A rehabilitation care planning method and system

Abstract

The present application relates to the technical field of healthcare information processing, and discloses a rehabilitation nursing plan method and system, comprising: obtaining a physiological feedback sequence of a subject through a state perception module, and generating a real-time state vector representing a physiological load state by an attribute mapping unit; a logic arbitration module calculates a plan load index of a current path based on a logic migration matrix, and when a limit threshold of the vector is triggered, a target primitive is selected from an alternative task pool to replace a conflict task according to a dimensional orthogonal relationship and a migration cost, the present application realizes the protection of the logical continuity of the rehabilitation path by topological potential constraint, avoids the phase jump of neural remodeling caused by plan switching, and independently maintains the constant total flux of rehabilitation tasks on the premise of guaranteeing the physiological safety envelope of the subject, thereby improving the consistency and evolution resilience of home rehabilitation decision.

Description

Technical Field

[0001] This invention relates to a rehabilitation nursing plan method and system, belonging to the field of medical and health care information processing technology. Background Technology

[0002] Current digital rehabilitation systems employ preset action sequences and static triggering methods. They set fixed action sequences based on the subject's initial assessment data and use standardized physiological parameters to monitor the progress of the plan. However, the human neurorehabilitation process exhibits time-varying and asymmetric inhibition characteristics. During the execution of the plan, local muscles experience functional dormancy or compensatory fatigue. Static plans cannot identify logical blockages caused by physiological states, resulting in a mismatch between preset nursing logic and real-time physiological capacity.

[0003] Improving approaches that increase the perceptual dimension or linearly reduce the intensity are difficult to handle nonlinear functional evolution. Increasing the perceptual dimension generates redundant data and increases the computational load on the processor. Reducing the intensity of movements to lower the rehabilitation stimulus below the preset threshold, in addition to setting the movement sequence, also has defects in the logical coherence of existing adaptive control strategies. For example, Chinese invention patent CN115359864A discloses a postoperative rehabilitation nursing method and system. The scheme introduces a neural network model and outputs nursing adjustment information by analyzing historical data of patients of the same family. The decision logic is still based on statistical probability discrete point intervention, which relies too much on the black box correction mode of the group's historical samples, ignores the topological continuity of the current rehabilitation path of the individual subject, and lacks endogenous assessment of the transfer cost between movement primitives. The system performs task switching based on discrete instructions, resulting in rehabilitation logic discontinuity. The phase oscillation of neural remodeling caused by the jump of control parameters causes the loss of rehabilitation gain adjustment process and compensatory damage induced by changes in the biomechanical environment.

[0004] Therefore, the technical problem to be solved by this invention is to address the lack of logical coherence in rehabilitation pathways in existing technologies, to construct an asynchronous arbitration mechanism that can dynamically rearrange tasks based on physiological fluctuations, to introduce logical topological constraints in action switching to smooth neural excitation representations, and to establish a deep pathway monitoring system that can identify compensatory behaviors to ensure that rehabilitation gains are not lost during the adjustment process. Summary of the Invention

[0005] To address the problems mentioned in the background art, the technical solution of the present invention is as follows: A rehabilitation nursing plan system, the system comprising:

[0006] The state awareness module is used to acquire the physiological feedback sequence of the subject during the execution of the current rehabilitation care plan;

[0007] The attribute mapping unit has a built-in weight mapping model for generating real-time state vectors based on physiological feedback sequences. The real-time state vectors include cardiopulmonary load components, motion precision components, and skeletal stress components.

[0008] The planning storage unit is used to store the candidate task pool and the associated weight matrix. The candidate task pool contains multiple nursing action primitives. Each nursing action primitive is associated with an attribute evaluation factor consisting of weights of cardiopulmonary load, action fineness and skeletal stress. The associated weight matrix limits the logical transfer weights between different nursing action primitives.

[0009] The logic arbitration module, connected to the state perception module, attribute mapping unit, and plan storage unit, performs the following operations: Calculates the plan load index, which characterizes the energy consumption status of the current rehabilitation pathway. The plan load index is the quotient of the sum of the products of the attribute evaluation factors and logical transfer weights associated with each nursing action primitive in the current rehabilitation nursing plan, relative to the execution time. When any component value in the real-time state vector exceeds a preset logical limit threshold, the currently executed nursing action primitive is marked as a conflicting task. Under the constraint of keeping the plan load index constant, based on the dimensional orthogonality between the attribute evaluation factors and the real-time state vector, the module selects the target primitive with the smallest logical transfer cost deviation from the candidate task pool, and replaces the conflicting task with the target primitive to generate an updated rehabilitation nursing plan.

