An adaptive running exercise management method and system based on the internet of things

Abstract

This invention discloses an adaptive running exercise management method and system based on the Internet of Things (IoT), relating to the field of sports service management strategy technology. An IoT-based adaptive running exercise management system includes: a request acquisition module, a strategy generation module, a structure adjustment module, a strategy pruning module, a parameter adjustment module, and a strategy update module. This invention constructs a segmented processing and sub-strategy pruning mechanism for candidate scheduling strategies. Under multi-dimensional constraints such as comprehensive sports mode switching costs, trend information, persistence characteristics, and historical path information, it structurally decomposes, filters, and reorganizes scheduling strategies, achieving controllable optimization and dynamic reconstruction of complex scheduling strategies.

Description

Technical Field

[0001] This invention relates to the field of sports service management strategy technology, and in particular to an adaptive running exercise management method and system based on the Internet of Things. Background Technology

[0002] With the continuous development of IoT technology and smart sports, data-driven exercise management is gradually becoming an important development direction in the field of sports and health. Through the collaborative work of wearable devices, mobile terminals, and cloud platforms, multi-dimensional status data of users during exercise can be collected in real time, and the user's exercise behavior can be analyzed and intervened accordingly. In typical aerobic exercise scenarios such as running, users need to switch between different exercise modes to achieve more scientific training results. Meanwhile, with the increasing demand for personalized services, how to combine users' historical behavioral characteristics with real-time status data to dynamically manage and schedule resources during exercise has become one of the current hot topics in technological research.

[0003] Most existing technologies focus on monitoring single exercise modes or generating simple exercise suggestions, lacking systematic modeling and optimization of the multi-mode switching process. During exercise mode switching, existing solutions typically fail to quantify the switching costs caused by changes in user status, resulting in a lack of scientific basis for the switching process and potentially impacting user experience and exercise effectiveness. At the resource scheduling level, existing systems employ fixed data collection frequencies and transmission strategies, failing to dynamically adjust according to changes in user status, thus causing data redundancy or missing key data to some extent. Existing technologies generally fail to effectively integrate user behavior characteristics, exercise participation, and historical path information into scheduling decisions, and lack mechanisms for multi-constraint screening and structural optimization of candidate strategies, making it difficult to achieve refined and personalized exercise service management.

[0004] In the process of IoT-based sports management, the comprehensive use of user status data, historical behavioral characteristics, and multi-sports mode switching information to quantitatively evaluate the sports mode switching process, and on this basis to achieve adaptive management of collaborative optimization between sports service resource scheduling and data collection and transmission behavior, thereby improving system resource utilization efficiency while ensuring sports performance, has become a core technical problem that urgently needs to be solved in this field. Summary of the Invention

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows: An IoT-based adaptive running exercise management method includes: Acquire user service request data in aerobic exercise mode, strength exercise mode, and relaxation exercise mode; Based on service request data, a user motion service profile is constructed, and a motion state transition model and a motion resource scheduling model are established. These are used to determine the motion mode switching cost during the motion mode switching process based on user state data collected by IoT terminals, and to generate a set of candidate scheduling strategies for optimizing the allocation of motion service resources. Based on the user's exercise service profile, the historical exercise service process of the user is analyzed, a user exercise participation model is established, and the changes in the user's service demand under different exercise modes are predicted. Based on the predicted changes in service demand, the exercise mode switching structure of the candidate scheduling strategy is adjusted, and the candidate scheduling strategy set is updated. Based on the trend information, continuous characteristics and historical path information of user status data within a preset time window, the candidate scheduling strategies are constrained and screened, and the selected candidate scheduling strategies are tailored for feasibility based on the cost of switching motion modes. Based on the cost of motion mode switching, the data collection and transmission behavior related to users is adaptively adjusted, and the adjustment parameters of the data collection and transmission behavior corresponding to the candidate scheduling strategy are determined, so that the adjustment parameters of the data collection and transmission behavior and the candidate scheduling strategy form a two-way constraint relationship. Based on the pruned candidate scheduling strategies, a motion service management strategy that optimizes resource allocation and service efficiency is generated. When changes in user motion status exceed preset trigger conditions, the motion resource scheduling model, motion status transition model, user motion participation model, and adjustment parameters are updated.

[0006] As a preferred technical solution of the present invention, the motion state transition model arranges the user state data collected by the IoT terminal in chronological order before and after the motion mode switch, divides the time window corresponding to the motion mode switch, extracts the change data of the user state data within the time window, obtains the transition state parameters representing the difference in motion state before and after the motion mode switch based on the change data, and determines the motion mode switch cost based on the transition state parameters; the motion resource scheduling model is constructed based on the user motion service profile and the motion mode switch cost, and generates a set of candidate scheduling strategies by performing correlation analysis between the user motion service profile and the motion mode switch cost.

[0007] As a preferred embodiment of the present invention, the determination of the motion mode switching cost includes: based on user status data within a time window, obtaining the baseline parameter value before motion mode switching and the target time period parameter value after motion mode switching corresponding to the user status data; calculating the parameter difference between the baseline parameter value and the target time period parameter value, as well as the rate of change of the user status data within the time window, wherein the parameter difference and the rate of change constitute the change data; normalizing the change data according to the maximum, minimum, or average value of the user status data within the time window to obtain the normalized parameter difference and rate of change; determining the transition state parameter characterizing the difference in motion state before and after motion mode switching based on the normalized parameter difference and rate of change; determining the corresponding weight coefficient according to the magnitude distribution of the normalized rate of change within the time window, and performing weighted accumulation processing on the transition state parameter based on the weight coefficient to obtain the motion mode switching cost.

[0008] As a preferred technical solution of the present invention, the adjustment of the exercise mode switching structure includes: extracting user preference features, exercise intensity distribution features, and historical response features under different exercise modes based on the user exercise service profile, and constructing analysis dimensions and weight coefficients corresponding to the user exercise service profile; analyzing the user's historical exercise service process based on the analysis dimensions and weight coefficients, performing weighted statistics on the user's participation frequency, participation duration, or interruption status under different exercise modes, establishing a user exercise participation model, and obtaining user participation parameters under different exercise modes; predicting changes in user service demand under different exercise modes based on the changes in participation parameters over different time periods; parameterizing the exercise mode switching structure of the candidate scheduling strategy according to the number of exercise mode switching, exercise mode duration, and switching order between adjacent exercise modes, and adjusting the parameterized exercise mode switching structure according to changes in service demand; and updating the candidate scheduling strategy set based on the adjusted exercise mode switching structure.

[0009] As a preferred embodiment of the present invention, the constraint and screening of candidate scheduling strategies based on trend information and persistence characteristics includes: arranging user status data within a time window in chronological order, calculating the sign of parameter differences between adjacent moments, and determining the direction of change of user status data; determining the trend information of user status data based on the direction of change of multiple consecutive moments; statistically analyzing the continuous time length during which the trend remains consistent within the time window to determine persistence characteristics; and comparing the trend information and persistence characteristics with preset trend conditions to eliminate candidate scheduling strategies that do not meet the preset trend conditions.

[0010] As a preferred technical solution of the present invention, the constraint and screening of candidate scheduling strategies based on historical path information includes: obtaining the user's motion mode switching sequence during the historical motion service process and arranging it in chronological order to form historical path information; dividing the motion mode switching sequence according to the switching relationship between adjacent motion modes in the historical path information to obtain motion mode switching paths; comparing the motion mode switching paths with preset path conditions and eliminating candidate scheduling strategies that do not meet the preset path conditions.

