A collaborative control method and system for an integrated rehabilitation device

Abstract

The present invention provides a collaborative control method and system for integrated rehabilitation devices, which acquire training data to be trained and communication data and perform preprocessing, and use the sub-devices in each rehabilitation device as nodes to construct a training sub-graph; sequentially fill the preprocessed training data to be trained into the corresponding training sub-graph according to the training order to obtain an initial training sub-graph; then determine the target rehabilitation device with the highest intelligence level to improve the convenience of user operation, and use the initial training sub-graph of the target rehabilitation device as the main graph and the remaining initial training sub-graphs as auxiliary graphs; according to the principle of merging nodes with the same training function, map all the auxiliary graphs to the main graph to form a complete training graph; use the corresponding communication conversion strategy to convert the complete training graph so that the rehabilitation devices complete rehabilitation training according to the complete training graph. The present invention can enable multiple existing rehabilitation devices to perform synchronous training, and the user operation is convenient, which can reduce economic and human input.

Description

Technical Field

[0001] The present invention belongs to the technical field of rehabilitation medical devices, and particularly relates to a collaborative control method and system for integrated rehabilitation devices. Background Art

[0002] The development of medical devices has provided great convenience for doctors and patients. In particular, medical devices for rehabilitation training have greatly improved the rehabilitation process of patients and effectively reduced the nursing difficulty of medical staff.

[0003] Currently, various corresponding rehabilitation devices have been developed in the market for different diseased parts, generally divided into upper limb training rehabilitation devices and lower limb training rehabilitation devices according to the body limbs. Upper limb rehabilitation training devices include wrist joint trainers, elbow joint trainers, forearm and wrist joint combined trainers, etc. Lower limb training rehabilitation devices include ankle joint trainers, standing trainers and other training devices. For some patients who need combined training, multifunctional training devices have also emerged in the market, such as combining a forearm rotation trainer, a wrist joint flexion and extension trainer, a shoulder joint rotation trainer, a compound wall puller, a shoulder ladder, a ribbed wood, a pulley suspension ring trainer, a pulley suspension ring trainer, etc. to form a composite rehabilitation training device.

[0004] For a small number of people with partial diseases, such as those with elbow joint diseases, a single elbow joint trainer can complete the rehabilitation training. However, for those with partial diseases in both the elbow joint and the lower limbs, the cost of purchasing a multifunctional training device is too high. Although purchasing a single rehabilitation training device can reduce costs, patients can only train in segments. For example, patients with both ankle joint and elbow joint diseases often have limited funds and first buy a single rehabilitation device and then another. If multiple diseased parts are trained simultaneously, human assistance is required, and the cost of human input is too high. Summary of the Invention

[0005] In order to solve the above problems existing in the prior art, the present invention provides a collaborative control method and system for integrated rehabilitation devices. The technical problems to be solved by the present invention are achieved through the following technical solutions:

[0006] Advantageous Effects:

[0007] In a first aspect, the present invention provides a collaborative control method for integrated rehabilitation devices, including:

[0008] S100, obtaining the to-be-trained data and communication data of all rehabilitation devices, and performing preprocessing to obtain the preprocessed to-be-trained data and communication data; the rehabilitation devices include upper limb rehabilitation devices and lower limb rehabilitation devices;

[0009] S200. For each rehabilitation device, use any sensing sub-device on the rehabilitation device as a sensing node, any control sub-device as a control node, and any motion sub-device controlled by the control sub-device as a motion node to construct a training sub-graph.

[0010] S300. In the order of training during training, sequentially fill the pre-processed data to be trained into the corresponding training sub-graphs to obtain initial training sub-graphs.

[0011] S400. Determine the target rehabilitation device with the highest intelligence level in the upper limb rehabilitation devices, use the initial training sub-graph of the target rehabilitation device as the main graph, and use the initial training sub-graphs of the remaining rehabilitation devices as auxiliary graphs.

[0012] S500. According to the principle of merging nodes with the same training function, map all the auxiliary graphs to the main graph to form a complete training graph.

[0013] S600. Use the pre-processed communication data to determine the corresponding communication conversion strategy, and use the communication conversion strategy to convert the complete training graph into a communication type applicable to the corresponding rehabilitation device and send it to the rehabilitation device, so that the rehabilitation device can complete the rehabilitation training according to the complete training graph.

[0014] In a second aspect, the present invention provides a collaborative control system for integrated rehabilitation devices. The collaborative control system is connected to all rehabilitation devices. The collaborative control system includes:

[0015] An acquisition module, configured to acquire the data to be trained and communication data of all rehabilitation devices, and perform pre-processing to obtain the pre-processed data to be trained and communication data. The rehabilitation devices include upper limb rehabilitation devices and lower limb rehabilitation devices.