[0010] Preferably, the logic arbitration module is used to acquire the initial physiological feedback data of the subject when performing the benchmark action within a preset sampling window after the rehabilitation plan is started, and calculate the statistical variance of the initial physiological feedback data to determine the background deviation value of the current execution cycle; the logic arbitration module is used to perform dynamic gain adjustment on the preset logic limit threshold used to determine whether the real-time state vector triggers the logic limit based on the background deviation value, so as to establish a dynamic judgment benchmark associated with the subject's immediate functional state.

[0011] Preferably, the system further includes an environmental perception module for acquiring the physical state parameters of the rehabilitation environment in which the subject is located, including environmental temperature and environmental humidity; and an attribute mapping unit for calling a preset compensation parameter matrix based on the physical state parameters, and using the compensation parameter matrix to perform asymmetric adjustment on the calculated weights of different components in the real-time state vector, so as to eliminate non-physiological disturbances caused by environmental factors to the physiological feedback sequence.

[0012] Preferably, the logic arbitration module is used to provide rhythm guidance signals when issuing rehabilitation nursing plans; the state perception module is used to extract phase lag data of the subject's movement trajectory relative to the rhythm guidance signal in real time; the logic arbitration module is used to determine the subject's neural compensation state based on the temporal evolution characteristics of the phase lag data, and accordingly to perform weighted correction on the real-time state vector, so as to drive the logic arbitration module to retrieve nursing action primitives with correction functions.

[0013] Preferably, the logic arbitration module is used to apply sub-sensory dimension parametric stimulus pulses to the subject during the execution of the rehabilitation nursing plan; the logic arbitration module is used to obtain the subject's response delay data to the parametric stimulus pulses through the state perception module; the logic arbitration module is used to determine that the subject's target rehabilitation pathway is in a compensatory operation state when the response delay data exceeds a preset delay threshold, and to retrieve nursing action primitives with compensatory inhibition function using the alternative task pool.

[0014] Preferably, the system further includes a pre-processing gateway, used to extract the power envelope corresponding to the subject's motion trajectory and the metabolic dissipation envelope corresponding to the physiological feedback sequence; the pre-processing gateway is used to calculate the temporal coherence coefficient of the power envelope and the metabolic dissipation envelope; the pre-processing gateway is used to lock the update logic of the real-time state vector when the temporal coherence coefficient is lower than a preset consistency threshold.

[0015] Preferably, the pool of candidate tasks in the planning storage unit is modeled as a directed weighted graph, where the weights represent the logical transfer costs between different nursing action primitives. The logical arbitration module is used to search for the optimal path in the directed weighted graph that satisfies the orthogonal association relationship and has the minimum logical transfer cost when screening target primitives, so as to maintain the topological consistency of the updated rehabilitation nursing plan in the direction of neural pathway training.

[0016] Preferably, the orthogonal relationship of dimensions is that among the attribute evaluation factors corresponding to the nursing action primitives, the dimension values ​​corresponding to specific components in the real-time state vector that exceed the preset logical limit threshold are lower than the preset component threshold. The attribute mapping unit has a built-in weight matrix based on the principle of physical decoupling. The weight matrix is ​​used to dynamically configure the linear contribution weights of the cardiopulmonary load component, the action fineness component, and the skeletal stress component in the real-time state vector according to the subject's historical rehabilitation data.

[0017] Preferably, the pre-processing gateway calculates the temporal coherence coefficient based on the law of energy conservation. The calculation formula is as follows: ,in, for Power envelope at time, for Metabolic dissipation envelope at time; preprocessing gateway for time-domain coherence coefficient Below At that time, it is determined that the current physiological feedback sequence contains decoherent noise.

[0018] A rehabilitation care planning method for implementing the aforementioned rehabilitation care planning system includes the following steps:

[0019] Step 1101: Obtain the physiological feedback sequence of the subject during the execution of the current rehabilitation care plan;

[0020] Step 1102: Generate a real-time state vector based on the physiological feedback sequence. The real-time state vector includes cardiopulmonary load components, motion precision components, and skeletal stress components.

[0021] Step 1103: Calculate the planned load index that characterizes the energy consumption status of the current rehabilitation pathway. The planned load index is the sum of the products of the attribute evaluation factors and logical transfer weights associated with each nursing action primitive in the current rehabilitation nursing plan, relative to the execution time.

[0022] Step 1104: Monitor whether the real-time state vector triggers the preset logic limit threshold;

[0023] Step 1105: When the value of any component in the real-time state vector exceeds the preset logical limit threshold, mark the currently executed nursing action primitive as a conflicting task.

