Multi-module Switching System Based on AI Model and Its Application in Health Care Robots

Abstract

The present invention relates to the field of health care robots, specifically a multi-module switching system based on an AI model and its application in health care robots, including: an attitude sensing module, a signal cleaning module, a model approximation module, a reaction output module, and a memory training module. The attitude sensing module is used to pre-transmit standard signal sites. The signal cleaning module is used to generate random Gaussian noise to train the AI model. The model approximation module is used to calculate the matching degree between the robot sensing signal and each module. The reaction output module is used to activate the robot's actions. The memory training module is used to construct a reaction memory bank and adjust the subsequent matching weights. The present invention can make the signals collected by the sensor clearer and more stable, improve the signal quality, reduce the input error of the sensor, optimize the resource utilization efficiency of the system, improve the overall performance, and at the same time improve the flexibility and adaptability of the AI model, and improve the stability and reliability of the health care robot system.

Description

Technical Field

[0001] The present invention relates to the field of health care robots, and specifically to a multi-module switching system based on an AI model and its application in health care robots. Background Art

[0002] A health care robot is an intelligent robot that provides rehabilitation and nursing services. This kind of robot can monitor the physiological state and activities of users through sensors and cameras, and provide customized rehabilitation plans and nursing services for the elderly or injured people, realizing functions such as rehabilitation training and health status monitoring.

[0003] The health care robot perceives the external state by identifying the potential points in the sensor, analyzes the electrical signals, and then drives the robot to perform action output. Since the physical states and behavioral characteristics of the service objects are changing, different modules need to be equipped for the robot. However, at present, most health care robots do not have a unified standard for the conditions of switching modules, nor can they provide personalized module switching solutions for users.

[0004] In addition, due to the relatively complex service environment of the health care robot, the complex environment is likely to cause different degrees of interference in the sensing signals of the sensors, making the robot unable to accurately obtain the external sensing information, affecting the module switching judgment and action output efficiency of the robot, and greatly reducing the adaptability and service quality of the robot. Summary of the Invention

[0005] The purpose of the present invention is to provide a multi-module switching system based on an AI model and its application in health care robots to solve the problems raised in the above background art.

[0006] To solve the above technical problems, the present invention provides the following technical solution: A multi-module switching system based on an AI model, including: an attitude sensing module, a signal cleaning module, a model approximation module, a reaction output module, and a memory training module;

[0007] The attitude sensing module is used to pre-transmit standard signal sites in the attitude sensor of the robot. Each type of electrical signal site corresponds to specific body parameters, behavior parameters, and attitude parameters of the health care object, obtain historical classification results, generate a historical site set of each type of electrical signal, and calculate the aggregation degree of the set. The aggregation degree reflects the average intensity of environmental noise;

[0008] The signal cleaning module is used to add random Gaussian noise with a divergence equal to the aggregation degree to the standard electrical signal sites according to the set aggregation degree to obtain a random noise set. Using the random noise set as a training library and the standard electrical signal as the output, an AI model is trained, and the trained AI model is used to analyze the electrical signal sites of the attitude sensor and convert them into corresponding standard electrical signal outputs;

[0009] The model approximation module is used to collect all the sensing data within one action cycle of the robot, simulate the multi-dimensional distribution of the electrical signal sites corresponding to all the sensing data in the virtual space, perform homogenization operations with the multi-dimensional distribution models between the various healthcare modules to obtain the approximation degrees of the multi-dimensional distribution to the various healthcare modules, and arrange all the modules in the order of the approximation degrees;

[0010] The reaction output module is used to generate a robot reaction library based on the approximation degrees of the modules and the corresponding outputs of the various modules. Using each recognized electrical signal site as a key value, match the corresponding outputs of the key value in the various healthcare modules, calculate the output value of the matching results in the various modules according to the approximation degrees of the modules and the matching consistency, and activate the robot reactions in sequence according to the output values;

[0011] The memory training module is used to perform convergence training on the robot reactions triggered by adjacent electrical signal sites to form a reaction memory library, and adjust the approximation degree weights of the healthcare modules and the diffusion divergence of the signal recognition model according to the matching degree of the activated electrical signal sites in the reaction memory library.

