Telemedicine network system and surgical robot system

Abstract

The invention provides a telemedicine network system and a surgical robot system. The telemedicine network system comprises a transit server level and a communication control station, the transit server hierarchy comprises a plurality of transit servers deployed between a remote end and a local end, a plurality of routing paths passing through the transit servers are formed, and each routing path passes through one transit server; the communication control station is deployed between the transit server hierarchy and the local end, and the plurality of routing paths are gathered to the communication control station and are connected to the local end through the communication control station; and the communication control station is configured to obtain a comprehensive score according to the reliability scores and the delay scores of the plurality of routing paths, and dynamically select the optimal routing path according to the comprehensive score of the plurality of routing paths based on a preset rule. Through the configuration, the reliability and the stability of the routing path are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of medical devices, and in particular to a remote medical network system and a surgical robot system. Background Art

[0002] In the telemedicine industry (such as remote surgery, remote diagnosis, etc.), the commonly used routing methods for remote real-time communications currently include deterministic routing, dynamic routing and redundant routing.

[0003] The transmission path of the data packet of deterministic routing is determined before the network is running, usually based on static configuration rather than real-time network status. Dynamic routing is a network communication mechanism that optimizes path selection in real time. Its core is to dynamically adjust the data transmission path according to the real-time status of the network (such as latency, bandwidth, and congestion). Redundant routing is a network communication mechanism that transmits critical data in parallel through multiple paths to achieve high reliability and zero-interruption fault tolerance. The core idea is to ensure that any single point failure will not affect the continuity of data transmission through physical or logical redundant design.

[0004] However, deterministic routing requires ensuring the status of each node on the routing path, which is costly. The routing path cannot be changed online, and if the network quality fluctuates, it cannot be responded to quickly. The routing nodes of dynamic routing are determined online, the routing path is less certain, the network quality is prone to fluctuations, and it takes time to update when the topology changes. Redundant routing only guarantees communication reliability, and lacks methods to optimize latency. Summary of the invention

[0005] The purpose of the present invention is to provide a telemedicine network system and a surgical robot system to solve the problem that the existing telemedicine real-time communication routing method is difficult to balance reliability and rapid response to network quality fluctuations.

[0006] To solve the above technical problems, the present invention provides a telemedicine network system, which includes: a relay server layer and a communication control station; The relay server level includes a plurality of relay servers deployed between the remote end and the local end, forming a plurality of routing paths through the relay servers, each of the routing paths passing through one of the relay servers; The communication control station is deployed between the relay server level and the local end, and the plurality of routing paths are aggregated to the communication control station and connected to the local end via the communication control station; The communication control station is configured to obtain a comprehensive score according to the reliability scores and delay scores of the multiple routing paths, and dynamically select the optimal routing path according to the comprehensive scores of the multiple routing paths based on preset rules.

[0007] Optionally, the reliability score is obtained by deep learning and using a multimodal hybrid model to quantitatively evaluate at least one of network quality period data, weather data, and current affairs data.

[0008] Optionally, the multimodal hybrid model includes: a long short-term memory network model, a convolutional neural network model and a bidirectional encoder representation model; wherein The network quality period data is evaluated by the long short-term memory network model; The weather data is evaluated by the convolutional neural network model; The current affairs data is evaluated by the bidirectional encoder representation model; The multimodal hybrid model dynamically weights and sums the evaluation results of the long short-term memory network model, the convolutional neural network model, and the bidirectional encoder representation model to obtain the reliability score.

[0009] Optionally, during the training of the multimodal hybrid model, the influence weights of the network quality period data, the weather data and the current affairs data are dynamically balanced through weighted summation, and the reliability score finally output is normalized to a percentage system.

[0010] Optionally, the delay score is obtained according to the current delay data based on a delay quantification scoring standard; the delay quantification scoring standard includes: If the delay data is not greater than the first delay threshold, the delay score is full marks; If the delay data is greater than a second delay threshold, the delay score is zero; If the delay data is greater than the first delay threshold and not greater than the second delay threshold, the delay score decreases nonlinearly according to the delay data, and a decreasing rate of the delay score increases according to an increase in the delay data.