[0024] Step 1106: Under the constraint of keeping the planned load index constant, based on the dimensional orthogonality between the attribute evaluation factor and the real-time state vector, select the target primitive with the smallest logical migration cost deviation between the candidate task pool and the conflicting task, and use the target primitive to replace the conflicting task to generate an updated rehabilitation care plan.

[0025] Compared with the prior art, the beneficial effects of the present invention are:

[0026] 1. In the rehabilitation nursing plan, nursing action attributes are defined using action primitives containing multi-dimensional weighted vectors. Real-time damage vectors are generated by combining physiological feedback data. When the logic arbitration module identifies physiological load components that trigger corresponding logical constraints, it selects target primitives from the candidate task pool to replace conflicting tasks based on the orthogonal correlation between attribute influencing factors and real-time damage vectors. This changes the traditional approach of only being able to stop or forcibly execute when encountering physiological fluctuations. By reorganizing logic, physiological limitations are avoided, and the total energy output of rehabilitation is maintained constant under the premise of ensuring the physiological safety envelope, so that the rehabilitation process evolves non-stop.

[0027] 2. The planning storage unit establishes a logical transition matrix with limited coupling weights between action primitives. The logic arbitration module calculates the real-time topological potential energy based on the accumulated power consumption. When the system performs action replacement, it uses the real-time topological potential energy value as a constraint condition to select the target primitive with the smallest difference from the original planned logical transition cost, maintains the topological continuity of neural excitation logic, avoids phase jumps in neural remodeling caused by frequent action switching, eliminates violent oscillations in the rehabilitation decision execution process, and autonomously maintains the total throughput of rehabilitation tasks while ensuring physical safety.

[0028] 3. The logic arbitration module injects sub-perceptual dimension parameter micro-perturbations into the execution terminal, collects the subject's corrective response data to the micro-perturbations, and compares the logic recovery delay with the neural activity model to identify the subject's compensatory behavior of using non-target muscles to assist in completing the movement trajectory. It penetrates the apparent physical trajectory achievement, observes the true load state of the neural pathway, realizes the logical penetration of rehabilitation execution quality, blocks the solidification of compensatory habits, and ensures that every nursing action truly acts on the target lesion area, thereby improving the accuracy of neural control in home rehabilitation. Attached Figure Description

[0029] Figure 1 This is a logical module architecture diagram of the rehabilitation nursing plan system of the present invention;

[0030] Figure 2 This is a trend diagram showing the evolution of each component of the real-time state vector of the present invention over training time.

[0031] Figure 3 This is a schematic diagram of the network topology and hardware deployment of the rehabilitation nursing plan system of the present invention. Detailed Implementation

[0032] The following detailed description of a rehabilitation nursing plan method and system provided by the present invention, in conjunction with the accompanying drawings and specific embodiments, should be noted. It should be understood that the following embodiments are intended to explain the present invention and are not intended to limit the scope of protection of the present invention.

[0033] This invention provides a rehabilitation nursing plan system. A state perception module collects the subject's raw physiological signals, and an attribute mapping unit performs cross-modal data fusion to generate a real-time state vector containing three-dimensional load components. A logic arbitration module performs asynchronous instruction rearrangement based on topology migration weights in the plan storage unit to maintain the logical continuity of the rehabilitation path while ensuring physiological safety. The state perception module acquires the subject's physiological feedback sequence during the execution of the current rehabilitation nursing plan, including motion trajectory data collected by an inertial measurement unit, heart rate variability data collected by a photoplethysmography (PPG) sensor, and electrical signal data collected by an electromyography (EMG) sensor. The attribute mapping unit has a built-in weight mapping model for generating a real-time state vector based on the physiological feedback sequence, which includes cardiopulmonary load components, movement fineness components, and skeletal stress components. Specifically, the attribute mapping unit extracts the temporal features from the heart rate variability, i.e. The interval standard deviation is mapped to the cardiopulmonary load component; the statistical variance of the angular velocity of the motion trajectory is calculated. This is mapped to a fineness component of movement; using the subject's body mass With limb acceleration Calculate the product of the skeletal stress components The mapping procedure maps time-series characteristics to normalized load components through a built-in transfer function, utilizing... In processing heart rate variability data, μ refers to the center offset determined based on the subject's resting-state physiological baseline, and k is the sensitivity coefficient, which is used for the fineness of movement component, and the statistical variance is calculated. After first-order dynamic smoothing filtering, and combined with a preset action complexity weight matrix, the data is converted into real-time performance loss values. The attribute mapping unit then calls the environmental compensation matrix. The formula for correcting for non-physiological fluctuations in ambient temperature and humidity is as follows: ,in, This is the corrected real-time state vector. This is the original real-time state vector. The element values ​​were determined based on a gradient experiment in a baseline environment of 25℃ and 50%RH, realizing the three-dimensional state vector feature space mapping of multi-source physiological flow and eliminating data drift caused by environmental interference.