[0012] Further, the attitude sensing module includes: a signal site unit and an environment aggregation unit;

[0013] The signal site unit is used to format the sensors of the healthcare robot, determine the standard signal sites according to the sensor hardware parameters, and be uniformly managed by the main control chip;

[0014] The environment aggregation unit is used to record the actual electrical signal sites of the sensors each time the sensors transmit signals back, generate a set linked list in the database, and calculate the aggregation degree of the set.

[0015] Further, the signal cleaning module includes: a random noise unit and a noise removal training unit;

[0016] The random noise unit is used to generate random Gaussian noise with a divergence equal to the aggregation degree according to the aggregation degree of the elements in the historical site set;

[0017] The noise removal training unit is used to add the random Gaussian noise into the standard signal sites to generate training noise, and use the training noise to train an AI classification model so that the AI classification model has the data cleaning ability.

[0018] Further, the model approximation module includes: a spatial distribution unit, a multi-simulation integration unit, and a model switching unit;

[0019] The spatial distribution unit is used to spread the electrical signal points of all sensors in the virtual space to form an actual signal model;

[0020] The multi-simulation integration unit is used to simulate the service intervals of each module of the robot in the virtual space, calculate the proportion of the part of the actual signal model located within the service interval, and output it as the approximation degree;

[0021] The model switching unit arranges all modules in the order of the approximation degree, and switches the service module of the robot to the first module.

[0022] Further, the reaction output module includes: a key-value matching unit and a behavior activation unit;

[0023] The key-value matching unit is used to identify the subsequent electrical signal points, use the subsequent electrical signal points as key values, weighted accumulate the output actions of each module according to the mapping amount and weight, and match the corresponding output in the current service module;

[0024] The behavior activation unit is used to activate the output action of the robot according to the matching result, and end the service cycle after a fixed duration when the robot makes the output action.

[0025] Further, the memory training module includes: a convergence classification unit, a memory bank unit, and a memory weighting unit;

[0026] The convergence classification unit is used to perform convergence training on the electrical signal data and the robot output using an AI model;

[0027] The memory bank unit is used to record the electrical signal points of each sensor and the robot actions during the service cycle in the database;

[0028] The memory weighting unit is used to adjust the matching weights of each module using the memory bank according to the matching degree of the activated electrical signal points in the memory bank during the subsequent action cycle.

[0029] The application of the multi-module switching system based on the AI model in the health care robot includes the following steps:

[0030] Step S1. Determine the standard signal points according to the hardware parameters of the attitude sensors in the health care robot, record the actual electrical signal points of the sensors each time the sensors transmit signals, and generate a signal set in the database;

[0031] Step S2. Calculate the aggregation degree of the signal set, generate random Gaussian noise with the same dispersion and aggregation degree, add the random Gaussian noise to the standard signal sites to obtain a noise library, and use the noise library to train the AI model to obtain an AI model with data cleaning ability;

[0032] Step S3. Collect all the sensing data of the robot during the service cycle, disperse the sensing data in the virtual space to obtain an actual signal model, calculate the mapping amount of the actual signal model in the space of each service module, and arrange the service modules in descending order according to the mapping amount;

[0033] Step S4. Switch the working module of the robot to the service module ranked first, use the subsequent electrical signal sites as key values, weighted accumulate the output actions of each module according to the mapping amount and weight to obtain the value of each action, activate the robot to output the action with the highest value, and end the current cycle after a fixed duration when the robot makes the output action;

[0034] Step S5. Record the electrical signal sites and actual action outputs of all sensors during the cycle to form a memory library, and adjust the weights of each action according to the sensing information uploaded during the cycle in the subsequent service cycles.

[0035] Further, Step S1 includes:

[0036] Step S11. Format each sensor of the healthcare robot, determine the standard signal sites according to the hardware parameters of the sensor, and the hardware parameters of the sensor include: sensor piezoelectric efficiency, temperature and humidity sensing efficiency, and human body monitoring data conversion efficiency, so that the standard signal site of each sensor corresponds to the specific body parameters, behavior parameters, and posture parameters of the healthcare object;

[0037] Step S12. When the sensor generates corresponding electrical signals according to the external physical changes, record the level intensity values of each electrical signal site in the database, record them in the database, and store all the recorded electrical signal site data in a set form to form a signal set A, where A = {T1, T2,..., Tn}, n represents the number of records, and Tn represents the level intensity of the electrical signal site in the nth record.