[0011] Optionally, the delay quantification scoring criteria include:

[0012] Where D is the delay score, t delay is the delay data, th1 is the first delay threshold, and th2 is the second delay threshold.

[0013] Optionally, the step of obtaining the comprehensive score includes: Setting a first scoring weight for the reliability scoring; setting a second scoring weight for the delay scoring; Setting a short board compensation item, wherein the short board compensation item is the smaller one of the reliability score and the delay score, and setting a compensation coefficient for the short board compensation item; The comprehensive score is the sum of the product of the reliability score and the first score weight, the product of the delay score and the second score weight, and the product of the short board compensation item and the compensation coefficient.

[0014] Optionally, at least one of the first scoring weight, the second scoring weight, and the compensation coefficient can be adjusted according to an application scenario of telemedicine.

[0015] Optionally, based on a preset rule, the step of dynamically selecting the optimal routing path according to the comprehensive scores of the plurality of routing paths includes: Before the telemedicine begins, selecting the routing path with the highest comprehensive score; After the telemedicine begins, monitoring the comprehensive score in real time; If the comprehensive score is not lower than the first score threshold, maintaining the current routing path; If the comprehensive score is lower than the first score threshold but not lower than the second score threshold, select the routing path with the highest comprehensive score outside the current routing path as the backup routing path, start pre-connection for the backup routing path, and perform a data transmission test; If the comprehensive score is lower than the second score threshold and the duration reaches the duration limit, switch to the backup routing path.

[0016] In order to solve the above technical problems, the present invention also provides a surgical robot system, which includes a remote end, a local end and the telemedicine network system as described above; the remote end communicates with the local end through the telemedicine network system.

[0017] In summary, in the telemedicine network system and surgical robot system provided by the present invention, the telemedicine network system includes: a relay server hierarchy and a communication control station; the relay server hierarchy includes multiple relay servers deployed between the remote end and the local end, forming multiple routing paths through the relay servers, and each routing path passes through one relay server; the communication control station is deployed between the relay server hierarchy and the local end, and the multiple routing paths are aggregated to the communication control station and connected to the local end via the communication control station; the communication control station is configured to obtain a comprehensive score based on the reliability scores and delay scores of the multiple routing paths, and based on preset rules, dynamically select the optimal routing path according to the comprehensive score of the multiple routing paths.

[0018] With this configuration, on the one hand, compared with deterministic routing, it is not necessary to control each node on the routing path through the setting of the relay server level, which can reduce the number of determined nodes and reduce the cost of routing lines; compared with dynamic routing, since the key nodes at the relay server level are controlled, the communication quality of each path can be monitored in real time, and redundancy is formed between them, which effectively improves the reliability and network communication quality. On the other hand, the selection of routing paths is based on the dual constraints of reliability and delay, which improves the reliability and stability of routing paths. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Those skilled in the art will appreciate that the drawings are provided for a better understanding of the present invention, but do not constitute any limitation on the scope of the present invention.

[0020] Figure 1 A schematic diagram of a locally controlled surgical robotic system.

[0021] Figure 2 is a schematic diagram of the data flow of a locally controlled surgical robotic system.

[0022] Figure 3 A schematic diagram of a remote-controlled surgical robotic system.

[0023] Figure 4 A schematic diagram of the data flow of a remotely controlled surgical robotic system.

[0024] Figure 5 Schematic diagram of a telemedicine network system according to an embodiment of the present invention.

[0025] Figure 6 It is a schematic diagram of the structure of a multimodal hybrid model according to an embodiment of the present invention.

[0026] Figure 7 4 is a schematic diagram of a delay quantization scoring standard according to an embodiment of the present invention.

[0027] Figure 8 It is a flowchart of selecting the optimal routing path based on preset rules according to an embodiment of the present invention.

[0028] Fig. 9 It is a schematic diagram of routing path switching according to an embodiment of the present invention.

[0029] In the accompanying drawings: 10-master control device; 11-master operating arm; 12-display device; 20-slave robot device; 21-robotic arm; 30-imaging device; 40-network system; 41-transit server level; 411-transit server; 42-communication control station; 50-remote end; 51-master control device; 60-local end; 61-slave robot device; 62-imaging device. DETAILED DESCRIPTION

[0030] In order to make the purpose, advantages and features of the present invention clearer, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the drawings are all in a very simplified form and are not drawn to scale, and are only used to conveniently and clearly assist in explaining the purpose of the embodiments of the present invention. In addition, the structure shown in the drawings is often a part of the actual structure. In particular, the emphasis of each drawing is different, and sometimes different scales are used.