[0016] A construction module, configured to, for each rehabilitation device, use any sensing sub-device on the rehabilitation device as a sensing node, any control sub-device as a control node, and any motion sub-device controlled by the control sub-device as a motion node to construct a training sub-graph.

[0017] A filling module, configured to, in the order of training during training, sequentially fill the pre-processed data to be trained into the corresponding training sub-graphs to obtain initial training sub-graphs.

[0018] A determination module, configured to determine the target rehabilitation device with the highest intelligence level in the upper limb rehabilitation devices, use the initial training sub-graph of the target rehabilitation device as the main graph, and use the initial training sub-graphs of the remaining rehabilitation devices as auxiliary graphs.

[0019] A composition module, configured to map all auxiliary graphs into the main graph according to the principle of merging nodes with the same training function, so as to form a complete training graph;

[0020] A rehabilitation module, configured to determine a corresponding communication conversion strategy by using the preprocessed communication data, and use the communication conversion strategy to convert the complete training graph into a communication type applicable to the corresponding rehabilitation device and send it to the rehabilitation device, so that the rehabilitation device completes rehabilitation training according to the complete training graph.

[0021] The present invention provides a collaborative control method and system for an integrated rehabilitation device, which acquires training data and communication data and performs preprocessing, uses sub-devices in each rehabilitation device as nodes to construct training sub-graphs; fills the preprocessed training data into the corresponding training sub-graphs in sequence according to the training order during training to obtain initial training sub-graphs; determines the target rehabilitation device with the highest intelligence level in the upper limb rehabilitation device to improve the convenience of user operation, and uses the initial training sub-graph of the target rehabilitation device as the main graph, and the initial training sub-graphs of the remaining rehabilitation devices as auxiliary graphs; maps all auxiliary graphs into the main graph according to the principle of merging nodes with the same training function to form a complete training graph; uses the corresponding communication conversion strategy to convert the complete training graph so that the rehabilitation device completes rehabilitation training according to the complete training graph. The present invention can enable multiple existing rehabilitation devices to train synchronously, and the user operation is convenient, which can reduce economic and human input.

[0022] The following will further describe the present invention in detail with reference to the drawings and embodiments. Description of the Drawings

[0023] Figure 1 is a schematic flowchart of a collaborative control method for an integrated rehabilitation device provided by the present invention;

[0024] Figure 2 is a schematic diagram of an elbow joint trainer provided by the present invention;

[0025] Figure 3 is a schematic flowchart of S100 provided by the present invention;

[0026] Figure 4 is a schematic flowchart of S200 provided by the present invention;

[0027] Figure 5 is a schematic flowchart of S300 provided by the present invention;

[0028] Figure 6 is a schematic flowchart of S500 provided by the present invention;

[0029] Figure 7 is a schematic flowchart of S600 provided by the present invention;

[0030] Figure 8 This is a schematic structural diagram of a cooperative control system for an integrated rehabilitation device provided by the present invention. Specific embodiments

[0031] The present invention will be further described in detail below in conjunction with specific embodiments, but the implementation manners of the present invention are not limited thereto.

[0032] As Figure 1 shown, the present invention provides a cooperative control method for an integrated rehabilitation device, including:

[0033] S100. Obtain the to-be-trained data and communication data of all rehabilitation devices, and perform preprocessing to obtain the preprocessed to-be-trained data and communication data; the rehabilitation devices include upper limb rehabilitation devices and lower limb rehabilitation devices;

[0034] The upper limb rehabilitation device in this application can be an elbow joint trainer as Figure 2 shown, and the lower limb rehabilitation device can be an ankle joint trainer. Both the upper limb rehabilitation device and the lower limb rehabilitation device are driven by motors to drive the devices to complete the rehabilitation training of the corresponding parts. The rehabilitation devices need to achieve a control effect, and communication is required between the corresponding devices. The sensors on the upper and lower limb rehabilitation devices are perception sub-devices, the controllers are control sub-devices, and the devices that directly contact the patient's body for training are controlled devices, such as the swing arm and wrist guard in the elbow joint trainer, etc. The controlled devices are generally mechanical structures. The to-be-trained data of the rehabilitation devices generally includes training duration, training location, various parameters during training, such as stretching degree, strength, movable angle, speed, bending degree, etc. Of course, the intelligent rehabilitation device can include the user's login information and the training data generated after the training is completed. These training data become the historical training data in the rehabilitation device log. In some solutions, the historical training data can be used to correct the patient's current training data (to-be-trained data).