[0034] The planning storage unit stores the candidate task pool and the associated weight matrix. The candidate task pool contains multiple nursing action primitives, each associated with an attribute evaluation factor consisting of weights from three load dimensions. The associated weight matrix limits the logical transition weights between different nursing action primitives. The logic arbitration module connects the attribute mapping unit and the planning storage unit, and is used to calculate the planned load index, which represents the energy consumption status of the current rehabilitation pathway. The planned load index is calculated as follows: the sum of the products of the attribute evaluation factor and the logical transition weight associated with each nursing action primitive in the current rehabilitation nursing plan is calculated, and then the sum is divided by the execution time. When the value of any component in the real-time state vector exceeds the preset logical limit threshold, the logic arbitration module marks the currently executed nursing action primitive as a conflicting task. Under the constraint of a constant planned workload index, based on the dimensional orthogonality between attribute evaluation factors and real-time state vectors, target primitives with the smallest logical transfer cost deviation from conflicting tasks are selected from the candidate task pool, and the conflicting tasks are replaced by the target primitives. Dimensional orthogonality means that among the attribute evaluation factors corresponding to nursing action primitives, the dimensional value corresponding to a specific component in the real-time state vector that exceeds a threshold is lower than a preset component threshold. The candidate task pool in the plan storage unit is modeled as a directed weighted graph. The weights in the directed weighted graph represent the logical transfer cost between different nursing action primitives. When selecting target primitives, the logic arbitration module searches for the optimal path in the directed weighted graph that satisfies the orthogonal association relationship and has the smallest logical transfer cost, so as to maintain the topological consistency of the updated rehabilitation nursing plan in the direction of neural pathway training.

[0035] To eliminate the illusion of trajectory achievement caused by subjects using non-target muscle groups for compensatory mechanisms, the logic arbitration module provides rhythmic guidance signals when issuing the rehabilitation care plan. The state perception module extracts the phase lag data of the subject's movement trajectory relative to the rhythmic guidance signal in real time. The logic arbitration module determines the subject's neural compensation state based on the temporal evolution characteristics of the phase lag data and accordingly performs weighted correction on the real-time state vector. Furthermore, during plan execution, the logic arbitration module applies sub-sensory dimension parametric excitation pulses to the subject and determines whether the target rehabilitation pathway is in a compensatory operational state by acquiring the subject's response delay data to these pulses. The system also... An environmental sensing module is included to acquire the ambient temperature and humidity of the rehabilitation environment in which the subject is located. An attribute mapping unit calls a preset compensation parameter matrix based on physical state parameters and uses this matrix to perform asymmetric adjustments on the calculated weights of different components in the real-time state vector to eliminate non-physiological disturbances caused by environmental factors to the physiological feedback sequence. To filter interference noise caused by sensor loosening or environmental vibration, the system includes a pre-processing gateway to extract the power envelope corresponding to the subject's motion trajectory and the metabolic dissipation envelope corresponding to the physiological feedback sequence. The pre-processing gateway calculates the temporal coherence coefficient of the power envelope and the metabolic dissipation envelope based on the law of energy conservation. The formula for calculating this coefficient is: Where T is the total number of sampling points within the preset sampling window. for Power envelope at time, for The metabolic dissipation envelope at any given time is used to determine the temporal coherence coefficient at the preprocessing gateway. Below At that time, the update logic for determining that the current physiological feedback sequence contains decoherent noise and locking the real-time state vector is implemented. Within the preset sampling window after the rehabilitation plan is started, the logic arbitration module acquires the initial physiological feedback data when the subject performs the baseline action and calculates the statistical variance of the data to determine the background deviation value of the current execution cycle. The logic arbitration module then applies a preset logic limit threshold based on the background deviation value. Dynamic gain adjustment is performed to establish a dynamic judgment criterion that is correlated with the subject's immediate functional state.

[0036] Example 1: In the scenario where a subject with upper limb dysfunction was performing a home rehabilitation task, the subject in the first... During the execution cycle, fine object grasping motion training is performed continuously. The state perception module, through the inertial measurement unit, displays the angular velocity sequence of the motion trajectory relative to the reference trajectory, from the initial state... Upgraded to The trajectory offset exceeds the correlation of the motion fineness component. A preset deviation threshold triggers the rescheduling procedure of the logic arbitration module. At this time, the subject is in a compensatory movement state using shoulder muscles, causing a logical misalignment between the rehabilitation path and the subject's real-time physiological load. The logic arbitration module extracts the execution parameters of the current rehabilitation care plan and calculates the planned load index of the current path. This indicator is the quotient of the sum of the products of the attribute evaluation factors associated with nursing action primitives and the logical transfer weights, relative to the execution time.