[0031] As used in the present invention, the singular forms "a", "an", "one" and "the" include plural objects, the term "or" is usually used to include the meaning of "and / or", the term "several" is usually used to include the meaning of "at least one", and the term "at least two" is usually used to include the meaning of "two or more". In addition, the terms "first", "second" and "third" are used only for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features. Therefore, the features defined as "first", "second" and "third" may explicitly or implicitly include one or at least two of the features, "one end" and "the other end" as well as "proximal end" and "distal end" generally refer to two corresponding parts, which include not only endpoints. In addition, as used in the present invention, "installed", "connected", "connected", and one element "set" on another element should be understood in a broad sense, usually only indicating that there is a connection, coupling, cooperation or transmission relationship between the two elements, and the connection, coupling, cooperation or transmission between the two elements can be direct or indirect through an intermediate element, and cannot be understood as indicating or implying the spatial position relationship between the two elements, that is, one element can be in any orientation such as inside, outside, above, below or one side of another element, unless the content clearly indicates otherwise. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances. In addition, directional terms such as above, below, up, down, upward, downward, left, right, etc. are used relative to the exemplary embodiments as shown in the figures, with the upward or upper direction toward the top of the corresponding figure, and the downward or lower direction toward the bottom of the corresponding figure.

[0032] The purpose of the present invention is to provide a telemedicine network system and a surgical robot system to solve the problem that the existing telemedicine real-time communication routing method is difficult to balance reliability and rapid response to network quality fluctuations. The following is a description with reference to the accompanying drawings.

[0033] Remote medical treatment (such as remote surgery, remote diagnosis, etc.) requires the realization of remote real-time communication of data. The following uses remote robotic surgery as an example to explain. In order to facilitate the understanding of remote robotic surgery, local robotic surgery is first explained here. Please refer to Figure 1 , which exemplarily shows a locally controlled master-slave teleoperated surgical robot system, where the master-slave teleoperation mainly refers to the master-end control device 10 and the slave-end robot device 20, which are configured in a master-slave control relationship. The operator (such as a doctor) can realize the drive control of the slave-end robot device 20 by operating the master operating arm 11 of the master control device 10 through master-slave mapping control. The slave-end robot device 20 includes a number of mechanical arms 21, which can be used to mount instruments (such as surgical instruments or other auxiliary instruments) and image acquisition devices (such as endoscopes). The movements of the master operating arm 11 driven by the operator are mapped to the movements of the instruments and image acquisition devices of the mechanical arm 21 through master-slave mapping, thereby realizing the master-slave mapping operation to perform the surgery. Further, the surgical robot system also includes an image device 30, and the image screen collected by the image acquisition device is fed back to the image device 30, and after being processed by the image device 30, it is transmitted to the display device 12 of the master control device 10 for display for the operator to observe. The operator can operate based on the observed image screen. In application, in addition to forward mapping operation control information, the master control device 10 and the slave robot device 20 can also reversely map feedback information, such as compensating the reaction force of the contact between the instrument and the tissue to the main operating arm 11, thereby forming a closed loop from operation to feedback.

[0034] Please refer to Figure 2 , which shows the data flow of the locally controlled surgical robot system. The data flow of local robotic surgery includes two aspects, one is the motion control signal data flow, and the other is the visual signal data flow. The motion control signal data flow is: starting from the main control arm 11 of the master control device 10, the master controller of the master control device 10, after receiving the operator's action signal, converts it into the motion control instruction of the instrument and sends it to the slave controller of the slave robot device 20, and the slave controller controls the mechanical arm 21 and the mounted instrument to make corresponding actions. The visual signal data flow is that the image acquisition device installed on the mechanical arm 21 of the slave robot device 20 collects the light signal and transmits it to the image host in the image device 30, where it is converted into a visual signal and then sent to the display device 12 of the master control device 10 for display, so that the operator can observe the real-time picture of the surgical field.