[0035] S200. For each rehabilitation device, use any perception sub-device on the rehabilitation device as a perception node, any control sub-device as a control node, and any motion sub-device controlled by the control sub-device as a motion node to construct a training sub-graph;

[0036] Due to various types of rehabilitation equipment and inconsistent levels of intelligence, etc., the various sensing sub-devices, control sub-devices, and motion sub-devices in the rehabilitation equipment are not synchronized among different rehabilitation equipment, and it is even more impossible to achieve synchronous perception or coordinated training during training. Exemplarily, the training process of the elbow joint rehabilitation equipment and the ankle joint trainer often rely on the patient's own initiative or family assistance to complete the coordinated training, and it is impossible to achieve the purpose of unified machine synchronous training. This step requires a node in the current training sub-graph of each sub-device to lay the foundation for the subsequent unified formation of a complete training graph. In this process, for rehabilitation equipment with a relatively high level of intelligence, it can be directly collected from the rehabilitation equipment. For rehabilitation equipment with a relatively low level of intelligence, these node information can be collected manually.

[0037] S300, in the order of training during training, sequentially fill the preprocessed data to be trained into the corresponding training sub-graph to obtain an initial training sub-graph;

[0038] The training order of the rehabilitation equipment in this application can be determined by starting from a specific controlled sub-device, or can be determined according to the order of the patient's training parts. The former can be obtained by using the training start time of the controlled sub-device, and the latter can be determined from the generation time of the historical training data in the log.

[0039] S400, determine the target rehabilitation equipment with the highest level of intelligence in the upper limb rehabilitation equipment, and use the initial training sub-graph of the target rehabilitation equipment as the main graph, and use the initial training sub-graphs of the remaining rehabilitation equipment as auxiliary graphs;

[0040] It should be noted that among multiple rehabilitation equipment, some may be the ones that the user purchased earliest, and some are later. According to the rules of technological development, generally, the rehabilitation equipment purchased later has a higher level of intelligence. Of course, there are also some rehabilitation equipment that is related to cost or performance. Generally, the rehabilitation equipment with a higher cost or higher performance has a higher level of intelligence. In this step, one or several rehabilitation equipment can be selected as the target rehabilitation equipment according to cost or performance.

[0041] S500, according to the principle of merging nodes with the same training function, map all auxiliary graphs to the main graph to form a complete training graph;

[0042] It should be noted that there may be sub-modules with repeated functions among multiple rehabilitation equipment. For example, the control sub-module controls the rotation amplitude of the motor to control the extension of the training during training. Therefore, nodes with the same training function can be merged. How the nodes in the auxiliary graph are merged according to the same function and finally converge to form a complete training graph.

[0043] S600 determines the corresponding communication conversion strategy using the preprocessed communication data, and uses the communication conversion strategy to convert the complete training graph into a communication type applicable to the corresponding rehabilitation device and send it to the rehabilitation device, so that the rehabilitation device completes the rehabilitation training according to the complete training graph.

[0044] It should be noted that different manufacturers of rehabilitation devices use different communication devices or communication interfaces for procurement, resulting in different communication protocols and data transmission formats. Therefore, to achieve the purpose of synchronous training, the complete training graph can be sent to each rehabilitation device synchronously. When the data format or communication format of these rehabilitation devices is different from that of the master device, that is, the target rehabilitation device, it is necessary to determine the corresponding communication conversion strategy according to the communication data, so as to convert the complete training graph into the communication type of the corresponding rehabilitation device and send it to each rehabilitation device.

[0045] As an optional implementation manner of the present invention, referring to Figure 3 , S100 includes:

[0046] S110, obtaining the data to be trained from the control end of each rehabilitation device and obtaining the communication data from the communication end; the communication type includes the interface type used for communication and the data transmission format corresponding to the interface type;

[0047] It should be noted that the interface type in this step can be existing interface types, such as RS485, USB, RS232, Ethernet interface, etc. Since the models of rehabilitation devices are different, the interface types will also be different.

[0048] S120, for each rehabilitation device, determining the communication type according to the communication data of the rehabilitation device; the training data at least includes the stretching degree, speed, pressure and duration during training;

[0049] In addition to including the stretching degree, speed, pressure, and duration, the training data in this step can also include the torsion angle, average joint torsion speed, maximum torsion angle, etc. According to different rehabilitation devices, these training data will change adaptively.

[0050] S130, labeling the communication data with communication type labels according to the labels corresponding to different communication types;

[0051] In this step, the communication data can be labeled with communication type labels according to different communication types, such as USB communication type, CAN communication type or RS485 communication type, laying a foundation for the follow-up.