[0011] As a preferred embodiment of the present invention, the feasibility pruning of candidate scheduling strategies includes: obtaining a set of candidate scheduling strategies, and for each candidate scheduling strategy in the set, segmenting it according to the switching nodes between adjacent motion modes in the candidate scheduling strategy to obtain at least two sub-strategies; judging the feasibility of each sub-strategy based on motion mode switching costs, trend information, persistence characteristics, and historical path information, retaining sub-strategies that meet the constraints, and eliminating sub-strategies that do not meet the constraints; and pruning the corresponding candidate scheduling strategy based on the retained sub-strategies or pruning based on the retained sub-strategies. The combined adjustments generate a pruned candidate scheduling strategy. The motion mode switching cost, trend information, persistence characteristics, and historical path information constrain the segmentation, sub-strategy selection, and pruning methods of the candidate scheduling strategy. The segmentation results and structural form of the candidate scheduling strategy limit the scope of influence and pruning methods of the motion mode switching cost, trend information, persistence characteristics, and historical path information in each sub-strategy, thus forming a two-way constraint relationship during the pruning process. The pruned candidate scheduling strategy replaces the corresponding candidate scheduling strategy in the original candidate scheduling strategy set, resulting in a pruned candidate scheduling strategy set.

[0012] As a preferred embodiment of the present invention, the determination of the adjustment parameters includes: for each candidate scheduling strategy in the candidate scheduling strategy set, determining the adjustment parameters of the data acquisition and transmission behavior corresponding to that candidate scheduling strategy, and using the adjustment parameters as one of the decision conditions of that candidate scheduling strategy; constraining the candidate scheduling strategies according to the adjustment parameters of the data acquisition and transmission behavior, so that candidate scheduling strategies that do not meet the adjustment parameters are not selected, modified, or pruned; simultaneously, limiting the value range, combination method, or effective condition of the adjustment parameters based on the motion mode switching structure, resource scheduling requirements, and execution order in the candidate scheduling strategies, so that the adjustment parameters of the data acquisition and transmission behavior are consistent with the candidate scheduling strategies; wherein, the adjustment parameters of the data acquisition and transmission behavior constrain the selection and execution of the candidate scheduling strategies, and the structure and scheduling requirements of the candidate scheduling strategies inversely constrain the setting method of the adjustment parameters, forming a bidirectional constraint relationship between the adjustment parameters and the candidate scheduling strategies.

[0013] As a preferred embodiment of the present invention, the adaptive adjustment of the data acquisition and transmission behavior includes: dividing the data acquisition frequency corresponding to the user status data into sampling levels according to the value of the motion mode switching cost in different numerical ranges; determining the data transmission priority or bandwidth allocation parameter corresponding to the user status data according to the sampling level; and adjusting the sampling level and the data transmission priority or bandwidth allocation parameter when the motion mode switching cost changes.

[0014] An IoT-based adaptive running exercise management system includes: Request Acquisition Module: Acquires service request data from users in aerobic exercise mode, strength exercise mode, and relaxation exercise mode; Strategy generation module: Constructs user motion service profiles, establishes motion state transition models and motion resource scheduling models, which are used to determine the motion mode switching cost during the motion mode switching process based on user state data collected by IoT terminals, and generate a set of candidate scheduling strategies for optimizing the allocation of motion service resources; The structure adjustment module analyzes the user's historical exercise service process based on the user's exercise service profile, establishes a user exercise participation model, and predicts changes in user service demand under different exercise modes. Based on the predicted changes in service demand, it adjusts the exercise mode switching structure of the candidate scheduling strategy and updates the candidate scheduling strategy set. Strategy pruning module: Based on the trend information of user status data changes within a preset time window, continuous characteristics, and historical path information during continuous motion mode switching, the module constrains and filters candidate scheduling strategies, and performs feasibility pruning on the filtered candidate scheduling strategies in combination with the cost of motion mode switching. Adjustment Parameter Module: Based on the cost of motion mode switching, it adaptively adjusts the data acquisition and transmission behavior related to the user, and determines the adjustment parameters of the data acquisition and transmission behavior corresponding to the candidate scheduling strategy, so that the adjustment parameters of the data acquisition and transmission behavior form a two-way constraint relationship with the candidate scheduling strategy. Strategy Update Module: Generates a motion service management strategy that optimizes resource allocation and service efficiency based on the pruned candidate scheduling strategies, and updates the motion resource scheduling model, motion state transition model, user motion participation model, and adjustment parameters when changes in user motion state exceed preset trigger conditions.

[0015] The present invention has the following advantages: This invention quantifies the differences in user state before and after switching motion modes, and dynamically calculates the switching cost by integrating parameter differences, rate of change, and weight distribution, thus transforming motion mode switching from an experience-driven to a data-driven, quantifiable decision-making process.

[0016] This invention constructs a segmented processing and sub-strategy pruning mechanism for candidate scheduling strategies. Under multi-dimensional constraints such as the cost of comprehensive motion mode switching, change trend information, persistence characteristics, and historical path information, the scheduling strategy is structurally decomposed, filtered, and reorganized, realizing the controllable optimization and dynamic reconstruction of complex scheduling strategies.

[0017] This invention introduces adjustment parameters for data acquisition and transmission behavior, and establishes a two-way constraint relationship between the adjustment parameters and candidate scheduling strategies. On the one hand, the adjustment parameters constrain the selection and execution of scheduling strategies; on the other hand, the scheduling strategies inversely limit the range of values ​​and effective conditions of the adjustment parameters, thereby achieving collaborative optimization between motion service scheduling and data acquisition and transmission.

[0018] This invention performs multidimensional weighted analysis on the frequency, duration, and response behavior of users in different exercise modes, and predicts changes in user service needs by combining time-varying characteristics. Based on this, the scheduling strategy structure is dynamically adjusted, realizing the transformation of the scheduling strategy from static configuration to adaptive optimization oriented towards user behavior evolution.

[0019] This invention implements hierarchical control of data acquisition frequency, data transmission priority, and bandwidth allocation based on motion mode switching costs, enabling data acquisition and transmission behavior to adaptively adjust according to changes in the user's motion state, thereby achieving dynamic collaborative allocation of computing and communication resources. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings in the following description are only schematic diagrams of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort. Figure 1 This is a schematic diagram of the structure of an adaptive running exercise management system based on the Internet of Things used in an embodiment of the present invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0022] Example 1: An adaptive running exercise management method based on the Internet of Things, comprising the following steps: Step S1: Obtain service request data from users in aerobic exercise mode, strength exercise mode, and relaxation exercise mode; In this embodiment, the service request data originates from the user's input, selection confirmations, and historical training configurations confirmed during the current training session, either before or during training. The service request data is a combination of data related to the current training session, including mode arrangements, training objectives, duration requirements, intensity requirements, switching requirements, and constraints. The service request data includes at least mode selection data, training objective data, duration arrangement data, intensity setting data, mode switching data, and preference constraint data. The mode selection data characterizes the types of exercise modes used by the user in this training session and their combinations; the training objective data characterizes the user's goal orientation for this training session; the duration arrangement data characterizes the user's settings for the total training duration and the duration of each exercise mode; the intensity setting data characterizes the user's settings for the training load range under different exercise modes; the mode switching data characterizes the user's settings for the mode switching order, switching trigger method, and switching confirmation method; and the preference constraint data characterizes the user's constraints and personalized requirements during the training process.

[0023] In a three-mode treadmill training scenario, users input or confirm the mode combination, training goals, and parameter requirements for the current training session through the training interface. Taking a typical training request as an example, the user sets the aerobic exercise mode as the starting mode, the strength training mode as the intermediate mode, and the relaxation exercise mode as the ending mode, setting the total training duration, the duration of each mode, and the corresponding speed or incline range. For another type of training request, the user only activates the aerobic and relaxation exercise modes, excluding the strength training mode. The resulting mode selection data directly limits the range of modes for this training session and serves as the mode boundary condition for generating subsequent candidate scheduling strategies. Correspondingly, training goal data characterizes training orientations such as fat loss, cardiovascular endurance improvement, lower limb strength training, or recovery relaxation; duration allocation data defines the time allocation basis for each mode in this training session; intensity setting data defines the speed, incline, or tempo ranges under different modes; mode switching data defines the connection order and switching requirements between different exercise modes; and preference constraint data defines the maximum allowed duration, the required mode types, or the parameter ranges that should not be used.