[0035] After learning about the locally controlled surgical robotic system, please refer to Figure 3, which shows a remote-controlled surgical robot system, which includes a remote terminal 50 (such as a remote operating room) and a local terminal 60 (such as a local operating room). The remote terminal 50 at least includes a master control device 51, and the master control device 51 of the remote terminal 50 is connected to the local terminal 60. Figure 1 The master control device 10 shown in the figure can be of the same or similar structure. The local end 60 at least includes a slave robot device 61 and an image device 62. The slave robot device 61 and the image device 62 of the local end 60 are Figure 1 The slave robot device 20 and the image device 30 shown can be of the same or similar structure. The remote end 50 and the local end 60 are configured to exchange data via the network system 40 to achieve remote operation.

[0036] Please refer to Figure 4 For the data flow of the remote-controlled surgical robot system, the motion control signal is sent from the master controller of the master control device 51 of the remote end 50, and is sent to the slave controller of the slave robot device 61 of the local end 60 via the network system 40. The visual signal is sent from the image device 62 of the local end 60, and is sent to the display device of the master control device 51 of the remote end 50 via the network system 40. Figure 4 It can be seen that during remote robotic surgery, the communication quality of the network system 40 has a great impact on the surgical process. Therefore, it is necessary to optimize the network routing during remote robotic surgery in real time, while reducing the delay and enhancing the reliability of the routing path, to provide guarantee for the smooth progress of remote surgery.

[0037] Of course, telemedicine is not limited to remote robotic surgery, but can also include remote diagnosis, etc., which also requires one-way or two-way real-time transmission of data (such as voice data or image data, etc.) through the network system 40.

[0038] To optimize the network system for telemedicine, please refer to Figure 5, an embodiment of the present invention provides a telemedicine network system 40, which includes: a relay server layer 41 and a communication control station 42; the relay server layer 41 includes a plurality of relay servers 411 deployed between a remote end 50 and a local end 60, forming a plurality of routing paths through the relay servers 411, and each of the routing paths passes through one of the relay servers 411; the communication control station 42 is deployed between the relay server layer 41 and the local end 60, and the plurality of routing paths are aggregated to the communication control station 42 and connected to the local end 60 via the communication control station 42; the communication control station 42 is configured to obtain a comprehensive score according to the reliability scores and delay scores of the plurality of routing paths, and dynamically select the optimal routing path according to the comprehensive scores of the plurality of routing paths based on preset rules.

[0039] With such configuration, on the one hand, by setting up the relay server level 41, compared with deterministic routing, it is not necessary to control each node on the routing path, which can reduce the number of determined nodes (each routing path only needs to control one relay server 411), and reduce the cost of routing lines; compared with dynamic routing, since the key nodes of the relay server level 41 (referring to the relay server 411 on each routing path) are controlled, the communication quality of each path can be monitored in real time, and redundancy is formed between each other, which effectively improves the reliability and network communication quality. On the other hand, the selection of routing paths is based on the dual constraints of reliability and delay, which improves the reliability and stability of routing paths.

[0040] The reliability score is explained below. In an alternative exemplary embodiment, the reliability score is obtained by deep learning, using a multimodal hybrid model to quantitatively evaluate at least one of the network quality time period data, weather data, and current affairs data. The inventors have found that the time period, weather conditions, politics, and other factors where the node (such as the transit server 411) is located have a certain impact on the communication quality. In the application, the time period situation of the current node location can be obtained by obtaining the network quality time period data (such as obtained through the device log), the weather conditions of the current node location can be obtained through weather data (such as can be captured through a meteorological API), and the political situation of the current node location can be obtained through current affairs data (such as can be captured through a news API).

[0041] After collecting network quality period data, weather data, and political data, the data can be preprocessed. Data preprocessing includes time alignment, feature engineering, and label generation.

[0042] Time alignment can unify data of different frequencies to the same time granularity.

[0043] Feature engineering may optionally include time period features, weather features, and political features. In one embodiment, time period features include using sin / cos to encode periodic information such as hours and weeks. Weather features include associating meteorological data with the geographic location of routing nodes, such as the real-time weather in the city where a data center is located. Political features include extracting event impact weights through NLP technology, such as event sentiment analysis + entity recognition based on bidirectional encoder representation (BERT).