[0037] The attribute mapping unit analyzes heart rate variability and trajectory offset data in the physiological feedback sequence to identify overload states in the corresponding granular dimension of the real-time state vector, and maintains the planned load index constant. Under the constraints, the target primitive is retrieved from the candidate task pool of the planned storage unit, and the granularity weight in the attribute evaluation factors of the target primitive is lower than that of the target primitive. And the weight of proximal limb stability is not less than The logic arbitration module calls the association weight matrix in the plan storage unit, performs path rearrangement in the directed weighted graph composed of nursing action primitives, and switches the current fine grasping action to a proximal limb support training action using the minimum transfer cost determined by the logical transfer weights. The subject's heart rate variability remained stable before and after the task switch. Within the preset range, the logical continuity of the rehabilitation care plan in the direction of neural pathway reconstruction is protected.

[0038] Example 2: The purpose of the experiment was to verify the efficacy of the logic arbitration module in reorganizing the rehabilitation path when subjects experienced nonlinear physiological fatigue evolution. The experimental platform used a device integrated with a wearable inertial measurement unit. Photoplethysmography (PPG) sensor The physical testing system, in which The sensor sampling frequency is set to The selection of this frequency is based on the fact that the bandwidth of the characteristic movements of human rehabilitation in balance is typically lower than that of human rehabilitation. A technical trade-off between the computational load of the embedded processor and the actual computational load is achieved by reducing redundant sampling points to decrease real-time computational pressure. Simultaneously, to simulate the electromagnetic environment of home rehabilitation, a signal-to-noise ratio of [value missing] is actively superimposed in the signal acquisition path. Gaussian white noise.

[0039] The experimental design includes an experimental group employing the topological association reconstruction mechanism of this scheme and a control group employing a linear reduction in action intensity strategy. and adoption Control group of overrange trajectory offset threshold strategy For each group, three levels of local fatigue intensity gradients—low, medium, and high—were set. During the experiment, when the subject entered the high fatigue intensity gradient, the deviation of the original motion trajectory extracted by the state perception module from the reference trajectory reached [value missing]. The logic arbitration module identifies that the value exceeds After setting the preset logical limit threshold, the planned load index for the current execution path is calculated as follows: This indicator is the sum of the products of the attribute evaluation factors associated with the nursing action primitives and the logical transfer weights, relative to the execution time; at this time, the logical arbitration module maintains the planned load indicator constant. Under constraints, the granularity weight of the attribute evaluation factors retrieved from the candidate task pool is lower than that of the other two tasks. And the weight of proximal limb stability is not less than The target primitives are shown in Table 1, which represents the rehabilitation performance indicators of each group under different fatigue intensity gradients.

[0040] Table 1: Rehabilitation Performance Indicators of Subjects under Different Working Conditions

[0041]

[0042] Referring to Table 1, when the experimental group faced high-intensity fatigue, i.e., the initial deviation triggered the corresponding logical constraint, the system used the correlation weight matrix in the plan storage unit to search for the path with the minimum logical transfer cost in the directed weighted graph, switching the current fine grasping action to a proximal limb support training action. Its trajectory deviation eventually converged to... Furthermore, the heart rate variability fluctuation remained within a certain range before and after the task switch. This effectively resolves the contradiction between physiological load and the logical misalignment of the rehabilitation plan, compared to the control group. Although modulating intensity alleviates physiological fatigue, the lack of logical matching based on dimensional orthogonality means that subjects still experience fatigue when performing subsequent actions. The trajectory deviated, and the planned workload index showed reduced effectiveness due to insufficient rehabilitation stimulation, compared to the control group. Because the preset logic threshold was set too high, the subjects continued to follow the original plan even after reaching the performance inflection point, resulting in an increase in heart rate pulse. And it produces neural compensatory behavior, confirming The technical rationale for using a threshold as the safety boundary of this invention.