[0024] The system records the content generated from user input, interface selection, template confirmation, and inheritance of historical configurations, and categorizes and organizes it according to preset data fields. These preset data fields include mode, target, duration, intensity, switching, and constraint fields. Through field-based processing, the original interactive content is transformed into structured service request data corresponding to this training session. Data directly input by the user is written into the current training record according to the corresponding fields; data retrieved from historical training records and confirmed in this session is written into the corresponding fields as confirmed request content; content not directly input by the user but confirmed through the default template is also included in the service request data.

[0025] The service request data is distinct from the user status data in subsequent steps. Service request data characterizes the user's subjective needs and defined boundaries for this training session; it is acquired before training begins or in the early stages of training. User status data, on the other hand, characterizes the actual motion state collected by the IoT terminal during training; it is generated during the training process. This current step only acquires and organizes the training request content and does not involve judging the user's status based on physiological or motion parameters.

[0026] In a specific implementation process, after starting a treadmill workout, the user selects an aerobic exercise mode, a strength training mode, and a cool-down mode in the training interface. The user sets the goal of fat loss and lower body strength training, sets the total training time to 30 minutes, and confirms the duration and intensity of the three modes. The user also stipulates that the final stage must include a cool-down mode. After recording and categorizing the above inputs and confirmations according to preset fields, the service request data corresponding to this training session is obtained.

[0027] Step S2: Construct a user motion service profile based on service request data, establish a motion state transition model and a motion resource scheduling model, which are used to determine the motion mode switching cost during the motion mode switching process based on user state data collected by IoT terminals, and to generate a set of candidate scheduling strategies for optimizing the allocation of motion service resources. The motion state transition model arranges user state data collected by IoT terminals in chronological order before and after motion mode switching, divides time windows corresponding to motion mode switching, extracts change data of user state data within time windows, obtains transition state parameters representing the difference in motion state before and after motion mode switching based on the change data, and determines the motion mode switching cost based on the transition state parameters; the motion resource scheduling model is constructed based on user motion service profiles and motion mode switching costs, and generates a set of candidate scheduling strategies by performing correlation analysis between user motion service profiles and motion mode switching costs.

[0028] The determination of the motion mode switching cost includes: based on user status data within a time window, obtaining the baseline parameter value before the motion mode switching and the target time period parameter value after the motion mode switching, respectively; calculating the parameter difference between the baseline parameter value and the target time period parameter value, as well as the rate of change of the user status data within the time window, wherein the parameter difference and the rate of change constitute the change data; normalizing the change data according to the maximum, minimum, or average value of the user status data within the time window to obtain the normalized parameter difference and rate of change; determining the transition state parameter characterizing the difference in motion state before and after the motion mode switching based on the normalized parameter difference and rate of change; determining the corresponding weight coefficient according to the magnitude distribution of the normalized rate of change within the time window, and performing weighted accumulation processing on the transition state parameter based on the weight coefficient to obtain the motion mode switching cost.

[0029] In this embodiment, when constructing a user exercise service profile based on service request data, different fields in the service request data are extracted accordingly. Specifically, mode selection data is used to determine the composition and order of exercise modes used by the user in this training session; training objective data is used to determine the user's training objective; duration allocation data is used to determine the time allocation basis for each exercise mode; intensity setting data is used to determine the parameter range under different exercise modes; mode switching data is used to determine the switching connection method between modes; and preference constraint data is used to determine the constraints that should be met during training. The resulting user exercise service profile includes at least mode preference information, training objective information, duration distribution information, intensity requirement information, and switching constraint information. Taking a three-mode treadmill training scenario as an example, when the user sets the aerobic exercise mode as the starting stage, the strength exercise mode as the intermediate stage, and the relaxation exercise mode as the ending stage, and simultaneously sets the total training duration, the duration of each mode, and the corresponding intensity range, the resulting service profile corresponds to the initial requirement outline of this training session.

[0030] After creating a user motion service profile, a motion state transition model is established. This model analyzes the degree of difference in user state changes before and after switching between different motion modes, and determines the cost of switching motion modes accordingly. The "motion state transition model" refers to the model processing procedure around the motion mode switching node, which involves time-series extraction, difference calculation, normalization, state representation, and weighted cumulative processing of user state data before and after the switch, generating a calculable switching cost around the specific switching node. In this embodiment, the motion state transition model, in the order of data processing, includes a time window construction unit, a state data extraction unit, a change data generation unit, a normalization processing unit, a transition state parameter generation unit, a weight determination unit, and a switching cost output unit.

[0031] The time window construction unit is used to divide the time window around the switching node between adjacent motion modes; the state data extraction unit is used to extract user state data within the time window; the change data generation unit is used to generate parameter difference based on the baseline parameter value before switching and the target time period parameter value after switching, and generate change rate by combining the change of user state data within the time window to form change data; the normalization processing unit is used to normalize the change data based on the maximum, minimum, or average value of user state data within the time window; the transition state parameter generation unit is used to generate transition state parameters representing the state difference before and after motion mode switching based on the normalized parameter difference and change rate; the weight determination unit is used to determine the corresponding weight coefficient according to the size distribution of the normalized change rate within the time window; the switching cost output unit is used to perform weighted accumulation processing on the transition state parameters based on the weight coefficient to obtain the motion mode switching cost of the corresponding motion mode switching node. The motion mode switching cost output by the motion state transition model is used as one of the input conditions of the motion resource scheduling model.

[0032] The user status data originates from data related to the current exercise state collected during training execution, including at least heart rate, cadence, running speed, incline, duration, and status data reflecting changes in exercise load. Within the corresponding time window, baseline parameter values ​​before the switch and target time period parameter values ​​after the switch are acquired, and the parameter difference between the two is calculated. Simultaneously, the rate of change of the user status data within this time window is calculated to form change data. For the change data, normalization is performed based on the maximum, minimum, or average value of the user status data within the time window to eliminate differences in units and value ranges between different types of status parameters. After normalization, the normalized parameter difference and rate of change are combined to form transitional state parameters. The determination of the transitional state parameters includes: dividing the time window into several continuous segments, calculating the normalized parameter difference and normalized rate of change for each segment, and adding the normalized parameter difference and normalized rate of change for each segment to obtain the local transitional state parameters for each segment; then summing or averaging the local transitional state parameters for each segment to obtain the transitional state parameters corresponding to the current exercise mode switching node. The weighting coefficients are determined based on the distribution of the normalized rate of change within the time window. These weighting coefficients are then used to weight and accumulate the transition state parameters to obtain the motion mode switching cost corresponding to the current motion mode switching node. Specifically, this involves: summarizing the normalized rates of change for each sampling moment or sampling segment within the time window to obtain the total rate of change; dividing the normalized rates of change for each sampling moment or sampling segment by the total rate of change to obtain the weighting coefficients for each sampling moment or sampling segment; and then multiplying the transition state parameters for each sampling moment or sampling segment by the corresponding weighting coefficients and accumulating them in chronological order to obtain the motion mode switching cost corresponding to the current motion mode switching node. The motion mode switching cost has a clear data source and processing path, originating from user state data within the time window before and after the switch, and is formed through data extraction, normalization, determination of transition state parameters, and weighted accumulation.

[0033] Simultaneously with establishing the motion state transition model, a motion resource scheduling model is also established. This model generates a set of candidate scheduling strategies based on the user's motion service profile and the cost of switching motion modes. The "motion resource scheduling model" refers to the model processing procedure that combines and constrains different motion modes during training, considering their arrangement, duration, resource allocation requirements, and switching order. The model receives demand boundary information from the user's motion service profile and motion mode switching cost information output by the motion state transition model, and then generates a set of candidate scheduling strategies corresponding to the current training session.