[0038] Further, Step S2 includes:

[0039] Step S21. Calculate the aggregation degree of the signal set, and the aggregation degree represents the average of the square of the difference between the elements in the set and the standard signal site, satisfying I = 1 / n · [(T1 - T0) 2 + (T2 - T0) 2 +... + (Tn - T0) 2 , where I is the aggregation degree and T0 is the level intensity of the standard signal site;

[0040] Step S22. Generate random Gaussian noise according to the degree of polymerization, so that the generated noise satisfies the following conditions:

[0041]

[0042] where F(t) represents the random Gaussian noise, t represents time, t0 represents the moment when the sensing signal occurs, and e is the base of the natural logarithm;

[0043] Step S23. Randomly change the value of t to obtain a set of Gaussian noise. Add each element in the set of random Gaussian noise into the standard signal sites to generate a noise library. Use the noise library to train the AI model so that when the AI model inputs the signal sites, it automatically outputs the standard signal site with the closest level intensity, realizing the cleaning process of the signal sites. The AI model includes: Mask R-CNN model, DeepLab model, U-Net model, and Watershed model.

[0044] Further, step S3 includes:

[0045] Step S31. Obtain the electrical signals of each sensor of the robot during the service cycle and clean the electrical signals. There is only a unique electrical signal for the same sensor during the service cycle. When the same sensor transmits multiple electrical signals during the service cycle, end the current service cycle and allocate the redundant electrical signals to the next service cycle;

[0046] Step S32. Generate an m-dimensional virtual space in the control chip and label the collected electrical signals in vector form, where m represents the number of electrical signals during the service cycle, to obtain the actual signal model V. The V = (d1, d2,..., dm), where dm represents the level intensity of the electrical signal transmitted by the mth sensor;

[0047] Step S33. Expand the robot module in interval form, represent the robot module as {[D1+, D1-], [D2+, D2-],..., [Dn+, Dn-]}, where D1+, D2+, and Dn+ respectively represent the upper bounds of the intervals of the level intensities of the electrical signals transmitted by sensors numbered 1, 2, and n, and D1-, D2-, and Dn- respectively represent the lower bounds of the intervals of the level intensities of the electrical signals transmitted by sensors numbered 1, 2, and n. Calculate the mapping vector of the actual signal model V in the n-dimensional space, record the modulus length of the part where the mapping vector falls within the intervals where each robot module is located as the mapping quantity of the actual signal model in the space of each module, and arrange the modules in descending order according to the mapping quantity.

[0048] Further, step S4 includes:

[0049] Step S41. Switch the working module of the robot to the first module with the highest mapping quantity. After receiving the subsequent signal, calculate the value of each action according to the following formula:

[0050]

[0051] where E represents the value of the action, k represents the number of service modules of the robot, u i represents the weight of the i-th service module, p i represents the probability that the action is called in the i-th service module, d i represents the influence parameter of the subsequent signal on the actual signal model, which is determined by the rules of the AI model, r i represents the mapping quantity of the actual signal model in the i-th service module;

[0052] Step S42. Select the action with the highest value as the output action for the current service cycle, activate the robot to execute the output action, and end the service cycle after a fixed duration of the output action of the robot.

[0053] Further, step S5 includes:

[0054] Step S51. Record the electrical signal sites of each sensor and the robot actions during the service cycle, and use the AI model to perform convergence training on the electrical signal data and the robot output. When the convergence training result is positive feedback, store the electrical signal sites of the sensor and the robot actions as a memory pattern in the memory bank;

[0055] Step S52. In the subsequent action cycle, calculate the matching degree of each electrical signal site in the actual signal model in the memory bank, and adjust the weights of the actions recorded in the memory unit in each service module according to the matching degree of the memory unit.

[0056] Compared with the prior art, the beneficial effects achieved by the present invention are:

[0057] 1. The present invention can pre-transmit the electrical signal sites in the attitude sensor, and for the historical site set of each type of electrical signal, add random noise with a distribution trend equal to the aggregation degree to the pre-transmitted sites according to the set aggregation degree, clean the recognition results of the sensor, which can make the signals collected by the sensor clearer and more stable, thereby improving the signal quality and reducing the input error of the sensor.

[0058] 2. The present invention can simulate the multi-dimensional distribution of electrical signal sites in space, fit the approach degrees of each health care module, use the identified electrical signal sites as key values, match the corresponding outputs of the key values in each health care module, and sequentially activate the behavioral responses of the robot, enabling the health care robot to quickly switch in different scenarios or tasks, enhancing its ability to adapt to different environments and tasks, flexibly configuring the resource usage of the health care robot, optimizing the resource utilization efficiency of the system, and improving the overall performance.