[0044] Label generation can use network performance metrics (such as latency increase percentage) as supervisory signals.

[0045] After completing data preprocessing, a multimodal hybrid model can be used for quantitative evaluation. Optionally, the multimodal hybrid model includes: a long short-term memory network model (LSTM), a convolutional neural network model (CNN) and a bidirectional encoder representation model (BERT); wherein the network quality period data is evaluated by the long short-term memory network model (LSTM), which captures periodic laws and outputs a period score T LSTM The weather data is evaluated by the convolutional neural network model (CNN), which extracts local meteorological patterns and outputs a weather score W CNN The current affairs data is evaluated by the bidirectional encoder representation model (BERT), which processes news events and policy documents through text analysis and outputs a political score P BERT .

[0046] Furthermore, the multimodal hybrid model dynamically weights and sums the evaluation results of the long short-term memory network model (LSTM), the convolutional neural network model (CNN), and the bidirectional encoder representation model (BERT) to obtain the reliability score R. The calculation formula is: R = α·T LSTM +β·W CNN +γ·P BERT Among them, α, β, and γ are dynamic weights, and their values ​​can be dynamically generated through the attention mechanism. The structure of this multimodal hybrid model is as follows: Figure 6 As shown. It is understandable that if one of the network quality period data, weather data and current affairs data does not need to participate in the quantitative evaluation, its dynamic weight can be reset to zero. The following is an example of using a multimodal hybrid model to output reliability scores.

[0047] Period Rating: T LSTM =85 (night surgery low load period); Weather rating: W CNN =40 (typhoon caused regional network instability); Political rating: PBERT =70 (no major policy risk); Dynamic weight: α=0.5, β=0.3, γ=0.2 (it is currently the typhoon season, and the weather weight is increased) The reliability score R = 0.5 × 85 + 0.3 × 40 + 0.2 × 70 = 68.5, which is 68.5 / 100 after normalization. It can be seen that the reliability score of this routing path is relatively low (68.5).

[0048] Before the multimodal hybrid model is applied, it needs to be trained. Optionally, during the training of the multimodal hybrid model, the influence weights of the network quality period data, the weather data and the current affairs data are dynamically balanced by weighted summation, and the reliability score outputted finally is normalized to a percentage system.

[0049] In an optional example, the mean square error (MSE) can be used as a loss function to optimize the regression accuracy of the reliability score of the multimodal hybrid model. L2 norm regularization can be optionally introduced to prevent model overfitting. In telemedicine applications, the sample size of routing data may be limited, and L2 regularization can avoid the model's sensitive dependence on a small amount of abnormal data (such as an extreme weather event). The weighted sum can be used to dynamically balance the influence weights of time periods, weather, and political factors. For example, in areas with high incidence of typhoons, the weather score weight β can be increased; in policy-sensitive areas, the political score weight γ can be increased. The final output reliability score is normalized to a percentage system, and the scoring scale is unified, which is convenient for weighted calculation with the delay score (also a percentage system, see the following description for details), so as to facilitate application in subsequent quantitative evaluation steps. In addition, normalization to a percentage system is conducive to intuitively displaying the reliability of the routing path (such as 90 points can be considered excellent, and 60 points can be considered to have certain risks) to assist in quick decision-making. Of course, the loss function is not limited to the mean square error (MSE). In some other embodiments, the mean absolute error MAE, Huber loss (referring to a mixed MSE and MAE method) or quantile loss (Quantile Loss) can also be used. Technical personnel in this field can select a suitable loss function according to actual needs, or combine multiple loss functions or customize a new loss function to meet specific task requirements.

[0050] The delay score is explained below. Telemedicine, especially application scenarios such as remote robotic surgery, is very sensitive to network delays. Therefore, it is necessary to design a nonlinear delay quantization scheme that severely reduces the score for high delays. In one embodiment, the delay score is obtained based on the current delay data and the delay quantization scoring criteria; the delay quantization scoring criteria include: If the delay data is not greater than the first delay threshold, the delay score is full marks; If the delay data is greater than a second delay threshold, the delay score is zero; If the delay data is greater than the first delay threshold and not greater than the second delay threshold, the delay score decreases nonlinearly according to the delay data, and the decreasing rate of the delay score increases according to the increase of the delay data, wherein the first delay threshold is less than the second delay threshold.