[0052] S140, grouping the aggregated communication data with the same communication type label into a group, and labeling the communication data in the same group with a group number label to obtain the preprocessed communication data;

[0053] S150. For each rehabilitation device, aggregate the data to be trained of the sub-devices with the same function on the rehabilitation device into a set of training data, and label the set of data to be trained with a device type label to obtain the preprocessed data to be trained.

[0054] In this step, for the same rehabilitation device, aggregating the data to be trained with the same function into a set and labeling it with a device type label can improve the efficiency of the subsequent merging process.

[0055] As an optional implementation manner of the present invention, referring to Figure 4 , S200 includes:

[0056] S210. For each rehabilitation device, use any sensing sub-device on the rehabilitation device as a sensing node, any control sub-device as a control node, and any motion sub-device controlled by the control sub-device as a motion node.

[0057] S220. Determine the most initial sensing node among all sensing nodes, and use this sensing node as the current node to find a target node having a connection relationship with the current node; the target node is a control node or a motion node.

[0058] It should be noted that the outermost sensor or sensing device is the initial sensing node, and this sensing node is controlled by a control sub-device or is connected to a motion sub-device. Each sensing sub-device is the current node, and all target nodes can be obtained by sequential searching.

[0059] S230. Use the line segments connecting the target node and the current node, and the target node and the target node as target edges.

[0060] S240. Assign a weight coefficient to each target edge according to the in-degree of each target node.

[0061] In this step, assign a weight coefficient to each target edge according to the in-degree of each target node. A higher in-degree indicates that the data transmitted between the sensor and the control sub-device or the motion sub-device is more important, or the control signal is more important, so a higher weight coefficient is assigned to it. Conversely, a lower weight coefficient is assigned to it.

[0062] S250. Use each target node as the current node, and repeat the process of S220 until all nodes are traversed to obtain the training sub-graph of the rehabilitation device.

[0063] In this step, for each rehabilitation device, all sensing nodes need to be traversed. In this way, connections are established between all sensing nodes and control nodes, motion nodes, between control nodes, and between control nodes and motion nodes, and finally the training sub-graph of the rehabilitation device is obtained.

[0064] As an alternative implementation manner of the present invention, referring to Figure 5 , S300 includes:

[0065] S310, obtaining historical training data from the logs of each rehabilitation device and determining the corresponding historical complete training graph of the historical training data;

[0066] It should be noted that the historical training data and the historical complete training graph can be obtained from the logs of the rehabilitation device, or from a database specifically storing the historical training data. It is more convenient for this application to obtain them from the logs.

[0067] S320, inputting the historical complete training graph corresponding to the historical training data into a graph neural network to obtain each layer of network nodes for the graph neural network to transfer the features of the historical training data, collecting the network nodes involved in transferring the features in the historical complete training graph in the graph neural network, and using the transfer relationship of the network nodes to determine the transfer order of each historical training data;

[0068] Among them, the current layer of network nodes receives the features transferred from the previous layer of network nodes, and the source network nodes directly receiving the historical complete training graph extract the features of the historical training data;

[0069] It should be noted that the graph neural network can transfer the features in the historical complete training graph layer by layer. The graph neural network can perform clustering according to edges or nodes during the transfer process, so that similar features can gradually converge together and different features gradually disperse. According to the clustering principle, similar features converge together, and vice versa. During the transfer process, due to different historical training data, the transfer order of the network nodes is different, and the transfer order of each historical training data can be obtained.

[0070] S330, for each rehabilitation device, classifying the historical training data of the rehabilitation device according to the sub-device type to obtain classified data;

[0071] S340, determining the nodes corresponding to the classified data in the corresponding training sub-graph, and using the transfer order of the classified data as the attribute information of the nodes;

[0072] S350, for each rehabilitation device, determining the execution order of each sub-device according to the execution order of the control sub-device in the rehabilitation device for executing the data to be trained after preprocessing and the response duration of the motion sub-device;

[0073] It should be noted that in each rehabilitation device, the control sub-device outputs a control signal to control the sensing sub-device or the motion sub-device. Due to different communication response times, that is, different response durations, the arrival times of the training data to be transmitted to the sub-devices are different. It is necessary to determine the execution order according to the execution sequence and response duration. The execution order of the motion sub-device with a longer response duration is relatively later, and the order of the training data to be preprocessed by the control sub-device is later. The longer the response duration, the later the execution order of the corresponding sub-device. The execution time of the control sub-device plus the response duration of the motion sub-device, plus the communication time is the execution time of each sub-device. According to the front and back of the execution time, the execution order of each sub-device can be determined.