[0059] 3. The present invention can perform convergence training on the robot responses triggered by adjacent electrical signal sites, and adjust the approach degree weights of the health care modules according to the classification strength of the activated electrical signal sites in the response memory bank, which can improve the flexibility and adaptability of the AI model and enhance the stability and reliability of the health care robot system. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] The drawings are used to provide a further understanding of the present invention, and constitute a part of the specification. They are used together with the embodiments of the present invention to explain the present invention, but do not constitute a limitation to the present invention. In the drawings:

[0061] Figure 1 is a schematic structural diagram of the multi-module switching system based on the AI model of the present invention;

[0062] Figure 2 is a schematic step diagram of the application of the multi-module switching system based on the AI model of the present invention in a health care robot. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0063] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0064] Please refer to Figure 1 , the present invention provides a technical solution: a multi-module switching system based on an AI model, including: an attitude sensing module, a signal cleaning module, a model approach module, a reaction output module, and a memory training module;

[0065] The attitude sensing module is used to pre-transmit standard signal sites in the attitude sensor of the robot. Each type of electrical signal site corresponds to specific body parameters, behavior parameters, and attitude parameters of the health care object, obtain historical classification results, generate a historical site set of each type of electrical signal, and calculate the aggregation degree of the set, and the aggregation degree reflects the average intensity of environmental noise;

[0066] The attitude sensing module includes: a signal site unit and an environmental aggregation unit;

[0067] The signal site unit is used to format the sensors of the health care robot, determine the standard signal sites according to the sensor hardware parameters, and be uniformly managed by the main control chip;

[0068] The environment aggregation unit is used to record the actual electrical signal sites of the sensors each time the sensors transmit signals back, generate a set linked list in the database, and calculate the aggregation degree of the set.

[0069] The signal cleaning module is used to add random Gaussian noise with a divergence equal to the aggregation degree to the standard electrical signal sites according to the set aggregation degree, obtain a random noise set, use the random noise set as a training library, take the standard electrical signal as the output, train the AI model, and use the trained AI model to analyze the electrical signal sites of the attitude sensor and convert them into corresponding standard electrical signals for output;

[0070] The signal cleaning module includes: a random noise unit and a de - impurity training unit;

[0071] The random noise unit is used to generate random Gaussian noise with a divergence equal to the aggregation degree according to the aggregation degree of the elements in the historical site set;

[0072] The de - impurity training unit is used to add the random Gaussian noise into the standard signal sites to generate training noise, and use the training noise to train the AI classification model so that the AI classification model has the data cleaning ability.

[0073] The model approximation module is used to collect all the sensing data within one action cycle of the robot, simulate the multi - dimensional distribution of the electrical signal sites corresponding to all the sensing data in the virtual space, perform an assimilation operation with the multi - dimensional distribution models between the various health care modules, obtain the approximation degrees between the multi - dimensional distribution and the various health care modules, and arrange all the modules in the order of the approximation degrees;

[0074] The model approximation module includes: a space distribution unit, a multi - simulation integration unit, and a model switching unit;

[0075] The space distribution unit is used to scatter the electrical signal sites of all the sensors in the virtual space to form an actual signal model;

[0076] The multi - simulation integration unit is used to simulate the service intervals of each module of the robot in the virtual space, calculate the proportion of the part of the actual signal model located within the service intervals, and output it as the approximation degree;

[0077] The model switching unit arranges all the modules in the order of the approximation degrees and switches the service module of the robot to the first module.

[0078] The reaction output module is used to generate a robot reaction library based on the proximity of the module and the corresponding outputs of each module. Using each recognized electrical signal site as a key value, it matches the corresponding output in each health care module for the key value, calculates the output value of the matching result in each module according to the proximity of the module and the matching consistency, and activates the robot reaction in sequence according to the output value.

[0079] The reaction output module includes: a key value matching unit and a behavior activation unit;

[0080] The key value matching unit is used to identify subsequent electrical signal sites, use the subsequent electrical signal sites as key values, weighted accumulate the output actions of each module according to the mapping amount and weight, and match the corresponding output in the current service module.