[0051] In application scenarios such as remote robotic surgery, two different levels of delay thresholds (first delay threshold and second delay threshold) can be set. Delays less than the first delay threshold (such as 50ms) have little effect on the surgery, so no matter what the actual value of the delay data is at this time, the delay score is full marks. Delays greater than the second delay threshold (200ms) will seriously affect the surgery, so no matter what the actual value of the delay data is at this time, the delay score is zero. When the delay data is between the first delay threshold and the second delay threshold, the delay score decreases nonlinearly according to the delay data, and the rate of decrease of the delay score increases according to the increase of the delay data. Specifically, in the area close to the first delay threshold (relatively low delay area, such as 50ms~100ms), the delay score decreases more slowly, and the slight delay difference has little effect on the score. In the area close to the second delay threshold (relatively high delay area, such as 150ms~200ms), the score drops sharply.

[0052] In one example, the delay quantification scoring criteria include:

[0053] Where D is the delay score, t delay is the delay data, th1 is the first delay threshold, and th2 is the second delay threshold. The delay data t according to the delay quantification scoring standard delay With delayed score delay The relationship is as Figure 7 As shown. It is understandable that those skilled in the art can set the first delay threshold and the second delay threshold according to the specific application scenario of actual telemedicine. Preferably, the delay score is also normalized to a percentage system to unify the dimensions.

[0054] Optionally, the step of obtaining the comprehensive score includes: setting a first scoring weight for the reliability score; setting a second scoring weight for the delay score; setting a short board compensation item, the short board compensation item being the smaller of the reliability score and the delay score, and setting a compensation coefficient for the short board compensation item; the comprehensive score is the sum of the product of the reliability score and the first scoring weight, the product of the delay score and the second scoring weight, and the product of the short board compensation item and the compensation coefficient.

[0055] Based on the reliability score R and delay score D of the routing path, the final comprehensive score of the routing path can be obtained through comprehensive analysis. The comprehensive score is oriented to the dual constraints of reliability and delay. At the same time, the coupling effect is eliminated by introducing short board compensation items to avoid the single dimension score dominating the decision. The calculation formula of the comprehensive score C is: C = w r ·R +w d ·D+μ·min(D,R) Among them, w r is the first scoring weight (i.e., the weight of the reliability score), w d is the second scoring weight (i.e., the weight of delayed scoring), μ is the compensation coefficient of the short board compensation item. By introducing the short board compensation item, it can avoid being misselected due to a single item scoring being too high.

[0056] Optionally, the first scoring weight w r The second scoring weight w d And at least one of the compensation coefficients μ is adjustable according to the application scenario of telemedicine. The weights in different application scenarios need to be set differently and can be dynamically adjusted according to the change of application scenarios. Preferably, the switching adjustment of the weight adopts the Sigmoid function transition (transition time ≤ 50ms) to avoid path oscillation caused by score jumps, so as to achieve smooth transition processing of the switching. The exemplary weight selection of some typical scenarios is shown in Table 1 below: Table 1 Application Scenario <![CDATA[w r ]]> <![CDATA[w d ]]> μ Weighting principle Only one-way image signal transmission 0.6 0.3 0.1 <![CDATA[At this time, the motion control signal is not transmitted through the network, so the weight w of the delay score d decreases]]> Simultaneously transmit image signals and motion control command signals 0.4 0.5 0.1 Motion control signals and image signals need to be transmitted simultaneously, and latency and reliability are important indicators. There is foreseeable network congestion 0.1 0.7 0.2 Compensate for reliability loss by forcing congested nodes to be avoided After obtaining the comprehensive scores of multiple routing paths, it is necessary to make a decision to select a routing path based on the comprehensive scores based on the preset rules. The decision logic optimization includes the main path optimization rules and the predictive switching mechanism.