[0074] S360. Using the execution order of each sub-device, add node connection relationships in the training sub-graph and determine the training order of each node in the training sub-graph;

[0075] It should be noted that the execution order of each sub-device is controlled by other sub-devices or the data comes from other sub-devices. Therefore, it is necessary to add connection relationships to its corresponding nodes.

[0076] S370. Fill the node attributes, training order, and training data to be processed into the corresponding nodes of the corresponding training sub-graph to obtain the initial training sub-graph.

[0077] As an optional implementation manner of the present invention, determining the target rehabilitation device with the highest intelligence level in the upper limb rehabilitation device in S400 includes:

[0078] Compare the communication efficiencies of all upper limb rehabilitation devices to determine the upper limb rehabilitation devices whose communication efficiencies exceed the threshold;

[0079] Among them, the threshold is a value set according to industry experience or actual situation and can also be changed.

[0080] Conduct a Turing test on all upper limb rehabilitation devices with communication efficiency thresholds, and select the upper limb rehabilitation device with the highest accuracy rate in the test results as the target rehabilitation device with the highest intelligence level.

[0081] In this implementation manner, when determining the highest intelligence level, first use the communication efficiency for screening and then conduct a Turing test. Of course, if the actual situation does not allow, the rehabilitation device with the highest cost can also be selected as the target rehabilitation device with the highest intelligence level.

[0082] As an optional implementation manner of the present invention, refer to Figure 6 , S500 includes:

[0083] S510. For any first node in each auxiliary graph, determine whether there is a second node in other auxiliary graphs whose training order is the same as that of the first node. If so, determine that the first node is the sibling node of the second node.

[0084] It should be noted that the purpose of this step is to confirm whether there are nodes with the same training order between auxiliary graphs. If so, it means that the two nodes are sibling nodes in the graph and are trained synchronously.

[0085] S520. For any first node in the auxiliary graph, determine whether there is a third node in the main graph whose node attribute is the same as that of the first node. If so, determine whether the training order of the third node is the same. If the same, confirm that the training functions of the third node and the first node are the same.

[0086] This step determines nodes with exactly the same node attributes in the auxiliary. Since the attribute information is the transfer order of classification data, it can be confirmed that the transfer orders of the device types and data features of the two are exactly the same in the graph neural network. If the training orders are also the same, it can be fully determined that the training functions of these two nodes are the same. The device type carries a label in the preprocessing link, so it can be directly viewed to determine the label.

[0087] S530. If there is no target node in the main graph whose node attribute is the same as that of the first node, determine whether the data to be trained of the target node is the same as that of the first node. If so, determine that the first node is the child node of the target node.

[0088] It should be noted that if the attributes are different, it means that both the device type and the transfer order may be different. Then, it can be checked whether the data to be trained is the same. If so, it is possible that the data of this node is passed to it by other nodes of the same generation as the first node, and it can be determined that the first node is the child node of the target node.

[0089] S540. Connect all auxiliary graphs according to sibling nodes and parent-child nodes to form a complete auxiliary graph, and map the complete auxiliary graph to the main graph according to the principle of merging nodes with the same training function to obtain a complete training graph.

[0090] As an optional implementation manner of the present invention, refer to Figure 7 , S600 includes:

[0091] S610. Receive a rehabilitation training request sent by any rehabilitation device and determine the preprocessed communication data corresponding to the rehabilitation device.

[0092] S620. Use the preprocessed communication data in S610 to determine the communication data format of the rehabilitation device.

[0093] S630. Determine the corresponding communication conversion strategy according to the communication data format;

[0094] It should be noted that different communication data formats have predetermined bit data, and their data frames specify what data a certain bit is. Therefore, it is necessary to refill the data bit positions according to the data format, and the process of converting the bit positions is the communication conversion strategy.

[0095] S640. Use the communication conversion strategy to convert the complete training graph into a communication type applicable to the corresponding rehabilitation device and send it to the rehabilitation device, so that the rehabilitation device can complete the rehabilitation training according to the complete training graph.

[0096] It should be noted that all rehabilitation devices receive a complete training graph carrying node attributes, training order, and data to be trained. Since it is for the same patient, there is no need for confidentiality, and the training data of other rehabilitation devices can be known among the rehabilitation devices. In this way, the same training order can synchronize the complete training during the training process. Users can operate conveniently and also reduce human assistance to a certain extent.

