Sensorless Control Method and Device for Pneumatic Soft Actuator, Pneumatic Soft Actuator and Electronic Device

Abstract

The present invention provides a sensorless control method, device, pneumatic soft actuator and electronic device for a pneumatic soft actuator, belonging to the field of soft robots. The method includes: obtaining the current injected air quality and pressure through a detection module on the air pipe arranged between the air pump module and the actuator body; inputting the injected air quality and pressure into a preset actuation estimation model, and using the actuation estimation model to obtain the current actuation estimation amount of the actuator body; obtaining an error value according to the actuation estimation amount and the target control amount; enabling the controller to adjust the gas injection amount of the air pump module based on the error value so as to adjust the actuation amount of the actuator body. In this way, there is no need to install a large-volume sensor on the actuator body, and the precise control of the actuator body can be achieved only based on the injected air quality and pressure, reducing the volume and gravity of the actuator body, and being not afraid of collision and humidity, thereby reducing the failure risk and improving flexibility, comfort and portability.

Description

Technical Field

[0001] The present invention relates to the field of soft robots, and more particularly, to a sensorless control method and device for a pneumatic soft actuator, a pneumatic soft actuator, and an electronic device. Background Art

[0002] Wearable robotic exoskeletons are revolutionizing human capabilities, with applications ranging from enhancing strength and endurance to assisting rehabilitation and improving daily activities. These devices are particularly effective in reducing physical fatigue, increasing load-bearing capacity, and helping physically disabled individuals regain mobility and live independently.

[0003] Due to advantages such as light weight, low cost, and simple driving, soft actuators are widely used in rehabilitation exoskeleton manipulators to assist the rehabilitation training and daily life of patients with hand function disorders. However, current control methods for soft actuators all require the installation of large-volume sensors such as angle sensors and torque sensors on the soft actuator body, resulting in poor flexibility, comfort, and portability. Moreover, the electronic components and circuits of the sensors require special anti-collision and moisture-proof treatments, leading to a high risk of failure. Summary of the Invention

[0004] In view of this, an object of the present invention is to provide a sensorless control method and device for a pneumatic soft actuator, a pneumatic soft actuator, and an electronic device, which can achieve control without installing large-volume sensors on the actuator body, greatly improving flexibility, comfort, and portability, and reducing the risk of failure.

[0005] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0006] In a first aspect, the present invention provides a sensorless control method for a pneumatic soft actuator, where the pneumatic soft actuator includes a controller, a gas pump module, a detection module, and an actuator body. The gas pump module is connected to the actuator body through a trachea, and the detection module is disposed on the trachea. The method includes:

[0007] Obtain the current measurement value of the detection module; wherein, the measurement value includes the injected air quality and pressure;

[0008] Input the injected air quality and the pressure into a preset actuation estimation model, and use the actuation estimation model to obtain the current actuation estimation amount of the actuator body; wherein, the actuation estimation model is a model obtained by training with the experimental data of the actuator body, and represents the relationship between the input amount and the actuation amount of the actuator body;

[0009] Obtain an error value based on the actuation estimation amount and the target control amount;

[0010] Input the error value into the controller, so that the controller adjusts the gas injection amount of the air pump module based on the error value to adjust the actuation amount of the actuator body.

[0011] Optionally, the method for obtaining the actuation estimation model includes:

[0012] Obtain multiple experimental data of the actuator body; wherein, the experimental data includes an input quantity and an output quantity, the output quantity includes an angle and / or a torque, and the input quantity includes the injected air mass and pressure;

[0013] Label each experimental data with the output quantity to obtain a sample set;

[0014] Use the sample set and Bayesian regularization to iteratively train a two-layer feedforward neural network to obtain an actuation estimation model.

[0015] Optionally, the sample set includes a first sample set labeled with an angle and a second sample set labeled with a torque;

[0016] The step of using the sample set and Bayesian regularization to iteratively train a two-layer feedforward neural network to obtain an actuation estimation model includes:

[0017] Use Bayesian regularization to iteratively train the two-layer feedforward neural network with the first sample set to obtain an angle estimation model;

[0018] Use Bayesian regularization to iteratively train the two-layer feedforward neural network with the second sample set to obtain a torque estimation model.

[0019] Optionally, the detection module includes a gas flow sensor and a pressure sensor;

[0020] The step of obtaining the current measurement value of the detection module includes:

[0021] Sample the gas flow sensor and the pressure sensor to obtain real-time gas flow rate and pressure;

[0022] Use the time period from the starting moment to the current moment as the target time period, and integrate the gas flow rates in the target time period in combination with the average molar mass of the gas output by the air pump to obtain the injected air mass at the current moment.

[0023] Optionally, the actuation estimation quantity includes a first angle quantity and a first torque quantity; when the target control quantity is an angle, the step of obtaining an error value according to the actuation estimation quantity and the target control quantity includes:

[0024] Input the first torque amount and the pressure into a preset angle prediction model, and use the angle prediction model to generate a second angle amount; wherein, the angle prediction model is a model obtained by training with experimental data of the actuator body;

[0025] Fuse the first angle amount and the second angle amount to obtain a current angle value;

[0026] Take the difference between the current angle value and the target control amount as the error value.