[0081] The behavior activation unit is used to activate the output action of the robot according to the matching result, and end the service cycle after a fixed duration when the robot makes the output action.

[0082] The memory training module is used to perform convergence training on the robot reactions triggered by adjacent electrical signal sites to form a reaction memory library, and adjust the proximity weight of the health care module and the diffusion divergence of the signal recognition model according to the matching degree of the activated electrical signal sites in the reaction memory library.

[0083] The memory training module includes: a convergence classification unit, a memory library unit, and a memory weighting unit;

[0084] The convergence classification unit is used to perform convergence training on the electrical signal data and the robot output using an AI model;

[0085] The memory library unit is used to record the electrical signal sites of each sensor and the robot actions during the service cycle in the database;

[0086] The memory weighting unit is used to adjust the matching weights of each module using the memory library according to the matching degree of the activated electrical signal sites in the memory library during subsequent action cycles.

[0087] As Figure 2 shown, the application of the multi-module switching system based on the AI model in the health care robot includes the following steps:

[0088] Step S1. Determine the standard signal sites according to the hardware parameters of the attitude sensors in the health care robot, and record the actual electrical signal sites of the sensors each time the sensors transmit signals, and generate a signal set in the database;

[0089] Step S1 includes:

[0090] Step S11. Format each sensor of the health care robot, determine the standard signal sites according to the hardware parameters of the sensors, and the hardware parameters of the sensors include: sensor piezoelectric efficiency, temperature and humidity sensing efficiency, and human body monitoring data conversion efficiency, so that the standard signal site of each sensor corresponds to the specific body parameters, behavior parameters, and posture parameters of the health care object;

[0091] Step S12. When the sensor generates corresponding electrical signals according to external physical changes, record the level intensity values of each electrical signal site in the database, record them in the database, and store all the recorded electrical signal site data in the form of a set to form a signal set A, where A = {T1, T2,..., Tn}, n represents the number of records, and Tn represents the level intensity of the electrical signal site in the nth record.

[0092] Step S2. Calculate the aggregation degree of the signal set, generate random Gaussian noise with the same dispersion degree and aggregation degree, add the random Gaussian noise to the standard signal site to obtain a noise library, and use the noise library to train the AI model to obtain an AI model with data cleaning ability;

[0093] Step S2 includes:

[0094] Step S21. Calculate the aggregation degree of the signal set, and the aggregation degree represents the average of the squared differences between the elements in the set and the standard signal site, satisfying I = 1 / n · [(T1 - T0) 2 + (T2 - T0) 2 +... + (Tn - T0) 2 , where I is the aggregation degree and T0 is the level intensity of the standard signal site;

[0095] Step S22. Generate random Gaussian noise according to the aggregation degree, so that the generated noise satisfies the following conditions:

[0096]

[0097] where F(t) represents random Gaussian noise, t represents time, t0 represents the moment when the sensing signal occurs, and e is the base of the natural logarithm;

[0098] Step S23. Randomly change the value of t to obtain a set of Gaussian noise, add each element in the set of random Gaussian noise to the standard signal site to generate a noise library, and use the noise library to train the AI model so that when the AI model inputs the signal site, it automatically outputs the standard signal site with the closest level intensity, realizing the cleaning process of the signal site. The AI model includes: Mask R-CNN model, DeepLab model, U-Net model, and Watershed model.

[0099] Step S3. Collect all the sensing data of the robot during the service cycle, scatter the sensing data in the virtual space to obtain the actual signal model, calculate the mapping amount of the actual signal model in the space of each service module, and arrange the service modules in descending order according to the mapping amount.

[0100] Step S3 includes:

[0101] Step S31. Obtain the electrical signals of each sensor of the robot during the service cycle and clean the electrical signals. There is only a unique electrical signal for the same sensor during the service cycle. When multiple electrical signals are transmitted by the same sensor during the service cycle, end the current service cycle and allocate the redundant electrical signals to the next service cycle.

[0102] Step S32. Generate an m-dimensional virtual space in the control chip and label the collected electrical signals in vector form, where m represents the number of electrical signals during the service cycle, to obtain the actual signal model V, where V = (d1, d2, …, dm), and dm represents the level intensity of the electrical signal transmitted by the mth sensor.