[0097] As an optional implementation manner of the present invention, S640 includes:

[0098] S641. Use the communication protocol agreed in the communication conversion strategy and the bit data format agreed by the communication protocol to convert the complete training graph into a communication type adapted to the corresponding rehabilitation device, and obtain the data packet to be transmitted;

[0099] S642. Transmit the data packet to be transmitted to the corresponding rehabilitation device, so that the rehabilitation device can parse the transmitted data packet to obtain the complete training graph, restore its own control node and motion node from the complete training graph, and make the control sub-device corresponding to the control node control the motion sub-device corresponding to the motion node in sequence according to the training order in the complete training graph to complete the rehabilitation training.

[0100] Among them, the collaborative control method of the integrated rehabilitation device is implemented on the processor integrated in the collaborative control system. The collaborative control system is also integrated with various types of communication modules. The communication protocol and communication conversion protocol are set inside the communication module. A communication interface adapted to the rehabilitation device is set outside the communication module. The communication interface is connected to the communication interface of the rehabilitation device by plugging and unplugging. The communication protocol stipulates the data transmission format between the collaborative control system and the communication interface, and the communication conversion protocol stipulates the data conversion method for communication between different rehabilitation devices.

[0101] Such as Figure 8As shown in the figure, the present invention provides a collaborative control system for an integrated rehabilitation device. The collaborative control system is connected to all rehabilitation devices. The collaborative control system includes:

[0102] An acquisition module 81, configured to acquire the to-be-trained data and communication data of all rehabilitation devices, and perform preprocessing to obtain the preprocessed to-be-trained data and communication data; the rehabilitation devices include upper limb rehabilitation devices and lower limb rehabilitation devices;

[0103] A construction module 82, configured to, for each rehabilitation device, use any sensing sub-device on the rehabilitation device as a sensing node, any control sub-device as a control node, and any motion sub-device controlled by the control sub-device as a motion node to construct a training sub-graph;

[0104] A filling module 83, configured to sequentially fill the preprocessed to-be-trained data into the corresponding training sub-graphs according to the training order during training to obtain initial training sub-graphs;

[0105] A determination module 84, configured to determine the target rehabilitation device with the highest intelligence level among the upper limb rehabilitation devices, and use the initial training sub-graph of the target rehabilitation device as the main graph, and use the initial training sub-graphs of the remaining rehabilitation devices as auxiliary graphs;

[0106] A composition module 85, configured to map all the auxiliary graphs to the main graph according to the principle of merging nodes with the same training function to form a complete training graph;

[0107] A rehabilitation module 86, configured to determine the corresponding communication conversion strategy by using the preprocessed communication data, and use the communication conversion strategy to convert the complete training graph into a communication type applicable to the corresponding rehabilitation device and send it to the rehabilitation device, so that the rehabilitation device completes the rehabilitation training according to the complete training graph.

[0108] In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the present invention, "a plurality" means two or more unless otherwise specifically defined.

[0109] Although the present application has been described in conjunction with various embodiments herein, however, in the process of implementing the claimed present application, those skilled in the art can understand and implement other variations of the disclosed embodiments by viewing the accompanying drawings, the disclosure content, and the appended claims. In the claims, the word "comprising" does not exclude other components or steps, and "a" or "one" does not exclude a plurality of cases.

[0110] The above content is a further detailed description of the present invention in combination with specific preferred embodiments, and it cannot be determined that the specific implementation of the present invention is only limited to these descriptions. For those of ordinary skill in the technical field to which the present invention pertains, without departing from the concept of the present invention, several simple deductions or substitutions can still be made, and all should be regarded as belonging to the protection scope of the present invention.