[0027] Optionally, the actuator estimation amount includes a first angle amount and a first torque amount; when the target control amount is torque, the step of obtaining the error value according to the actuator estimation amount and the target control amount includes:

[0028] Input the first angle amount and the pressure into a preset torque prediction model, and use the torque prediction model to generate a second torque amount; wherein, the torque prediction model is a model obtained by training with experimental data of the actuator body;

[0029] Fuse the first torque amount and the second torque amount to obtain a current torque value;

[0030] Take the difference between the current torque value and the target control amount as the error value.

[0031] Optionally, the actuator body includes a flexible integrated board and a plurality of air bags arranged on the flexible integrated board;

[0032] One end of the air bag is arranged on the flexible integrated board, and the distance between the arrangement positions of every two adjacent air bags is equal to a preset value;

[0033] Each air bag is provided with air holes, and the air bag is communicated with the trachea through the air holes.

[0034] In a second aspect, the present invention provides a sensorless control device for a pneumatic soft actuator. The pneumatic soft actuator includes a controller, an air pump module, a detection module, and an actuator body. The air pump is communicated with the actuator body through a trachea, and the detection module is arranged on the trachea. The device includes a detection module, an estimation module, an error acquisition module, and a control adjustment module;

[0035] The detection module is configured to obtain the current measurement value of the detection module; wherein, the measurement value includes the injected air quality and pressure;

[0036] The estimation module is configured to input the injected air quality and the pressure into a preset actuation estimation model, and obtain the current actuation estimation quantity of the actuator body by using the actuation estimation model; wherein, the actuation estimation model is a model obtained by training with the experimental data of the actuator body, and represents the relationship between the input quantity and the actuation quantity of the actuator body;

[0037] The error acquisition module is configured to obtain an error value according to the actuation estimation quantity and the target control quantity;

[0038] The control adjustment module is configured to input the error value into the controller, so that the controller adjusts the gas injection quantity of the air pump module based on the error value to adjust the actuation quantity of the actuator body.

[0039] In a third aspect, the present invention provides a pneumatic soft actuator, including a controller, an air pump module, a detection module and an actuator body. The actuator body includes a flexible integrated board and a plurality of air bags arranged on the flexible integrated board;

[0040] One end of the air bag is arranged on the flexible integrated board, and the distance between the arrangement positions of every two adjacent air bags is equal to a preset value. Each air bag is provided with air holes;

[0041] The air pump module is communicated with the air holes of the air bag through an air pipe, and the detection module is arranged on the air pipe;

[0042] The detection module is configured to continuously detect the gas injected by the air pump module into the actuator body to obtain a real-time measurement value;

[0043] The controller is configured to implement the sensorless control method of the pneumatic soft actuator as described in the first aspect.

[0044] In a fourth aspect, the present invention provides an electronic device, including a processor and a memory. The memory stores a computer program that can be executed by the processor, and the processor can execute the computer program to implement the sensorless control method of the pneumatic soft actuator as described in the first aspect.

[0045] In a fifth aspect, the present invention provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the sensorless control method of the pneumatic soft actuator as described in the first aspect.

[0046] The sensorless control method, device, pneumatic soft actuator and electronic device provided by the embodiments of the present invention. The method includes: obtaining the current injected air quality and pressure through a detection module disposed on the air pipe between the air pump module and the actuator body; inputting the injected air quality and pressure into a preset actuation estimation model, and using the actuation estimation model to obtain the current actuation estimation amount of the actuator body, where the actuation estimation model is a model obtained by training with the experimental data of the actuator body, and represents the relationship between the input amount and the actuation amount of the actuator body; obtaining an error value according to the actuation estimation amount and the target control amount; inputting the error value into a controller, so that the controller adjusts the gas injection amount of the air pump module based on the error value to adjust the actuation amount of the actuator body. In this way, there is no need to install a large-volume sensor on the actuator body, and the precise control of the actuator body can be realized only based on the injected air quality and pressure, greatly reducing the volume and gravity of the actuator body, thereby improving the flexibility, comfort and portability of the actuator body. At the same time, the limitation of the sensor weight and size on the force output of the soft actuator is reduced or even eliminated, and the risk of mechanical and electronic failures is reduced.

[0047] To make the above objects, features and advantages of the present invention more obvious and understandable, the following specifically enumerates preferred embodiments and cooperates with the attached drawings to make a detailed description as follows. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required in the embodiments. It should be understood that the following drawings only show some embodiments of the present invention, and therefore should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts.

[0049] Figure 1 Shows the structural schematic diagram of the pneumatic soft actuator provided by the embodiments of the present invention.

[0050] Figure 2 Shows the structural schematic diagram of the actuator body provided by the embodiments of the present invention.

[0051] Figure 3 Shows the structural schematic diagram of the electronic device provided by the embodiments of the present invention.

[0052] Figure 4 Shows one of the flow schematic diagrams of the sensorless control method provided by the embodiments of the present invention.

