariate joint representation vector, including: Respectively extracting the time series features of the time series of the air humidity value, the time series of the particle density, the time series of the wind speed value, the time series of the heart rate value, and the time series of the respiratory frequency to obtain an air humidity time series implicit feature coding vector, a particle density time series implicit feature coding vector, a wind speed time series implicit feature coding vector, a heart rate time series implicit feature coding vector, and a respiratory frequency time series implicit feature coding vector; Combining the air humidity time series implicit feature coding vector, the particle density time series implicit feature coding vector and the wind speed time series implicit feature coding vector to obtain the environmental parameter time series multivariate joint representation vector; The heart rate time series implicit feature coding vector and the respiratory frequency time series implicit feature coding vector are feature combined to obtain the human body state time series multivariate joint representation vector.

4. The event information push method based on the smart cycling helmet according to claim 3 is characterized in that: Respectively extracting the time series features of the time queue of the air humidity value, the time queue of the particle density, the time queue of the wind speed value, the time queue of the heart rate value, and the time queue of the respiratory frequency to obtain an air humidity time series implicit feature coding vector, a particle density time series implicit feature coding vector, a wind speed time series implicit feature coding vector, a heart rate time series implicit feature coding vector, and a respiratory frequency time series implicit feature coding vector, including: The time queue of the air humidity value, the time queue of the particle density, the time queue of the wind speed value, the time queue of the heart rate value and the time queue of the respiratory frequency are encoded in time series based on a time series convolutional network to obtain the air humidity time series implicit feature coding vector, the particle density time series implicit feature coding vector, the wind speed time series implicit feature coding vector, the heart rate time series implicit feature coding vector and the respiratory frequency time series implicit feature coding vector.

5. The method for pushing event information based on a smart cycling helmet according to claim 4, characterized in that: Performing a bidirectional attention-based interactive response analysis on the environmental parameter temporal multivariate joint representation vector and the human body state temporal multivariate joint representation vector to obtain a human body state-environmental parameter temporal fine-grained interactive response encoding vector, including: Constructing a bidirectional attention field between the environmental parameter temporal multivariate joint representation vector and the human body state temporal multivariate joint representation vector to obtain a human body state-environmental parameter temporal interaction forward attention score field and a human body state-environmental parameter temporal interaction reverse attention score field; Performing attention balance modulation on the human body state-environment parameter time series interaction positive attention score field and the human body state-environment parameter time series interaction reverse attention score field to obtain a human body state-environment parameter time series interaction positive and negative bidirectional attention balance field; Based on the positive and negative bidirectional attention balance field of the human body state-environmental parameter temporal series interaction, feature mapping interactive response analysis is performed on the environmental parameter temporal series multivariate joint representation vector and the human body state temporal series multivariate joint representation vector to obtain the human body state-environmental parameter temporal series fine-grained interactive response encoding vector.

6. The method for pushing event information based on a smart cycling helmet according to claim 5, characterized in that: Constructing a bidirectional attention field between the environmental parameter temporal multivariate joint representation vector and the human body state temporal multivariate joint representation vector to obtain a human body state-environmental parameter temporal interaction forward attention score field and a human body state-environmental parameter temporal interaction reverse attention score field, including: Performing homography projection transformation on the environmental parameter time series multivariate joint representation vector and the human body state time series multivariate joint representation vector to obtain an environmental parameter time series multivariate joint feature homography projection coding vector and a human body state time series multivariate joint feature homography projection coding vector; Calculate the forward attention score field of the environmental parameter temporal multivariate joint feature homography projection coding vector relative to the human body state temporal multivariate joint feature homography projection coding vector to obtain the human body state-environmental parameter temporal interactive forward attention score field; The reverse attention score field of the human body state temporal multivariate joint feature homography projection coding vector relative to the environmental parameter temporal multivariate joint feature homography projection coding vector is calculated to obtain the human body state-environmental parameter temporal interaction reverse attention score field.

7. The method for pushing event information based on a smart cycling helmet according to claim 6, characterized in that: Based on the human body state-environment parameter temporal interactive positive and negative bidirectional attention balance field, the environmental parameter temporal multivariate joint representation vector and the human body state temporal multivariate joint representation vector are subjected to feature mapping interactive response analysis to obtain the human body state-environment parameter temporal fine-grained interactive response encoding vector, including: The environment parameter temporal multivariate joint feature homography projection coding vector and the human body state temporal multivariate joint feature homography projection coding vector are respectively mapped to the human body state-environment parameter temporal interactive positive and negative bidirectional attention balance field to obtain the environment parameter temporal multivariate joint feature homography projection attention modulation coding vector and the human body state temporal multivariate joint feature homography projection attention modulation coding vector; The human body state-environmental parameter temporal fine-grained interaction response coding vector is obtained by calculating the point-by-point division between the environmental parameter temporal multivariate joint feature homography projection attention modulation coding vector and the human body state temporal multivariate joint feature homography projection attention modulation coding vector.

8. The method for pushing event information based on a smart cycling helmet according to claim 7, characterized in that: Determining the riding risk index based on the human body state-environmental parameter time series fine-grained interactive response encoding vector includes: The human body state-environmental parameter time series fine-grained interaction response encoding vector is input into a cycling risk index estimator based on a feedforward neural network model to obtain the cycling risk index.

9. An event information push system based on a smart cycling helmet, characterized in that: include: A multivariate time series feature extraction module is used to perform multivariate time series joint feature extraction on relevant environmental state data information and relevant comfort condition information to obtain an environmental parameter time series multivariate joint representation vector and a human body state time series multivariate joint representation vector; An interactive response analysis module, used for performing interactive response analysis based on bidirectional attention on the environmental parameter temporal multivariate joint representation vector and the human body state temporal multivariate joint representation vector to obtain a human body state-environmental parameter temporal fine-grained interactive response encoding vector; The cycling risk assessment module is used to determine the cycling risk index based on the human body state-environmental parameter time series fine-grained interactive response coding vector.

Images