Claims

1. A collaborative control method for integrated rehabilitation equipment, characterized in that: include: S100, acquiring the to-be-trained data and communication data of all rehabilitation devices, and performing preprocessing to obtain the preprocessed to-be-trained data and communication data; The rehabilitation equipment includes upper limb rehabilitation equipment and lower limb rehabilitation equipment; S200, for each rehabilitation device, taking any sensing sub-device on the rehabilitation device as a sensing node, taking any controlling sub-device as a controlling node, and taking any motion sub-device controlled by the controlling sub-device as a motion node, so as to construct a training subgraph; S300, according to the training order during training, sequentially filling the pre-processed data to be trained into the corresponding training sub-graph to obtain an initial training sub-graph; S400, determining a target rehabilitation device with the highest intelligence level among the upper limb rehabilitation devices, and using the initial training subgraph of the target rehabilitation device as a main graph, and using the initial training subgraphs of the remaining rehabilitation devices as auxiliary graphs; S500, mapping all auxiliary graphs into the main graph according to the principle of merging nodes with the same training function to form a complete training graph; S600, using the preprocessed communication data to determine a corresponding communication conversion strategy, and using the communication conversion strategy to convert the complete training graph into a communication type suitable for a corresponding rehabilitation device and send the conversion to the rehabilitation device, so that the rehabilitation device completes the rehabilitation training according to the complete training graph; S500 includes: S510, for any first node of each auxiliary graph, determine whether there is a second node in other auxiliary graphs that has the same training order as the first node, and if so, determine that the first node is a sibling node of the second node; S520, for any first node in the auxiliary graph, determine whether the main graph has a third node with the same node attribute as the first node, and if so, determine whether the training order of the third node is the same, and if so, confirm that the training function of the third node is the same as that of the first node; S530, if there is no target node in the main graph having the same node attribute as the first node, determining whether the to-be-trained data of the target node is the same as that of the first node; if the to-be-trained data is the same as that of the first node, determining that the first node is a child node of the target node; S540, all auxiliary graphs are connected according to brother nodes and parent-child nodes to form a complete auxiliary graph, and according to the principle of merging nodes with the same training function, the complete auxiliary graph is mapped to the main graph to obtain a complete training graph.

2. The collaborative control method of integrated rehabilitation equipment according to claim 1, characterized in that: S100 includes: S110, acquiring training data from the control terminal of each rehabilitation device and acquiring communication data from the communication terminal; the communication type includes the interface type used for communication and the data transmission format corresponding to the interface type; S120, for each rehabilitation device, determining a communication type according to the communication data of the rehabilitation device; the training data at least includes stretching degree, speed, pressure and duration during training; S130, marking the communication data with a communication type label according to labels corresponding to different communication types; S140, grouping the aggregated communication data having the same communication type tag into one group, and adding a group number tag to the communication data in the same group, to obtain the communication data after preprocessing; S150, for each rehabilitation device, aggregate the to-be-trained data of sub-devices with the same function on the rehabilitation device into a group of training data, and label the group of to-be-trained data with a device type label to obtain preprocessed to-be-trained data.

3. The collaborative control method of integrated rehabilitation equipment according to claim 1, characterized in that: S200 includes: S210, for each rehabilitation device, taking any sensing sub-device on the rehabilitation device as a sensing node, taking any controlling sub-device as a controlling node, and taking any moving sub-device controlled by the controlling sub-device as a moving node; S220, determining the most initial sensing node among all sensing nodes, and taking the sensing node as the current node, and searching for a target node having a connection relationship with the current node; the target node is a control node or a motion node; S230, taking the line segments connecting the target node and the current node, and the target node and the target node as the target edge; S240, assigning a weight coefficient to each target edge according to the in-degree of each target node; S250, taking each target node as the current node, and repeating the process of S220 until all nodes are traversed to obtain a training subgraph of the rehabilitation equipment.

4. The collaborative control method of integrated rehabilitation equipment according to claim 1, characterized in that: S300 includes: S310, acquiring historical training data from the log of each rehabilitation device and determining a historical complete training graph corresponding to the historical training data; S320, inputting the historical complete training graph corresponding to the historical training data into the graph neural network, obtaining each layer of network nodes of the graph neural network for transmitting the characteristics of the historical training data, collecting the network nodes involved in the transmission of the characteristics of the historical complete training graph in the graph neural network, and determining the transmission order of each historical training data by using the transmission relationship of the network nodes; The current layer network nodes receive the features transmitted from the previous layer network nodes, and directly receive the source network nodes of the historical complete training graph to extract the features of the historical training data; S330, for each rehabilitation device, classify the historical training data of the rehabilitation device according to the sub-device type to obtain classification data; S340, determining a node corresponding to the classification data in the corresponding training subgraph, and using the transmission order of the classification data as attribute information of the node; S350, for each rehabilitation device, determining the execution order of each sub-device according to the order in which the control sub-devices in the rehabilitation device execute the pre-processed training data and the response time of the motion sub-device; S360, using the execution order of each sub-device, adding node connection relationships in the training sub-graph, and determining the training order of each node in the training sub-graph; S370, filling the node attributes, training order and data to be trained into the corresponding nodes of the corresponding training subgraph to obtain the initial training subgraph.

5. The collaborative control method of integrated rehabilitation equipment according to claim 1, characterized in that: The target rehabilitation devices with the highest intelligence among the upper limb rehabilitation devices in S400 include: The communication efficiencies of all upper limb rehabilitation devices are compared to determine the upper limb rehabilitation devices whose communication efficiencies exceed a threshold; A Turing test is performed on upper limb rehabilitation devices of all communication efficiency thresholds, and the upper limb rehabilitation device with the highest accuracy in the test results is selected as the target rehabilitation device with the highest intelligence.