[0043] Example 3: This example combines Figures 1 to 3 A description of a rehabilitation nursing plan method and system, such as Figure 1As shown, this system acquires the subject's physiological feedback sequence and motion trajectory data through the state perception module, and synchronously acquires the physical state parameters of the rehabilitation environment, including ambient temperature and humidity, through the environmental perception module. The pre-processing gateway extracts the power envelope corresponding to the subject's motion trajectory and the metabolic dissipation envelope corresponding to the physiological feedback sequence, and then calculates the temporal coherence coefficient to identify and filter non-pathological noise generated by sensor shaking or environmental vibration. The attribute mapping unit, based on the built-in weight mapping model, transforms the multimodal raw data into a three-dimensional load real-time state vector containing cardiopulmonary load components, motion precision components, and skeletal stress components. The plan storage unit stores a pool of alternative tasks consisting of multiple nursing action primitives, an association weight matrix that limits the coupling weight between actions, and a directed weighted graph that represents the logical transfer cost, providing static logical support for decision rearrangement. The logic arbitration module calculates the plan load index based on the data output by the above units, and performs asynchronous task filtering and replacement when the components in the real-time state vector trigger the preset logical limit threshold. Finally, it outputs an updated rehabilitation nursing plan that achieves topological consistency of the rehabilitation path and smooth phase switching of neural remodeling.

[0044] like Figure 2 As shown in the figure, the horizontal axis represents the training time in minutes, and the vertical axis represents the component values, which are used to quantify the subject's immediate functional load status during the execution of the rehabilitation plan. The figure shows the temporal evolution characteristics of different load dimensions in the real-time state vector. The solid trajectory corresponding to the cardiopulmonary load component and the dotted trajectory corresponding to the skeletal stress component show an upward trend as the training time increases, which is used to characterize the evolution of cumulative fatigue and physical stress. The dashed trajectory corresponding to the fine motor skill component shows an overall downward trend as the subject's functional fluctuations. When the value of a specific component in the real-time state vector reaches or exceeds the preset logical limit threshold represented by the horizontal dashed line at 0.8, the system marks the currently executed action as a conflicting task, and the logic arbitration module retrieves the target primitive with the function of correction or compensatory inhibition for logical reorganization under the constraint of maintaining the plan load index constant.

[0045] like Figure 3As shown, the physiological sensing nodes deployed on the subject's side integrate pulse, electromyography, and inertial measurement units, and together with the environmental sensing nodes responsible for real-time temperature and humidity monitoring, they construct a wireless body area network to achieve non-invasive acquisition of multi-source physiological and environmental parameters. The acquired data stream is transmitted through a local wireless communication link to a smartphone or dedicated gateway that serves as the local intelligent computing hub. This hub has a built-in logic arbitration engine and task pool, responsible for executing computationally intensive asynchronous instruction reordering logic. On the one hand, the local intelligent computing hub sends rhythm guidance signals and control commands to the interactive execution nodes that include audio guidance and haptic feedback functions. On the other hand, it uploads the encrypted rehabilitation data stream to the remote medical care cloud platform for rehabilitation assessment. During the cross-network data interaction process, the system performs anonymization preprocessing at the data acquisition end, so that the cloud platform only receives the desensitized numerical quantification indicators, thereby effectively blocking the risk of leakage of the subject's original physiological characteristic information while realizing remote medical and health care information processing.

[0046] Example 4: In the initial deployment scenario of a subject's first access to the rehabilitation nursing plan system, the logic arbitration module invokes the standardized calibration procedure to address the issue of range differences in the physiological feedback sequence among different individuals. The subject completes the preset time. A low-intensity baseline motion sequence is generated, during which the state-aware module collects raw physiological data streams, and the attribute mapping unit extracts the heart rate within the baseline cycle. The standard deviation of the interval is The variance of the angular velocity of the trajectory The normalized mapping coefficients of the real-time state vector are calculated. The attribute mapping unit obtains the rate of change component by subtracting the baseline value from the real-time physiological parameter and then dividing by the baseline value. That is, the corresponding dimension value in the real-time state vector is calculated using the following formula: ,in, The first real-time state vector One portion, These are the real-time physiological parameters for the current cycle. To obtain the corresponding benchmark value determined by the calibration procedure, the logic arbitration module obtains the statistical distribution probability of the motion trajectory offset in the benchmark action sequence and calculates its... The upper limit of the confidence interval is used as the initial anchor point for the preset logical limit threshold, with the mean of the baseline offset being [value missing]. At that time, the logic arbitration module determines the initial threshold based on the statistical variance. The system executes each time The baseline offset is updated after each training cycle to establish a dynamic judgment envelope that is associated with the subject's immediate functional state.

[0047] In addition, for the logical transfer weight determination step of the association weight matrix in the planned storage unit, the system adopts a quantitative assignment method based on biomechanical association degree. The logical arbitration module determines the weights according to the muscle synergy between nursing action primitives. When two nursing action primitives involve the same target muscle group, the logical arbitration module sets the corresponding logical transfer weights at... to Within the specified range, when the movement involves antagonistic muscle groups, this weight is set at... to Within the specified range, the attribute mapping unit maps the subject's body mass parameters. With real-time acquisition of limb extremity acceleration Input the bone stress component calculation model: ,in, The stress component of the skeleton, in units of , This refers to body mass parameters, in units of... , This refers to the acceleration at the extremities, measured in units of... .