[0057] Optionally, based on a preset rule, the step of dynamically selecting the optimal routing path according to the comprehensive scores of the plurality of routing paths includes: Before the telemedicine begins, selecting the routing path with the highest comprehensive score; After the telemedicine begins, monitoring the comprehensive score in real time; If the comprehensive score is not lower than the first score threshold, maintaining the current routing path; If the comprehensive score is lower than the first score threshold but not lower than the second score threshold, select the routing path with the highest comprehensive score outside the current routing path as the backup routing path, start pre-connection for the backup routing path, and perform a data transmission test; If the comprehensive score is lower than the second score threshold and the duration reaches the duration limit, switch to the backup routing path.

[0058] Main path optimization rule: Before telemedicine begins, the comprehensive scores of each routing path are sampled for a period of time and the average is taken. Before telemedicine begins, the routing path with the highest comprehensive score is selected. Then, during the telemedicine process, the comprehensive scores of each routing path are updated in real time. If the network quality fluctuates during the telemedicine process, the routing path is switched according to the predictive switching mechanism.

[0059] Predictive switching mechanism: Set two different levels of scoring thresholds (a first scoring threshold and a second scoring threshold, where the first scoring threshold is greater than the second scoring threshold).

[0060] Please refer to Figure 8 ,In an example, the first scoring threshold is set to 75, and the second scoring threshold is set to 70. When the comprehensive score C ≥ 75, the telemedicine data communication can be carried out normally and the current routing path is maintained.

[0061] When 70≤C<75, the routing path with the highest comprehensive score other than the current routing path is selected as the backup routing path, a pre-connection is started for it, and a data transmission test is performed.

[0062] When C < 70 and continues to exceed the time limit (such as 10s), it is forced to switch to the backup routing path. The switching delay is preferably no more than 8ms. Of course, if the comprehensive score of the current routing path returns to above 75 after passing through the interval of 70 ≤ C < 75, it is not necessary to switch to the backup routing path, and the pre-connection and data transmission test of the backup routing path is stopped.

[0063] Optionally, if during the medical process, after the comprehensive score of the current routing path falls into the interval of 70≤C<75, if no other routing path meets the requirements (such as the comprehensive score ≥75), a network quality risk is prompted.

[0064] Please refer to Fig. 9 , which shows an example of routing path switching. Before the start of telemedicine, the routing path with the highest score is remote end 50 - relay server 411a - communication control station 42 - local end 60. It is selected as the optimal routing path for data transmission.

[0065] In telemedicine, if network congestion occurs between the relay server 411a and the communication control station 42, or if a disaster occurs and the network cable is damaged, the comprehensive score of the routing path will quickly drop below 70. At this time, according to the preset rules, the optional routing path with the highest comprehensive score can be immediately switched to, for example, the remote end 50 - relay server 411b - communication control station 42 - local end 60. This ensures that the impact of network quality fluctuations on telemedicine is minimized.

[0066] The embodiment of the present invention provides a surgical robot system, which includes a remote terminal 50, a local terminal 60, and the telemedicine network system 40 as described above; the remote terminal 50 communicates with the local terminal 60 through the telemedicine network system 40. The remote terminal 50 and the local terminal 60 are configured in a master-slave control relationship. The structure and principle of other components of the surgical robot system can be understood by referring to the above and combining with the prior art, and will not be described in detail here.

[0067] In summary, in the telemedicine network system and surgical robot system provided by the present invention, the telemedicine network system includes: a relay server hierarchy and a communication control station; the relay server hierarchy includes multiple relay servers deployed between the remote end and the local end, forming multiple routing paths through the relay servers, and each routing path passes through one relay server; the communication control station is deployed between the relay server hierarchy and the local end, and the multiple routing paths are aggregated to the communication control station and connected to the local end via the communication control station; the communication control station is configured to obtain a comprehensive score based on the reliability scores and delay scores of the multiple routing paths, and based on preset rules, dynamically select the optimal routing path according to the comprehensive score of the multiple routing paths. With this configuration, on the one hand, compared with deterministic routing, it is not necessary to control each node on the routing path through the setting of the relay server level, which can reduce the number of determined nodes and reduce the cost of routing lines; compared with dynamic routing, since the key nodes at the relay server level are controlled, the communication quality of each path can be monitored in real time, and redundancy is formed between them, which effectively improves the reliability and network communication quality. On the other hand, the selection of routing paths is based on the dual constraints of reliability and delay, which improves the reliability and stability of routing paths.