[0103] Step S33. Expand the robot modules in interval form, represent the robot modules as {[D1+, D1-], [D2+, D2-], …, [Dn+, Dn-]}, where D1+, D2+ and Dn+ respectively represent the upper bounds of the intervals of the level intensities of the electrical signals transmitted by the sensors numbered 1, 2 and n, and D1-, D2- and Dn- respectively represent the lower bounds of the intervals of the level intensities of the electrical signals transmitted by the sensors numbered 1, 2 and n. Calculate the mapping vector of the actual signal model V in the n-dimensional space, and record the modulus length of the part of the mapping vector that falls within the interval where each robot module is located as the mapping amount of the actual signal model in the space of each module, and arrange the modules in descending order according to the mapping amount.

[0104] Step S4. Switch the working module of the robot to the service module ranked first, use the subsequent electrical signal positions as key values, weighted accumulate the output actions of each module according to the mapping amount and weight, obtain the value of each action, activate the action with the highest value output by the robot, and end the current cycle after a fixed duration when the robot makes the output action.

[0105] Step S4 includes:

[0106] Step S41. Switch the working module of the robot to the first module with the highest mapping amount, and after receiving the subsequent signals, calculate the value of each action according to the following formula:

[0107]

[0108] where, E represents the value of the action, k represents the number of service modules of the robot, u i represents the weight of the ith service module, pi represents the probability that the representative action is called in the i-th service module, d i represents the influence parameter of the subsequent signal on the actual signal model, which is determined by the rules of the AI model, r i represents the mapping amount of the actual signal model in the i-th service module;

[0109] Step S42. Select the action with the highest value as the output action for the current service cycle, activate the robot to execute the output action, and end the service cycle after a fixed duration when the robot makes the output action.

[0110] Step S5. Record the electrical signal sites of all sensors and the actual action outputs within the cycle to form a memory bank, and adjust the weights of each action according to the sensing information uploaded within the cycle in subsequent service cycles.

[0111] Step S5 includes:

[0112] Step S51. Record the electrical signal sites of each sensor and the robot actions within the service cycle, and use the AI model to perform convergence training on the electrical signal data and the robot output. When the convergence training result is positive feedback, store the electrical signal sites of the sensors and the robot actions as a memory pattern in the memory bank;

[0113] Step S52. In subsequent action cycles, calculate the matching degree of each electrical signal site in the actual signal model in the memory bank, and adjust the weights of the actions recorded in the memory unit in each service module according to the matching degree of the memory unit.

[0114] Embodiment: There are 3 sensors in the health care robot, which respectively monitor the body temperature, body posture and humidity of the object. There are also 3 service modules in the robot, namely nursing, medical treatment and alarm. Within one service cycle, the electrical signal sites of the 3 sensors are obtained, which are 2.1 mV, 0.8 mV and 1.4 mV respectively, and the mapping amounts in the 3 service modules are 0.21, 0.45 and 1.12 respectively. Then the robot is switched to the alarm state, and when receiving subsequent signals in subsequent service cycles, the robot actions are executed according to the logic of the alarm state.

[0115] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to this process, method, article or device.

[0116] Finally, it should be noted that the above are only preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or perform equivalent replacements for some of the technical features. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A multi-module switching system based on an AI model, characterized in that, The system includes the following modules: an attitude sensing module, a signal cleaning module, a model approximation module, a reaction output module, and a memory training module; The attitude sensing module is used to pre-transmit standard signal sites in the attitude sensors of the robot. Each type of electrical signal site corresponds to specific body parameters, behavior parameters, and attitude parameters of the health care object. It obtains historical classification results, generates a historical site set for each type of electrical signal, and calculates the aggregation degree of the set. The aggregation degree reflects the average intensity of environmental noise; The signal cleaning module is used to add random Gaussian noise with a dispersion degree equal to the aggregation degree to the standard electrical signal sites according to the set aggregation degree to obtain a random noise set. Using the random noise set as a noise library and the standard electrical signal as the output to train the AI model, and using the trained AI model to analyze the electrical signal sites of the attitude sensors and convert them into corresponding standard electrical signal outputs; The model approximation module is used to collect all the sensing data within one action cycle of the robot, simulate the multi-dimensional distribution of the electrical signal sites corresponding to all the sensing data in the virtual space, perform an identical operation with the multi-dimensional distribution models between the health care modules, obtain the mapping amounts of the multi-dimensional distribution in each health care module, and arrange all the modules in the order of approximation degree; The reaction output module is used to generate a robot reaction library according to the approximation degree of the module and the corresponding outputs of each module. Using each recognized electrical signal site as a key value, match the corresponding outputs of the key value in each health care module, weighted accumulate the output actions of each module according to the mapping amount and weight, obtain the value of each action, and activate the robot actions in sequence according to the value; The memory training module is used to perform convergence training on the robot reactions triggered by adjacent electrical signal sites to form a memory library, and adjust the weights of the actions in each module according to the matching degree of the activated electrical signal sites in the memory library.