[0048] Example 5: In a non-standardized wearing environment of the rehabilitation nursing plan system, due to the installation deviation between the axis of the wearable inertial measurement unit and the anatomical axis of the subject's limb, a projection component is generated in the angular velocity of the motion trajectory. The logic arbitration module drives the system to execute a standardized calibration procedure to establish a spatial model of the subject's limb. After the sensor is fixed, the subject completes two unloaded extension movements along a preset path according to the system guidance. During this period, the state perception module extracts the rotation matrix of different sensor nodes in the spatial quaternion domain. The attribute mapping unit calculates the gravity vector offset value of each sensor node relative to the proximal joint center point. This procedure establishes a mapping relationship between the anatomical coordinate system and the sensor physical coordinate system, eliminating real-time state vector drift caused by changes in wearing position. The attribute mapping unit converts the original angular velocity sequence... Input rotation matrix Perform spatial attitude reconstruction and standardize motion characteristics The calculation process is expressed as follows: in, For standardized motion characteristics, For rotation matrix, The original angular velocity sequence is used; the logic arbitration module calculates the motion precision component in the real-time state vector based on standardized motion features, and determines the precision when the residual gravity component of the rotation matrix is ​​below a certain value. When the calibration tolerance threshold is reached, the system determines that the hardware deployment status conforms to the operational benchmark of logical arbitration.

[0049] Example 6: When subjects performed lower limb pumping training and the wireless communication link experienced a transient packet loss rate exceeding [percentage missing], In the boundary conditions, the logic arbitration module calls the online self-test procedure to assess the integrity of the physiological feedback sequence, and the pre-processing gateway is open for a duration of [duration missing]. The data retention window is used to perform first-order linear extrapolation within the window using the real-time state vector of the previous execution cycle to fill sampling gaps caused by signal interruptions; if the state awareness module is in Internally restore data connection and determine signal phase lag after reconnection. Less than The preset out-of-step threshold, This is the signal phase lag, measured in radians. The system maintains the execution sequence of the current rehabilitation care plan, ensuring that the care plan is not logically interrupted during physical link fluctuations.

[0050] When the data connection interruption lasts for more than Furthermore, when the on-site safety backoff procedure is triggered, the system calls the safety backoff matrix in the plan storage unit to perform a weight reset operation on the associated weight matrix, and the logic arbitration module corrects the cardiopulmonary load component in the real-time state vector to... The protection benchmark value is determined, and the skeletal stress component is selected from the attribute evaluation factors in the candidate task pool. The static support primitive replacement conflict task enables the standardized motion characteristics of the subjects. Switch to resting state monitoring mode. To standardize motion characteristics, until the state perception module continuously acquires them. A complete and coherent physiological feedback cycle.

[0051] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0052] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A rehabilitation nursing plan system, characterized in that, The system includes: The state awareness module is used to acquire the physiological feedback sequence of the subject during the execution of the current rehabilitation care plan; The attribute mapping unit has a built-in weight mapping model for generating real-time state vectors based on physiological feedback sequences. The real-time state vectors include cardiopulmonary load components, motion precision components, and skeletal stress components. The planning storage unit is used to store the candidate task pool and the associated weight matrix. The candidate task pool contains multiple nursing action primitives. Each nursing action primitive is associated with an attribute evaluation factor consisting of weights of cardiopulmonary load, action fineness and skeletal stress. The associated weight matrix limits the logical transfer weights between different nursing action primitives. The logic arbitration module, connected to the state perception module, attribute mapping unit, and plan storage unit, performs the following operations: Calculates the plan load index, which characterizes the energy consumption status of the current rehabilitation pathway. The plan load index is the sum of the products of the attribute evaluation factors and logical transfer weights associated with each nursing action primitive in the current rehabilitation care plan, relative to the execution time. When any component value in the real-time state vector exceeds a preset logical limit threshold, the currently executed nursing action primitive is marked as a conflicting task. Under the constraint of keeping the plan load index constant, based on the dimensional orthogonality between the attribute evaluation factors and the real-time state vector, the target primitive with the smallest logical transfer cost deviation from the candidate task pool is selected, and the conflicting task is replaced using the target primitive to generate an updated rehabilitation care plan. The dimensional orthogonality relationship is that the dimensional value corresponding to the specific component in the real-time state vector that exceeds the preset logical limit threshold is lower than the preset component threshold among the attribute evaluation factors corresponding to the nursing action primitive. In addition, the logic arbitration module is used to apply sub-sensory dimension parametric stimulus pulses to the subject during the execution of the rehabilitation nursing plan; obtain the subject's response delay data to the parametric stimulus pulses through the state perception module; when the response delay data exceeds the preset delay threshold, determine that the subject's target rehabilitation pathway is in a compensatory operation state, and use the alternative task pool to retrieve nursing action primitives with compensatory inhibition function. The pool of alternative tasks in the planning storage unit is modeled as a directed weighted graph, where the weights represent the logical transfer costs between different nursing action primitives. The logical arbitration module is used to search for the optimal path in the directed weighted graph that satisfies the orthogonal association relationship and minimizes the logical transfer cost when screening target primitives, so as to maintain the topological consistency of the updated rehabilitation nursing plan in the direction of neural pathway training.