[0034] The exercise resource scheduling model includes a profile input layer, a strategy combination layer, a constraint filtering layer, and a strategy output layer. The profile input layer receives mode preference information, training target information, duration distribution information, intensity requirement information, and switching constraint information from the user's exercise service profile, and determines the mode composition range, duration range, parameter allocation range, and switching order range corresponding to the current training. The strategy combination layer arranges and combines aerobic exercise modes, strength exercise modes, and relaxation exercise modes within the specified ranges to form multiple initial strategy structures. The constraint filtering layer receives the exercise mode switching cost output by the exercise state transition model and writes the exercise mode switching cost corresponding to each adjacent exercise mode connection segment into the corresponding initial strategy structure, restricting mode connection segments that do not meet the switching constraints, thereby filtering out initial strategy structures that do not meet the current training requirements. The strategy output layer outputs a set of candidate scheduling strategies that meet the constraints.

[0035] The user's motion service profile is input into the profile input layer to determine the allowed motion mode types, number of modes, mode order range, duration range of each mode, and training parameter range for each mode. Next, the policy combination layer generates multiple initial policy structures based on these ranges. Each initial policy structure includes at least the motion mode arrangement result, the duration allocation result for each motion mode, and the switching order result between adjacent motion modes. Then, the motion mode switching cost output by the motion state transition model for each motion mode switching node is input into the constraint filtering layer. The motion mode switching cost corresponding to each switching node is matched to the corresponding connection segment in each initial policy structure, and it is determined whether each connection segment meets the switching conditions corresponding to the current training. Finally, the initial policy structures that meet the switching constraints are retained to form a candidate scheduling policy set.

[0036] Each candidate scheduling strategy in the candidate scheduling strategy set includes at least the order of motion mode combinations, the duration of each motion mode, the switching nodes between adjacent motion modes, and the resource allocation requirements corresponding to each motion mode. The resource allocation requirements are used to characterize the training parameter arrangement content corresponding to the execution phase of different motion modes, and include at least speed allocation requirements, gradient allocation requirements, and phase duration requirements.

[0037] The output of the motion state transition model is connected to the input of the motion resource scheduling model. Specifically, the motion state transition model outputs the corresponding motion mode switching cost for each motion mode switching node; the motion resource scheduling model receives the motion mode switching cost and generates a set of candidate scheduling strategies based on the user's motion service profile. For each adjacent motion mode connection segment in the candidate scheduling strategy, the motion mode switching cost of the corresponding switching node is used as the strategy generation constraint for that connection segment. When generating candidate scheduling strategies, the connection method between modes is limited based on the motion mode switching cost corresponding to each connection segment. If the motion mode switching cost corresponding to a certain mode connection segment does not meet the current training requirements, the initial strategy structure containing that connection segment will not enter the candidate scheduling strategy set; if the motion mode switching cost corresponding to a certain mode connection segment meets the current training requirements, that connection segment will be retained and combined with other mode connection segments that meet the conditions to form a candidate scheduling strategy.

[0038] Step S3: Analyze the user's historical exercise service process based on the user's exercise service profile, establish a user exercise participation model, and predict changes in user service demand under different exercise modes; adjust the exercise mode switching structure of the candidate scheduling strategy according to the predicted changes in service demand, and update the candidate scheduling strategy set. The adjustment of the exercise mode switching structure includes: extracting user preference features, exercise intensity distribution features, and historical response features under different exercise modes based on the user's exercise service profile, and constructing analysis dimensions and weight coefficients corresponding to the user's exercise service profile; analyzing the user's historical exercise service process based on the analysis dimensions and weight coefficients, performing weighted statistics on the user's participation frequency, participation duration, or interruption status under different exercise modes, establishing a user exercise participation model, and obtaining user participation parameters under different exercise modes; predicting changes in user service demand under different exercise modes based on changes in participation parameters over different time periods; parameterizing the exercise mode switching structure of the candidate scheduling strategy according to the number of exercise mode switching, exercise mode duration, and switching order between adjacent exercise modes, and adjusting the parameterized exercise mode switching structure according to changes in service demand; and updating the candidate scheduling strategy set based on the adjusted exercise mode switching structure.

[0039] In this embodiment, the "user's historical exercise service process" refers to the training execution records formed by the user during past treadmill training sessions, encompassing aerobic exercise, strength training, and relaxation modes. This historical exercise service process is a data set including participation, duration, and interruption information for each exercise mode in historical training. To maintain consistency with the aforementioned data definitions, the historical exercise service process includes at least participation frequency data, participation duration data, and interruption data for different exercise modes. Participation frequency data characterizes the distribution of the number of times a user enters a particular exercise mode within a historical period; participation duration data characterizes the distribution of the actual duration of a user's training in different exercise modes; and interruption data characterizes instances where a user prematurely ends, skips a switch, or fails to complete the predetermined duration during the execution of a particular exercise mode. All of the above data originates from historical training execution records.

[0040] When analyzing historical exercise service processes, analytical dimensions and weight coefficients related to the current training are extracted based on the user's exercise service profile. "Analytical dimensions" are analytical items used to classify and statistically analyze historical participation behaviors, ensuring that different types of participation behaviors are processed within the same analytical framework. The adjustment of the exercise mode switching structure includes: extracting user preference features, exercise intensity distribution features, and historical response features under different exercise modes based on the user's exercise service profile, and constructing preference analysis dimensions, intensity analysis dimensions, and response analysis dimensions respectively; statistically analyzing the feature values ​​corresponding to each analysis dimension, and dividing the feature values ​​corresponding to each analysis dimension by the sum of the feature values ​​corresponding to all analysis dimensions to obtain the weight coefficients corresponding to each analysis dimension; and analyzing the user's historical exercise service processes based on the analytical dimensions and weight coefficients.

[0041] After determining the analysis dimensions and weighting coefficients, the user's historical exercise service process is analyzed, and the frequency, duration, and interruption of participation in aerobic exercise, strength training, and relaxation exercise modes are extracted. Based on the analysis dimensions and weighting coefficients, the extracted data is weighted and statistically analyzed to establish a user exercise participation model. The "user exercise participation model" is a model processing result used to characterize the user's actual participation level in different exercise modes, transforming scattered participation behavior data from historical training into participation parameters for subsequent strategy adjustments. The "participation level" is a quantitative result formed based on actual mode participation behaviors recorded in historical training data, through classification extraction, dimensional statistical analysis, and weighting.

[0042] The user exercise participation model is a hierarchical parameterized model, including a data input layer, a feature transformation layer, a weighted calculation layer, and a result output layer. The data input layer is used to input user participation frequency data, participation duration data, and interruption data for different exercise modes during historical exercise services. The feature transformation layer converts the participation frequency data, participation duration data, and interruption data into frequency participation values, duration participation values, and interruption correction values, respectively. The weighted calculation layer performs weighted calculations on the frequency participation values, duration participation values, and interruption correction values ​​based on the weight coefficients corresponding to the analysis dimensions, obtaining the participation parameters corresponding to aerobic exercise mode, strength exercise mode, and relaxation exercise mode, respectively. The result output layer outputs the participation parameters corresponding to each exercise mode.

[0043] In this embodiment, user participation parameters for different exercise modes are obtained through a user exercise participation model. These participation parameters characterize the user's actual participation level in the corresponding exercise mode under specific analytical conditions. Specifically, for aerobic exercise, participation parameters are formed based on the number of entries, actual duration, and interruption frequency in historical training; for strength exercise, participation parameters are formed based on the participation frequency, sustained stability, and early exit occurrences in high-intensity mode; and for relaxation exercise, participation parameters are formed based on the entry ratio, duration distribution, and retention rate before the end of the recovery phase. Each exercise mode corresponds to a set of participation parameters, and these parameters correspond to the actual participation behavior already formed in the user's historical training.