2. The multi-module switching system based on the AI model according to claim 1, wherein: The attitude sensing module includes: a signal site unit and an environment aggregation unit; The signal site unit is used to format each sensor of the health care robot, determine the standard signal sites according to the sensor hardware parameters, and be uniformly managed by the main control chip; The environment aggregation unit is used to record the actual electrical signal sites of the sensors each time the sensors return signals, generate a set linked list in the database, and calculate the aggregation degree of the set; The signal cleaning module includes: a random noise unit and an impurity removal training unit; The random noise unit is used to generate random Gaussian noise with a dispersion degree equal to the aggregation degree according to the aggregation degree of the elements in the historical site set; The impurity removal training unit is used to add the random Gaussian noise into the standard signal sites to generate training noise, and use the training noise to train the AI classification model so that the AI classification model has the data cleaning ability.

3. The multi-module switching system based on the AI model according to claim 2, wherein: The model approximation module includes: a space distribution unit, a multi-simulation combination unit, and a model switching unit; The space distribution unit is used to scatter the electrical signal sites of all sensors in the virtual space to form an actual signal model; The multi-simulation combination unit is used to simulate the service intervals of each module of the robot in the virtual space, calculate the proportion of the part of the actual signal model located within the service interval, and output it as the approximation degree; The model switching unit arranges all modules in the order of proximity and switches the service module of the robot to the first module.

4. The multi-module switching system based on the AI model according to claim 3, wherein: The reaction output module includes: a key-value matching unit and a behavior activation unit; The key-value matching unit is used to identify the subsequent electrical signal sites, use the subsequent electrical signal sites as key values, weighted accumulate the output actions of each module according to the mapping amount and weight, and match the corresponding output in the current service module; The behavior activation unit is used to activate the output action of the robot according to the matching result, and end the service cycle after a fixed duration of the robot's output action.

5. The multi-module switching system based on the AI model according to claim 4, wherein: The memory training module includes: a convergence classification unit, a memory bank unit, and a memory weighting unit; The convergence classification unit is used to perform convergence training on the electrical signal data and the robot output using an AI model; The memory bank unit is used to record the electrical signal sites of each sensor and the robot actions during the service cycle in the database; The memory weighting unit is used to adjust the matching weights of each module using the memory bank according to the matching degree of the activated electrical signal sites in the memory bank during the subsequent action cycle.

6. The application of the multi-module switching system based on the AI model according to claim 1 in a health care robot, characterized in that, It includes the following steps: Step S1. Determine the standard signal sites according to the hardware parameters of the attitude sensors in the health care robot, and record the actual electrical signal sites of the sensors each time the sensors transmit signals, and generate a signal set in the database; Step S2. Calculate the aggregation degree of the signal set, generate random Gaussian noise with the same dispersion degree as the aggregation degree, add the random Gaussian noise to the standard signal sites to obtain a noise library, and use the noise library to train the AI model to obtain an AI model with data cleaning ability; Step S3. Collect all the sensing data of the robot during the service cycle, scatter each sensing data in the virtual space to obtain an actual signal model, calculate the mapping amount of the actual signal model in the space of each service module, and arrange each service module in descending order according to the mapping amount; Step S4. Switch the working module of the robot to the service module ranked first, use the subsequent electrical signal sites as key values, weighted accumulate the output actions of each module according to the mapping amount and weight to obtain the value of each action, activate the action with the highest output value of the robot, and end the current cycle after a fixed duration of the robot's output action; Step S5. Record the electrical signal sites of all sensors and the actual action outputs during the cycle to form a memory bank, and adjust the weights of each action according to the sensing information uploaded during the cycle in the subsequent service cycle.