[0068] It should be noted that the above embodiments can be combined with each other. The above description is only a description of the preferred embodiments of the present invention, and does not limit the scope of the present invention. Any changes and modifications made by ordinary technicians in the field of the present invention based on the above disclosure are within the protection scope of the present invention.

Claims

1. A telemedicine network system, characterized in that: include: Relay server level and communication control station; The relay server level includes a plurality of relay servers deployed between the remote end and the local end, forming a plurality of routing paths through the relay servers, each of the routing paths passing through one of the relay servers; The communication control station is deployed between the relay server level and the local end, and the plurality of routing paths are aggregated to the communication control station and connected to the local end via the communication control station; The communication control station is configured to obtain a comprehensive score according to the reliability scores and delay scores of the multiple routing paths, and dynamically select the optimal routing path according to the comprehensive scores of the multiple routing paths based on preset rules.

2. The telemedicine network system according to claim 1, characterized in that: The reliability score is obtained by deep learning and using a multimodal hybrid model to quantitatively evaluate at least one of network quality period data, weather data, and current affairs data.

3. The telemedicine network system according to claim 2, characterized in that: The multimodal hybrid model includes: a long short-term memory network model, a convolutional neural network model and a bidirectional encoder representation model; wherein The network quality period data is evaluated by the long short-term memory network model; The weather data is evaluated by the convolutional neural network model; The current affairs data is evaluated by the bidirectional encoder representation model; The multimodal hybrid model dynamically weights and sums the evaluation results of the long short-term memory network model, the convolutional neural network model, and the bidirectional encoder representation model to obtain the reliability score.

4. The telemedicine network system according to claim 2, characterized in that: During the training process of the multimodal hybrid model, the influence weights of the network quality period data, the weather data and the current affairs data are dynamically balanced through weighted summation, and the reliability score finally output is normalized to a percentage system.

5. The telemedicine network system according to claim 1, characterized in that: The delay score is obtained according to the current delay data based on the delay quantification scoring standard; the delay quantification scoring standard includes: If the delay data is not greater than the first delay threshold, the delay score is full marks; If the delay data is greater than a second delay threshold, the delay score is zero; If the delay data is greater than the first delay threshold and not greater than the second delay threshold, the delay score decreases nonlinearly according to the delay data, and a decreasing rate of the delay score increases according to an increase in the delay data.

6. The telemedicine network system according to claim 5, characterized in that: The delay quantification scoring criteria include: Where D is the delay score, t delay is the delay data, th1 is the first delay threshold, and th2 is the second delay threshold.

7. The telemedicine network system according to claim 1, characterized in that: The steps of obtaining the comprehensive score include: Setting a first scoring weight for the reliability scoring; setting a second scoring weight for the delay scoring; Setting a short board compensation item, wherein the short board compensation item is the smaller one of the reliability score and the delay score, and setting a compensation coefficient for the short board compensation item; The comprehensive score is the sum of the product of the reliability score and the first score weight, the product of the delay score and the second score weight, and the product of the short board compensation item and the compensation coefficient.

8. The telemedicine network system according to claim 7, characterized in that: At least one of the first scoring weight, the second scoring weight, and the compensation coefficient may be adjustably set according to an application scenario of telemedicine.

9. The telemedicine network system according to claim 1, characterized in that: Based on the preset rules, the step of dynamically selecting the optimal routing path according to the comprehensive scores of the multiple routing paths includes: Before the telemedicine begins, selecting the routing path with the highest comprehensive score; After the telemedicine begins, monitoring the comprehensive score in real time; If the comprehensive score is not lower than the first score threshold, maintaining the current routing path; If the comprehensive score is lower than the first score threshold but not lower than the second score threshold, select the routing path with the highest comprehensive score outside the current routing path as the backup routing path, start pre-connection for the backup routing path, and perform a data transmission test; If the comprehensive score is lower than the second score threshold and the duration reaches the duration limit, switch to the backup routing path.

10. A surgical robot system, characterized in that: It comprises a remote end, a local end and a telemedicine network system according to any one of claims 1 to 9; the remote end communicates with the local end through the telemedicine network system.

Images