[0044] Based on the changes in engagement parameters over different time periods, the system predicts changes in user service needs under different exercise modes. Specifically, this includes: dividing the user's historical exercise service process into multiple statistical time periods in chronological order, and determining the engagement parameters corresponding to different exercise modes within each statistical time period; calculating the difference in engagement parameters for the same exercise mode between adjacent statistical time periods, and determining the direction and magnitude of change in engagement for the corresponding exercise mode based on the difference in engagement parameters; and predicting changes in user service needs under different exercise modes based on the direction and magnitude of change in engagement for each exercise mode over multiple statistical time periods.

[0045] The adjustment of the parameterized motion mode switching structure according to changes in service demand includes: parameterizing the motion mode switching structure of the candidate scheduling strategy according to the number of motion mode switching, the duration of motion mode, and the switching order between adjacent motion modes; and adjusting the parameterized motion mode switching structure according to changes in service demand. Specifically, this includes: representing the motion mode switching structure in the candidate scheduling strategy as switching number parameters, duration parameters, and sequence position parameters; correcting the corresponding switching number parameters, duration parameters, and sequence position parameters according to changes in service demand corresponding to different motion modes; and reconstructing the motion mode switching structure corresponding to the candidate scheduling strategy based on the corrected parameter results.

[0046] Step S4: Based on the trend information, persistence characteristics and historical path information of user status data within a preset time window, constrain and filter candidate scheduling strategies, and perform feasibility tailoring on the filtered candidate scheduling strategies in combination with the cost of motion mode switching. The constraint and screening of candidate scheduling strategies based on trend information and persistence characteristics includes: arranging user status data in chronological order within a time window, calculating the sign of parameter differences between adjacent moments to determine the direction of change in user status data; determining the trend information of user status data based on the direction of change over multiple consecutive moments; calculating the continuous time length during which the trend remains consistent within the time window to determine persistence characteristics; and comparing the trend information and persistence characteristics with preset trend conditions to eliminate candidate scheduling strategies that do not meet the preset trend conditions.

[0047] The constraint and screening of candidate scheduling strategies based on historical path information includes: obtaining the user's motion mode switching sequence during historical motion service processes and arranging it in chronological order to form historical path information; dividing the motion mode switching sequence according to the switching relationship between adjacent motion modes in the historical path information to obtain motion mode switching paths; comparing the motion mode switching paths with preset path conditions and eliminating candidate scheduling strategies that do not meet the preset path conditions.

[0048] The feasibility-based pruning of candidate scheduling strategies includes: obtaining a set of candidate scheduling strategies; segmenting each candidate scheduling strategy according to the switching nodes between adjacent motion modes to obtain at least two sub-strategies; assessing the feasibility of each sub-strategy based on motion mode switching costs, trend information, persistence characteristics, and historical path information; retaining sub-strategies that meet the constraints and eliminating those that do not; pruning the corresponding candidate scheduling strategies based on the retained sub-strategies or combining and adjusting the retained sub-strategies to generate pruned candidate scheduling strategies; wherein the motion mode switching costs, trend information, persistence characteristics, and historical path information constrain the segmentation, sub-strategy selection, and pruning methods of the candidate scheduling strategies; the segmentation results and structural form of the candidate scheduling strategies limit the scope of influence and pruning methods of the motion mode switching costs, trend information, persistence characteristics, and historical path information in each sub-strategy, thereby forming a two-way constraint relationship in the pruning process; and replacing the corresponding candidate scheduling strategies in the original candidate scheduling strategy set with the pruned candidate scheduling strategies to obtain the pruned candidate scheduling strategy set.

[0049] In this embodiment, "change trend information" refers to the continuous change direction of user status data along the time sequence within a preset time window; "persistence characteristics" refers to the continuous duration during which this change direction remains consistent within the preset time window; and "historical path information" refers to the motion mode switching sequence and its path relationship formed by the user during historical motion services. All three types of information are used to further constrain the updated candidate scheduling strategies. The change trend information and persistence characteristics reflect the immediate adaptability of each candidate scheduling strategy under the current training state, while the historical path information reflects the consistency between each candidate scheduling strategy and the user's past mode switching habits.

[0050] When constraining and filtering candidate scheduling strategies based on trend information and persistence characteristics, user state data within a preset time window is arranged chronologically. The "preset time window" is a continuous time interval set around the current execution phase of the motion mode or mode switching node, used to extract continuous state changes in the current training state. For user state data at adjacent moments within the time window, the sign of the corresponding parameter difference is calculated to determine the direction of change of the current user state data between adjacent moments. If the parameter difference is positive, it corresponds to an upward direction; if the parameter difference is negative, it corresponds to a downward direction; if the parameter difference is close to zero, it corresponds to a stationary direction. Based on this, the overall direction of change of user state data within the time window is summarized according to the direction of change over multiple consecutive moments, yielding trend information. This trend information is a summary result of the direction of state change over a continuous period.

[0051] After obtaining the trend information, the duration for which this trend remains consistent within the time window is calculated to determine its persistence characteristic. If a state parameter exhibits the same direction of change across multiple consecutive sampling times, the duration of this consistent change is taken as the persistence characteristic of that state parameter. This distinguishes between short-term fluctuations and continuous changes, ensuring that the selection criteria for candidate scheduling strategies are no longer limited to state changes at a single moment.

[0052] The constraint and screening of candidate scheduling strategies based on trend information and persistence characteristics includes: arranging user status data within a time window in chronological order, calculating the sign of parameter differences between adjacent moments to determine the direction of change in user status data; determining the trend information of user status data based on the direction of change over multiple consecutive moments; statistically analyzing the continuous duration for which the trend remains consistent within the time window to determine persistence characteristics; pre-setting corresponding trend direction conditions and persistence duration conditions for each candidate scheduling strategy, comparing the trend information with the trend direction conditions, and comparing the persistence characteristics with the persistence duration conditions; retaining the corresponding candidate scheduling strategy when the comparison result meets the preset trend conditions, otherwise eliminating the corresponding candidate scheduling strategy.

[0053] The constraint and screening of candidate scheduling strategies based on historical path information includes: obtaining the user's motion mode switching sequence during historical motion services and arranging it in chronological order to form historical path information; dividing the motion mode switching sequence according to the switching relationship between adjacent motion modes in the historical path information to obtain motion mode switching paths; pre-setting a corresponding target path sequence or allowed path sequence for each candidate scheduling strategy, converting the motion mode switching path into an actual path sequence, and comparing it with the corresponding target path sequence or allowed path sequence; when the comparison result meets the preset path conditions, retaining the corresponding candidate scheduling strategy; otherwise, eliminating the corresponding candidate scheduling strategy.

[0054] "Feasibility trimming" refers to segmenting and analyzing candidate scheduling strategies by switching nodes, and determining their feasibility at a local level to form trimmed candidate scheduling strategies. Specifically, a set of filtered candidate scheduling strategies is obtained, and for each candidate scheduling strategy, it is segmented according to the switching nodes between adjacent motion modes to obtain at least two sub-strategies. A "sub-strategy" is a local strategy structure composed of adjacent motion mode segments in the original candidate scheduling strategy, used to evaluate different switching segments of the original candidate scheduling strategy separately.

[0055] The feasibility pruning of candidate scheduling strategies includes: obtaining a set of candidate scheduling strategies, and for each candidate scheduling strategy in the set, segmenting it according to the switching nodes between adjacent motion modes in the candidate scheduling strategy to obtain at least two sub-strategies; and performing feasibility judgment on each sub-strategy based on motion mode switching cost, change trend information, persistence characteristics, and historical path information, specifically: obtaining the motion mode switching cost, change trend information, persistence characteristics, and historical path information corresponding to each sub-strategy, and comparing them with the corresponding cost conditions, trend conditions, persistence conditions, and path conditions; and retaining the sub-strategy when all comparison results meet the corresponding constraints, otherwise removing the sub-strategy.