7. The application of the multi-module switching system based on the AI model in the health care robot according to claim 6, wherein: Step S1 includes: Step S11. Format each sensor of the health care robot, and determine the standard signal sites according to the hardware parameters of the sensors. The hardware parameters of the sensors include: sensor piezoelectric efficiency, temperature and humidity sensing efficiency, and human body monitoring data conversion efficiency, so that the standard signal site of each sensor corresponds to the specific body parameters, behavior parameters, and attitude parameters of the health care object; Step S12. When the sensor generates corresponding electrical signals according to external physical changes, record the level intensity values of each electrical signal site in the database, store them in the database, and store all the recorded electrical signal site data in a set form to form a signal set A, where A = {T1, T2, …, Tn}, n represents the number of records, and Tn represents the level intensity of the electrical signal site in the nth record.

8. Application of the multi-module switching system based on the AI model in the healthcare robot according to claim 7, characterized in that: Step S2 includes: Step S21. Calculate the aggregation degree of the signal set. The aggregation degree represents the average of the squared differences between the elements in the set and the standard signal site, and satisfies I = 1 / n · [(T1 - T0) 2 + (T2 - T0) 2 + … + (Tn - T0) 2 , where I is the aggregation degree and T0 is the level intensity of the standard signal site; Step S22. Generate random Gaussian noise according to the degree of polymerization, so that the generated noise satisfies the following conditions: ； where F(t) represents random Gaussian noise, t represents time, t0 represents the moment when the sensing signal occurs, and e is the base of the natural logarithm; Step S23. Randomly change the value of t to obtain a set of Gaussian noise, add each element in the set of random Gaussian noise to the standard signal site to generate a noise library, and use the noise library to train the AI model, so that when the AI model inputs a signal site, it automatically outputs the standard signal site with the closest level intensity, realizing the cleaning process of the signal site. The AI model includes: Mask R-CNN model, DeepLab model, U-Net model, and Watershed model.

9. The application of the multi-module switching system based on the AI model in the health care robot according to claim 8, wherein: Step S3 includes: Step S31. Obtain the electrical signals of each sensor of the robot during the service cycle and clean the electrical signals. There is only a unique electrical signal for the same sensor during the service cycle. When the same sensor transmits multiple electrical signals during the service cycle, end the current service cycle and allocate the redundant electrical signals to the next service cycle; Step S32. Generate an m-dimensional virtual space in the control chip and label the collected electrical signals in vector form, where m represents the number of electrical signals during the service cycle, to obtain an actual signal model V, where V = (d1, d2, …, dm), and dm represents the level intensity of the electrical signal transmitted by the mth sensor; Step S33. Expand the robot module in the form of intervals, represent the robot module as {[D1+, D1-], [D2+, D2-], …, [Dn+, Dn-]}, where D1+, D2+, and Dn+ respectively represent the upper bounds of the intervals of the level intensities of the electrical signals transmitted by sensors numbered 1, 2, and n, and D1-, D2-, and Dn- respectively represent the lower bounds of the intervals of the level intensities of the electrical signals transmitted by sensors numbered 1, 2, and n. Calculate the mapping vector of the actual signal model V in the n-dimensional space, and record the modulus length of the part of the mapping vector that falls within the interval where each robot module is located as the mapping amount of the actual signal model in each module space, and arrange the modules in descending order of the mapping amount.

10. The application of the multi-module switching system based on the AI model in the health care robot according to claim 9, characterized in that: Step S4 includes: Step S41. Switch the working module of the robot to the first module with the highest mapping amount, and after receiving subsequent signals, calculate the value of each action according to the following formula: ； Among them, E represents the value of the action, k represents the number of robot service modules, and u i represents the weight of the i-th service module, and p i represents the probability that the action is called in the i-th service module, and d i represents the influence parameter of the subsequent signal on the actual signal model, which is determined by the rules of the AI model, and r i represents the mapping amount of the actual signal model in the i-th service module; Step S42. Select the action with the highest value as the output action of the current service cycle, activate the robot to execute the output action, and end the service cycle after a fixed duration when the robot makes the output action; Step S5 includes: Step S51. Record the electrical signal points and robot actions of each sensor during the service cycle, and use the AI model to conduct convergence training on the electrical signal data and robot outputs. When the convergence training result is positive feedback, store the electrical signal points and robot actions of the sensor as a memory pattern in the memory bank. Step S52. In subsequent action cycles, calculate the matching degree of each electrical signal point in the actual signal model in the memory bank, and adjust the weights of the actions recorded in the memory unit in each service module according to the matching degree of the memory units.

Images