6. The collaborative control method of integrated rehabilitation equipment according to claim 1, characterized in that: S600 includes: S610, receiving a rehabilitation training request sent by any rehabilitation device, and determining pre-processed communication data corresponding to the rehabilitation device; S620, using the communication data preprocessed in S610, determining the communication data format of the rehabilitation device; S630, determining a corresponding communication conversion strategy according to the communication data format; S640: Using the communication conversion strategy, the complete training graph is converted into a communication type suitable for a corresponding rehabilitation device and sent to the rehabilitation device, so that the rehabilitation device completes the rehabilitation training according to the complete training graph.

7. The collaborative control method of integrated rehabilitation equipment according to claim 6, characterized in that: S640 includes: S641, using the communication protocol agreed upon in the communication conversion strategy and the bit data format agreed upon in the communication protocol, convert the complete training graph into a communication type suitable for the corresponding rehabilitation device to obtain a data packet to be transmitted; S642, transmit the data packet to be transmitted to the corresponding rehabilitation device, so that the rehabilitation device parses the transmission data packet to obtain the complete training graph, and restores its own control nodes and motion nodes from the complete training graph, and enables the control sub-device corresponding to the control node to control the motion sub-device corresponding to the motion node in sequence according to the training order in the complete training graph to complete the rehabilitation training.

8. The collaborative control method of integrated rehabilitation equipment according to claim 1, characterized in that: The collaborative control method of the integrated rehabilitation equipment is implemented on a collaborative control system integrated with a processor. The collaborative control system also integrates multiple types of communication modules. The communication module is internally provided with a communication protocol and a communication conversion protocol. The communication module is externally provided with a communication interface adapted to the rehabilitation equipment. The communication interface is connected to the communication interface of the rehabilitation equipment by plugging and unplugging. The communication protocol stipulates the data transmission format between the collaborative control system and the communication interface. The communication conversion protocol stipulates the data conversion method between different rehabilitation equipment.

9. A collaborative control system for integrated rehabilitation equipment, characterized in that: The collaborative control system connects all rehabilitation equipment, and the collaborative control system includes: An acquisition module is configured to acquire the to-be-trained data and communication data of all rehabilitation devices, and perform preprocessing to obtain the preprocessed to-be-trained data and communication data; the rehabilitation devices include upper limb rehabilitation devices and lower limb rehabilitation devices; A construction module is configured to construct a training subgraph by taking, for each rehabilitation device, any sensing sub-device on the rehabilitation device as a sensing node, any controlling sub-device as a controlling node, and any motion sub-device controlled by the controlling sub-device as a motion node; A filling module is configured to fill the pre-processed training data into the corresponding training sub-graph in sequence according to the training order during training to obtain an initial training sub-graph; A determination module is configured to determine a target rehabilitation device with the highest intelligence among the upper limb rehabilitation devices, and use the initial training subgraph of the target rehabilitation device as the main graph, and use the initial training subgraphs of the remaining rehabilitation devices as auxiliary graphs; A composition module is configured to map all auxiliary graphs into the main graph according to the principle of merging nodes with the same training function to form a complete training graph; The rehabilitation module is configured to determine a corresponding communication conversion strategy using the preprocessed communication data, and use the communication conversion strategy to convert the complete training graph into a communication type suitable for the corresponding rehabilitation device and send it to the rehabilitation device, so that the rehabilitation device completes the rehabilitation training according to the complete training graph; The step of determining the target rehabilitation device with the highest intelligence level among the upper limb rehabilitation devices, and using the initial training subgraph of the target rehabilitation device as the main graph, and using the initial training subgraphs of the remaining rehabilitation devices as the auxiliary graphs includes: S510, for any first node of each auxiliary graph, determine whether there is a second node in other auxiliary graphs that has the same training order as the first node, and if so, determine that the first node is a sibling node of the second node; S520, for any first node in the auxiliary graph, determine whether the main graph has a third node with the same node attribute as the first node, and if so, determine whether the training order of the third node is the same, and if so, confirm that the training function of the third node is the same as that of the first node; S530, if there is no target node in the main graph having the same node attribute as the first node, determining whether the to-be-trained data of the target node is the same as that of the first node; if the to-be-trained data is the same as that of the first node, determining that the first node is a child node of the target node; S540, all auxiliary graphs are connected according to brother nodes and parent-child nodes to form a complete auxiliary graph, and according to the principle of merging nodes with the same training function, the complete auxiliary graph is mapped to the main graph to obtain a complete training graph.

Images