[0056] The process of pruning or combining the retained sub-strategies to adjust the corresponding candidate scheduling strategies includes: deleting the motion pattern segments corresponding to the eliminated sub-strategies, determining whether the end motion pattern of the adjacent retained sub-strategies and the start motion pattern of the subsequent retained sub-strategies meet the pattern connection condition; when the pattern connection condition is met, connecting the adjacent retained sub-strategies in their original time order; when the pattern connection condition is not met, not connecting them; for multiple retained sub-strategies, performing the above connection judgment and connection processing in sequence to obtain the pruned candidate scheduling strategies.

[0057] The bidirectional constraint relationship proposed in this step, where "motion mode switching costs, trend information, persistence characteristics, and historical path information constrain the segmentation, sub-strategy selection, and pruning methods of candidate scheduling strategies, and the segmentation results and structural forms of candidate scheduling strategies limit the scope of influence and pruning methods of the motion mode switching costs, trend information, persistence characteristics, and historical path information in each sub-strategy," is implemented through a technical process of "segmentation by switching node—segmental judgment of sub-strategies—pruning or reorganization based on the judgment results." Specifically, motion mode switching costs, trend information, persistence characteristics, and historical path information serve as judgment inputs, directly limiting whether a sub-strategy is retained and how it is pruned; the sub-strategy boundaries formed after segmentation of the candidate scheduling strategy conversely limit the aforementioned judgment information to participate in the judgment only within the corresponding sub-strategy scope. The scope of influence of the judgment information is limited by the segmentation results, and the pruning method after segmentation is constrained by the judgment information, thus forming a corresponding bidirectional constraint relationship during the pruning process.

[0058] Step S5: Based on the cost of switching motion modes, adaptively adjust the data collection and transmission behaviors related to users, and determine the adjustment parameters of the data collection and transmission behaviors corresponding to the candidate scheduling strategies, so that the adjustment parameters of the data collection and transmission behaviors form a two-way constraint relationship with the candidate scheduling strategies. The determination of the adjustment parameters includes: for each candidate scheduling strategy in the candidate scheduling strategy set, determining the adjustment parameters for the data acquisition and transmission behavior corresponding to that candidate scheduling strategy, and using the adjustment parameters as one of the decision conditions for that candidate scheduling strategy; constraining the candidate scheduling strategies based on the adjustment parameters for data acquisition and transmission behavior, so that candidate scheduling strategies that do not meet the adjustment parameters are not selected, modified, or pruned; simultaneously, limiting the value range, combination method, or effective condition of the adjustment parameters based on the motion mode switching structure, resource scheduling requirements, and execution order in the candidate scheduling strategies, so that the adjustment parameters for data acquisition and transmission behavior are consistent with the candidate scheduling strategies; wherein, the adjustment parameters for data acquisition and transmission behavior constrain the selection and execution of candidate scheduling strategies, and the structure and scheduling requirements of candidate scheduling strategies inversely constrain the setting method of adjustment parameters, forming a two-way constraint relationship between the adjustment parameters and the candidate scheduling strategies.

[0059] In this embodiment, "data acquisition and transmission behavior" refers to the acquisition and transmission of state data during the current training process. The data acquisition behavior targets data items related to the exercise state during training, including heart rate, cadence, running speed, incline, duration, and state data reflecting changes in exercise load. The data transmission behavior targets the output arrangement of the aforementioned state data during training, including transmission intervals, transmission order, and data transmission requirements at specific nodes. The "adjustment parameters" are a set of parameters that limit the aforementioned data acquisition and transmission behaviors, corresponding to the data guarantee conditions for the actual execution of candidate scheduling strategies.

[0060] When adaptively adjusting user-related data acquisition and transmission behavior based on motion mode switching costs, for each candidate scheduling strategy in the candidate scheduling strategy set, the switching nodes between adjacent motion modes are identified, and the corresponding motion mode switching costs are extracted. These switching costs are directly used as the basis for determining the adjustment of data acquisition and transmission behavior. "Adaptive adjustment" refers to setting different data acquisition intensities and data transmission arrangements based on the differences in switching costs among switching nodes in different candidate scheduling strategies.

[0061] The adjustment parameters include at least two categories: data acquisition parameters and data transmission parameters. Data acquisition parameters define the acquisition method for each state data item during the execution of the candidate scheduling strategy, and include at least the acquisition time period, acquisition frequency, key acquisition nodes, and continuous acquisition duration. Data transmission parameters define the output method for the corresponding state data, and include at least the transmission time period, transmission interval, priority transmission nodes, and continuous transmission duration. All of these parameters are set corresponding to the mode switching structure in the candidate scheduling strategy and are determined around key nodes before and after mode switching and during mode execution.

[0062] For each candidate scheduling strategy in the candidate scheduling strategy set, adjustment parameters for the data acquisition and transmission behavior corresponding to that candidate scheduling strategy are determined, and these adjustment parameters are used as one of the decision conditions for that candidate scheduling strategy. In this embodiment, the candidate scheduling strategy must meet the requirements of its corresponding data acquisition and transmission behavior adjustment parameters. If the adjustment parameters corresponding to a candidate scheduling strategy are inconsistent with the current training conditions, or cannot meet the requirements for state monitoring and data transmission during the execution of the candidate scheduling strategy, then that candidate scheduling strategy will no longer be a preferred selection target.

[0063] Based on the adjustment parameters of data acquisition and transmission behavior, candidate scheduling strategies are constrained to prevent, modify, or eliminate those that do not meet the adjustment parameters. "Does not meet the adjustment parameters" means that a candidate scheduling strategy cannot maintain consistency with its corresponding data acquisition and transmission requirements under its existing motion mode switching structure, resource scheduling requirements, or execution order. For example, if a candidate scheduling strategy includes multiple consecutive switching nodes with high switching costs, the acquisition frequency, transmission frequency, and key monitoring periods corresponding to this strategy must be increased accordingly. If the strategy does not have matching data acquisition and transmission conditions in its actual execution arrangement, then the candidate scheduling strategy must be restricted, modified, or eliminated in this step.

[0064] Based on the motion mode switching structure, resource scheduling requirements, and execution order in the candidate scheduling strategy, the value range, combination method, or effective conditions of the adjustment parameters are limited to ensure that the adjustment parameters for data acquisition and transmission behavior are consistent with the candidate scheduling strategy. The connection relationship between various motion modes in the candidate scheduling strategy determines which time periods are key acquisition segments and which switching nodes are key transmission nodes; the duration of different modes in the candidate scheduling strategy determines the continuous effective range of acquisition and transmission parameters; the resource scheduling requirements in the candidate scheduling strategy determine the acquisition focus of corresponding status data items at different stages; and the execution order in the candidate scheduling strategy determines the triggering order and switching order of the adjustment parameters.

[0065] The "two-way constraint relationship" in this step is achieved through specific parameter determination, condition writing, strategy constraints, and parameter reverse limitation processes. The adjustment parameters for data acquisition and transmission behavior, as one of the decision conditions for candidate scheduling strategies, directly participate in the selection, modification, and pruning of candidate scheduling strategies. The motion mode switching structure, resource scheduling requirements, and execution order in the candidate scheduling strategies, in turn, limit the value range, combination form, and effective scope of the adjustment parameters. The adjustment parameters determine which candidate scheduling strategies enter the subsequent execution scope, while the candidate scheduling strategies determine how the adjustment parameters are set and where they take effect.

[0066] Step 6: Generate a motion service management strategy that optimizes resource allocation and service efficiency based on the pruned candidate scheduling strategy, and update the motion resource scheduling model, motion state transition model, user motion participation model and adjustment parameters when the user's motion state changes exceed the preset trigger conditions.

[0067] In this embodiment, the "motion service management strategy" is the execution result formed based on the pruned candidate scheduling strategy. It includes a comprehensive strategy encompassing motion mode arrangement, resource allocation, execution order, and corresponding data collection and transmission requirements. The generated motion service management strategy is based on the pruned candidate scheduling strategy output in the aforementioned steps, combined with the determined adjustment parameters to form executable training management content.