2. The rehabilitation nursing plan system according to claim 1, characterized in that, The logic arbitration module is used to acquire the initial physiological feedback data of the subject when performing the baseline action within the preset sampling window after the rehabilitation plan is started, and to calculate the statistical variance of the initial physiological feedback data to determine the background deviation value of the current execution cycle. Based on the background deviation value, dynamic gain adjustment is performed on the preset logic limit threshold used to determine whether the real-time state vector triggers the logic limit, so as to establish a dynamic judgment benchmark associated with the subject's immediate functional state.

3. The rehabilitation nursing plan system according to claim 1, characterized in that, The system also includes an environmental perception module, which is used to acquire the physical state parameters of the rehabilitation environment in which the subject is located, including environmental temperature and humidity; and an attribute mapping unit, which is used to call a preset compensation parameter matrix based on the physical state parameters, and to use the compensation parameter matrix to perform asymmetric adjustment on the calculated weights of different components in the real-time state vector.

4. The rehabilitation nursing plan system according to claim 1, characterized in that, The logic arbitration module provides rhythm guidance signals when issuing rehabilitation care plans; the state awareness module extracts the phase lag data of the subject's movement trajectory relative to the rhythm guidance signals in real time. The logic arbitration module is used to determine the subject's neural compensation state based on the temporal evolution characteristics of phase lag data, and accordingly to perform weighted correction on the real-time state vector, so as to drive the logic arbitration module to retrieve nursing action primitives with correction function.

5. A rehabilitation nursing plan system according to claim 1, characterized in that, The system also includes a pre-processing gateway, which is used to extract the power envelope corresponding to the subject's motion trajectory and the metabolic dissipation envelope corresponding to the physiological feedback sequence; calculate the temporal coherence coefficient of the power envelope and the metabolic dissipation envelope; and lock the update logic of the real-time state vector when the temporal coherence coefficient is lower than the preset consistency threshold.

6. The rehabilitation nursing plan system according to claim 1, characterized in that, The attribute mapping unit has a built-in weight matrix based on the principle of physical decoupling; the weight matrix is ​​used to dynamically configure the linear contribution weights of cardiopulmonary load component, motion precision component and skeletal stress component in real-time state vector according to the subject's historical rehabilitation data.

7. The rehabilitation nursing plan system according to claim 5, characterized in that, The preprocessing gateway calculates the temporal coherence coefficient based on the law of energy conservation. The calculation formula is as follows: ,in, for Power envelope at time, for Metabolic dissipation envelope at time; preprocessing gateway for time-domain coherence coefficient Below At that time, it is determined that the current physiological feedback sequence contains decoherent noise.

8. A method for implementing a rehabilitation nursing plan, used to implement the rehabilitation nursing plan system of claim 1, characterized in that, Includes the following steps: Step 1101: Obtain the physiological feedback sequence of the subject during the execution of the current rehabilitation care plan; Step 1102: Generate a real-time state vector based on the physiological feedback sequence. The real-time state vector includes cardiopulmonary load components, motion precision components, and skeletal stress components. Step 1103: Calculate the planned load index that characterizes the energy consumption status of the current rehabilitation pathway. The planned load index is the sum of the products of the attribute evaluation factors and logical transfer weights associated with each nursing action primitive in the current rehabilitation nursing plan, relative to the execution time. Step 1104: Monitor whether the real-time state vector triggers the preset logic limit threshold; Step 1105: When the value of any component in the real-time state vector exceeds the preset logical limit threshold, mark the currently executed nursing action primitive as a conflicting task. Step 1106: Under the constraint of keeping the planned load index constant, based on the dimensional orthogonality between the attribute evaluation factor and the real-time state vector, select the target primitive with the smallest logical migration cost deviation between the candidate task pool and the conflicting task, and use the target primitive to replace the conflicting task to generate an updated rehabilitation care plan.

Images