[0068] When generating a motion service management strategy based on the pruned candidate scheduling strategies, a set of pruned candidate scheduling strategies is obtained, and the candidate scheduling strategy content that meets the current training requirements and constraints is determined from it. "Pruned candidate scheduling strategies" refer to the strategy results after adjustments to the motion mode switching structure, constraint screening and feasibility pruning, and parameter constraint adjustment. According to the motion mode composition, switching order between adjacent motion modes, duration of each motion mode, and corresponding resource scheduling requirements in the determined candidate scheduling strategies, the execution content of this training process is organized to form the motion service management strategy corresponding to the current training. The motion service management strategy is the execution-oriented expression of the pruned candidate scheduling strategies, and its output is the specific management content adopted in the current training process.

[0069] The "resource allocation" refers to the configuration of training parameters and data support for each stage of the current training process. Resource allocation in this step includes at least the allocation of duration, intensity, and mode switching nodes for different exercise modes, as well as the corresponding data acquisition and transmission arrangements. Specifically, duration allocation defines the duration of aerobic, strength, and relaxation exercises within the current training session; intensity allocation defines the speed, incline, or training tempo range for each exercise mode; mode switching node arrangements define the connection points between different exercise modes; and data acquisition and transmission arrangements define the data acquisition and output methods for each stage.

[0070] During the current training process, when a user's motion state changes beyond a preset trigger condition, the motion resource scheduling model, motion state transition model, user motion participation model, and adjustment parameters are updated. The "preset trigger condition" refers to a criterion used to determine whether the current training state has deviated from the applicable scope of the existing management strategy. Its purpose is to ensure that the models and parameters during the current training process can be adjusted promptly as the user's state changes, rather than remaining unchanged. The preset trigger condition is based on user state data collected during training, specifically corresponding to whether the degree of change in user state data relative to the existing strategy within the current time window reaches an update threshold.

[0071] The determination of preset trigger conditions still revolves around user status data. User status data within a preset time window during the current training process is acquired and compared with the status conditions upon which the current motion service management strategy is based. If the magnitude, direction, or duration of change in user status data within the time window reaches a preset update threshold, it is determined that the preset trigger condition has been exceeded. After determining that the change in user motion status exceeds the preset trigger condition, the motion resource scheduling model is updated. "Updating the motion resource scheduling model" refers to revising the mode arrangement boundaries, duration allocation basis, intensity allocation basis, or switching connection basis in the original candidate scheduling strategy generation basis based on the current user status change and the current training execution results, so that the subsequently generated or invoked candidate scheduling strategies are consistent with the current training state. The motion state transition model is updated, which involves re-extracting new change data, determining new transition state parameters, and recalculating the corresponding motion mode switching cost based on the user status data corresponding to the current mode switching node.

[0072] "Updating the user participation model" refers to incorporating the actual participation during the current training process into the participation behavior analysis, supplementing and correcting the participation parameters under different exercise modes. Since the actual participation behavior in the current training constitutes a new training record, it is added as a new part of the historical exercise service process for subsequent participation analysis. Updating and adjusting parameters involves resetting the data collection and transmission behavior parameters corresponding to each candidate scheduling strategy or the current exercise service management strategy based on the updated exercise resource scheduling model, exercise state transition model, and user participation model, ensuring that the new adjusted parameters are consistent with the updated mode arrangement and switching requirements.

[0073] The "update" step in this process involves continuous data processing following the steps described above. When a user's motion state changes beyond the preset trigger conditions, the motion state transition model and motion mode switching cost are updated first. Then, the updated switching cost, along with the current training requirements, is applied to the motion resource scheduling model. The user's motion participation model is updated in conjunction with new participation behaviors during the current training process, and the adjustment parameters corresponding to the candidate scheduling strategy or the current motion service management strategy are further corrected.

[0074] Example 2, an IoT-based adaptive running exercise management system, see [link / reference] Figure 1 As shown, the specific modules are as follows: The request acquisition module is used to acquire service request data from users in aerobic exercise mode, strength exercise mode, and relaxation exercise mode, and to acquire user status data collected by IoT terminals; wherein, the user status data includes, but is not limited to, parameters such as heart rate, exercise intensity, cadence, or speed.

[0075] The strategy generation module, connected to the request acquisition module, is used to construct a user motion service profile based on the service request data and user status data, and to establish a motion state transition model and a motion resource scheduling model on this basis. The motion state transition model is used to determine the motion mode switching cost during the motion mode switching process based on the user status data, and the motion resource scheduling model is used to generate a set of candidate scheduling strategies to achieve the initial allocation of motion service resources.

[0076] The structure adjustment module, connected to the strategy generation module, is used to analyze the user's historical exercise service process based on the user's exercise service profile, construct a user exercise participation model, and predict changes in the user's service demand under different exercise modes based on the user exercise participation model; and adjust the exercise mode switching structure in the candidate scheduling strategy set according to the predicted changes in service demand, and update the candidate scheduling strategy set.

[0077] The parameter adjustment module, connected to the strategy generation module and the structure adjustment module, is used to adaptively adjust the user-related data acquisition and transmission behavior based on the motion mode switching cost, and determine the adjustment parameters of the data acquisition and transmission behavior corresponding to the candidate scheduling strategy. The adjustment parameters are used to constrain the selection and execution of the candidate scheduling strategy, while the structure and scheduling requirements of the candidate scheduling strategy conversely limit the value range and effective conditions of the adjustment parameters, so that the two form a bidirectional constraint relationship.

[0078] The strategy pruning module, connected to the structure adjustment module and the parameter adjustment module, is used to constrain and filter the candidate scheduling strategy set based on the change trend information, persistence characteristics and historical path information of user status data within a preset time window, and to perform feasibility pruning on the filtered candidate scheduling strategies in combination with the motion mode switching cost, so as to obtain the pruned candidate scheduling strategy set.

[0079] The strategy update module, connected to the strategy pruning module, is used to generate a motion service management strategy for optimizing resource allocation and service efficiency based on the pruned candidate scheduling strategies. When the user's motion state changes beyond the preset triggering conditions, the module updates the motion resource scheduling model, motion state transition model, user motion participation model, and adjustment parameters to achieve dynamic adaptive optimization of the system.

[0080] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An adaptive running exercise management method based on Internet of Things, characterized in that, include: Acquire user service request data in aerobic exercise mode, strength exercise mode, and relaxation exercise mode; Based on service request data, a user motion service profile is constructed, and a motion state transition model and a motion resource scheduling model are established. These are used to determine the motion mode switching cost during the motion mode switching process based on user state data collected by IoT terminals, and to generate a set of candidate scheduling strategies for optimizing the allocation of motion service resources. Based on the user's exercise service profile, the historical exercise service process of the user is analyzed, a user exercise participation model is established, and the changes in the user's service demand under different exercise modes are predicted. Based on the predicted changes in service demand, the exercise mode switching structure of the candidate scheduling strategy is adjusted, and the candidate scheduling strategy set is updated. Based on the trend information, continuous characteristics and historical path information of user status data within a preset time window, the candidate scheduling strategies are constrained and screened, and the selected candidate scheduling strategies are tailored for feasibility based on the cost of switching motion modes. Based on the cost of motion mode switching, the data collection and transmission behavior related to users is adaptively adjusted, and the adjustment parameters of the data collection and transmission behavior corresponding to the candidate scheduling strategy are determined, so that the adjustment parameters of the data collection and transmission behavior and the candidate scheduling strategy form a two-way constraint relationship. Based on the pruned candidate scheduling strategies, a motion service management strategy that optimizes resource allocation and service efficiency is generated. When changes in user motion status exceed preset trigger conditions, the motion resource scheduling model, motion status transition model, user motion participation model, and adjustment parameters are updated. 2.The self-adaptive running exercise management method based on the Internet of Things according to claim 1, wherein, The motion state transition model arranges user state data collected by IoT terminals in chronological order before and after motion mode switching, divides time windows corresponding to motion mode switching, extracts change data of user state data within time windows, obtains transition state parameters representing the difference in motion state before and after motion mode switching based on the change data, and determines the motion mode switching cost based on the transition state parameters; the motion resource scheduling model is constructed based on user motion service profiles and motion mode switching costs, and generates a set of candidate scheduling strategies by performing correlation analysis between user motion service profiles and motion mode switching costs.

3. The adaptive running exercise management method based on the Internet of Things according to claim 2, characterized in that, The determination of the motion mode switching cost includes: based on user status data within a time window, obtaining the baseline parameter value before the motion mode switching and the target time period parameter value after the motion mode switching, respectively; calculating the parameter difference between the baseline parameter value and the target time period parameter value, as well as the rate of change of the user status data within the time window, wherein the parameter difference and the rate of change constitute the change data; normalizing the change data according to the maximum, minimum, or average value of the user status data within the time window to obtain the normalized parameter difference and rate of change; determining the transition state parameter characterizing the difference in motion state before and after the motion mode switching based on the normalized parameter difference and rate of change; determining the corresponding weight coefficient according to the magnitude distribution of the normalized rate of change within the time window, and performing weighted accumulation processing on the transition state parameter based on the weight coefficient to obtain the motion mode switching cost.

4. The adaptive running exercise management method based on the Internet of Things according to claim 1, characterized in that, The adjustment of the exercise mode switching structure includes: extracting user preference features, exercise intensity distribution features, and historical response features under different exercise modes based on the user's exercise service profile, and constructing analysis dimensions and weight coefficients corresponding to the user's exercise service profile; analyzing the user's historical exercise service process based on the analysis dimensions and weight coefficients, performing weighted statistics on the user's participation frequency, participation duration, or interruption status under different exercise modes, establishing a user exercise participation model, and obtaining user participation parameters under different exercise modes; predicting changes in user service demand under different exercise modes based on changes in participation parameters over different time periods; parameterizing the exercise mode switching structure of the candidate scheduling strategy according to the number of exercise mode switching, exercise mode duration, and switching order between adjacent exercise modes, and adjusting the parameterized exercise mode switching structure according to changes in service demand; and updating the candidate scheduling strategy set based on the adjusted exercise mode switching structure.

5. The adaptive running exercise management method based on the Internet of Things according to claim 1, characterized in that, The constraint and screening of candidate scheduling strategies based on trend information and persistence characteristics includes: arranging user status data in chronological order within a time window, calculating the sign of parameter differences between adjacent moments to determine the direction of change in user status data; determining the trend information of user status data based on the direction of change over multiple consecutive moments; calculating the continuous time length during which the trend remains consistent within the time window to determine persistence characteristics; and comparing the trend information and persistence characteristics with preset trend conditions to eliminate candidate scheduling strategies that do not meet the preset trend conditions.

6. The adaptive running exercise management method based on the Internet of Things according to claim 1, characterized in that, The constraint and screening of candidate scheduling strategies based on historical path information includes: obtaining the user's motion mode switching sequence during historical motion service processes and arranging it in chronological order to form historical path information; dividing the motion mode switching sequence according to the switching relationship between adjacent motion modes in the historical path information to obtain motion mode switching paths; comparing the motion mode switching paths with preset path conditions and eliminating candidate scheduling strategies that do not meet the preset path conditions.

7. The adaptive running exercise management method based on the Internet of Things according to claim 1, characterized in that, The feasibility-based pruning of candidate scheduling strategies includes: obtaining a set of candidate scheduling strategies; segmenting each candidate scheduling strategy according to the switching nodes between adjacent motion modes to obtain at least two sub-strategies; assessing the feasibility of each sub-strategy based on motion mode switching costs, trend information, persistence characteristics, and historical path information; retaining sub-strategies that meet the constraints and eliminating those that do not; pruning the corresponding candidate scheduling strategies based on the retained sub-strategies or combining and adjusting the retained sub-strategies to generate pruned candidate scheduling strategies; wherein the motion mode switching costs, trend information, persistence characteristics, and historical path information constrain the segmentation, sub-strategy selection, and pruning methods of the candidate scheduling strategies; the segmentation results and structural form of the candidate scheduling strategies limit the scope of influence and pruning methods of the motion mode switching costs, trend information, persistence characteristics, and historical path information in each sub-strategy, thereby forming a two-way constraint relationship in the pruning process; and replacing the corresponding candidate scheduling strategies in the original candidate scheduling strategy set with the pruned candidate scheduling strategies to obtain the pruned candidate scheduling strategy set.

8. The adaptive running exercise management method based on the Internet of Things according to claim 1, characterized in that, The determination of the adjustment parameters includes: for each candidate scheduling strategy in the candidate scheduling strategy set, determining the adjustment parameters for the data acquisition and transmission behavior corresponding to that candidate scheduling strategy, and using the adjustment parameters as one of the decision conditions for that candidate scheduling strategy; constraining the candidate scheduling strategies based on the adjustment parameters for data acquisition and transmission behavior, so that candidate scheduling strategies that do not meet the adjustment parameters are not selected, modified, or pruned; simultaneously, limiting the value range, combination method, or effective condition of the adjustment parameters based on the motion mode switching structure, resource scheduling requirements, and execution order in the candidate scheduling strategies, so that the adjustment parameters for data acquisition and transmission behavior are consistent with the candidate scheduling strategies; wherein, the adjustment parameters for data acquisition and transmission behavior constrain the selection and execution of candidate scheduling strategies, and the structure and scheduling requirements of candidate scheduling strategies inversely constrain the setting method of adjustment parameters, forming a two-way constraint relationship between the adjustment parameters and the candidate scheduling strategies.

9. The adaptive running exercise management method based on the Internet of Things according to claim 1, characterized in that, The adaptive adjustment of the data acquisition and transmission behavior includes: classifying the sampling level of the data acquisition frequency corresponding to the user status data according to the value of the motion mode switching cost in different numerical ranges; determining the data transmission priority or bandwidth allocation parameter corresponding to the user status data according to the sampling level; and adjusting the sampling level and data transmission priority or bandwidth allocation parameter when the motion mode switching cost changes.

10. An adaptive running exercise management system based on the Internet of Things, characterized in that, The system employs an IoT-based adaptive running exercise management method as described in any one of claims 1 to 9, comprising: Request Acquisition Module: Acquires service request data from users in aerobic exercise mode, strength exercise mode, and relaxation exercise mode; Strategy generation module: Constructs user motion service profiles, establishes motion state transition models and motion resource scheduling models, which are used to determine the motion mode switching cost during the motion mode switching process based on user state data collected by IoT terminals, and generate a set of candidate scheduling strategies for optimizing the allocation of motion service resources; The structure adjustment module analyzes the user's historical exercise service process based on the user's exercise service profile, establishes a user exercise participation model, and predicts changes in user service demand under different exercise modes. Based on the predicted changes in service demand, it adjusts the exercise mode switching structure of the candidate scheduling strategy and updates the candidate scheduling strategy set. Strategy pruning module: Based on the trend information of user status data changes within a preset time window, continuous characteristics, and historical path information during continuous motion mode switching, the module constrains and filters candidate scheduling strategies, and performs feasibility pruning on the filtered candidate scheduling strategies in combination with the cost of motion mode switching. Adjustment Parameter Module: Based on the cost of motion mode switching, it adaptively adjusts the data acquisition and transmission behavior related to the user, and determines the adjustment parameters of the data acquisition and transmission behavior corresponding to the candidate scheduling strategy, so that the adjustment parameters of the data acquisition and transmission behavior form a two-way constraint relationship with the candidate scheduling strategy. Strategy Update Module: Generates a motion service management strategy that optimizes resource allocation and service efficiency based on the pruned candidate scheduling strategies, and updates the motion resource scheduling model, motion state transition model, user motion participation model, and adjustment parameters when changes in user motion state exceed preset trigger conditions.

Images