[0053] Figure 5 Shows Figure 4 The flow schematic diagram of some sub-steps of step 11 in

[0054] Figure 6 Figure 2 shows the second schematic flowchart of the sensorless control method provided by the embodiment of the present invention.

[0055] Figure 7 Figure 3 shows the schematic structural diagram of the measurement experiment system provided by the embodiment of the present invention.

[0056] Figure 8 Figure 4 shows the first schematic diagram of the control system architecture provided by the embodiment of the present invention.

[0057] Figure 9 Figure 5 shows the second schematic diagram of the control system architecture provided by the embodiment of the present invention.

[0058] Figure 10 Figure 6 shows Figure 4 the first schematic flowchart of some sub-steps of step 15 in

[0059] Figure 11 Figure 7 shows Figure 4 the second schematic flowchart of some sub-steps of step 15 in

[0060] Figure 12 Figure 8 shows the dynamic simulation model of a pneumatic soft actuator.

[0061] Figure 13 Figure 9 shows the simulation result diagram when the actuator body is set at the ankle joint.

[0062] Figure 14 Figure 10 shows the simulation result diagram when the actuator body is set at the knee joint.

[0063] Figure 15 Figure 11 shows the schematic block diagram of the sensorless control device provided by the embodiment of the present invention.

[0064] Description of reference numerals: 10 - pneumatic soft actuator; 110 - controller; 120 - air pump; 130 - solenoid valve; 140 - actuator body; 141 - flexible integrated board; 142 - airbag; 143 - air duct joint; 150 - air flow sensor; 160 - pressure sensor; 20 - electronic device; 210 - memory; 220 - processor; 230 - communication module; 30 - measurement experiment system; 310 - support platform; 320 - lever; 330 - moving column; 340 - tension pulley; 350 - soft rope; 360 - first angle gauge; 370 - second angle gauge; 380 - tensiometer; 390 - barometer; 391 - thermometer; 40 - sensorless control device; 410 - detection module; 420 - estimation module; 430 - error acquisition module; 440 - control adjustment module. Detailed implementation manners

[0065] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying 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 the embodiments. Components of the embodiments of the present invention described and illustrated in the accompanying drawings here can be arranged and designed in various different configurations.

[0066] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed present invention, but merely represents selected embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0067] It should be noted that 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 terms "comprising", "including" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "including one..." does not exclude the existence of additional identical elements in the process, method, article or device including the said element.

[0068] Currently, actuators in wearable robotic exoskeletons usually have complex structures. For example, for easy control, sensors for measuring torque, angle, etc. are provided on the actuator body, and the weight and size of the actuator body increase accordingly. This not only results in poor flexibility, comfort, and portability of the actuator body, but also limits the force output and brings risks of mechanical and electronic failures.

[0069] In view of the above problems, the embodiments of the present invention provide a sensorless control method, device, pneumatic soft actuator, and electronic device for a pneumatic soft actuator, which can achieve control without installing large-volume sensors on the actuator body, greatly reducing the mass and size of the actuator body, thereby improving the above problems.

[0070] Please refer to Figure 1 and Figure 2 , the embodiments of the present invention provide a pneumatic soft actuator 10, including a controller 110, an air pump module, a detection module 410, and an actuator body 140. The actuator body 140 includes a flexible integrated board 141 and a plurality of airbags 142 provided on the flexible integrated board 141.

[0071] Refer toFigure 2 One end of the airbag 142 is disposed on the flexible integrated board 141, and the distance between the setting positions of every two adjacent airbags 142 is equal to a preset value. Each airbag 142 is provided with an air hole. That is, a plurality of airbags 142 are arranged at equal intervals on the flexible integrated board 141, and the sizes and shapes of these airbags 142 are the same.

[0072] The air pump module is communicated with the air holes of the airbag 142 through a trachea, and the detection module 410 is disposed on the trachea.

[0073] The actuator body 140 is used to be disposed on any part of the application object. For example, it can be disposed on the knee joint of the human body or on the ankle joint of the human body.

[0074] The detection module 410 is used to continuously detect the gas injected into the actuator body 140 by the air pump module to obtain a real-time measurement value.

[0075] The controller 110 is used to implement the sensorless control method of the pneumatic soft actuator 10 provided by the embodiment of the present invention.

[0076] Refer to Figure 2 In order to facilitate the inflation of the airbag 142 and avoid air leakage at the same time, an air pipe joint 143 can be disposed on the air hole of the airbag 142, and the air pipe joint 143 is fixed on the airbag 142 through TPU glue. In this way, the air pipe joint 143 can be stably fixed, and air leakage between the air hole of the airbag 142 and the air pipe joint 143 can be avoided.

[0077] In addition, refer to Figure 1 The air pump module includes an air pump 120 and a solenoid valve 130. The air pump 120 is communicated with the actuator body 140 (i.e., the air pipe joints 143 of each airbag 142) through a trachea, and the solenoid valve 130 is disposed on the trachea.

[0078] Refer to Figure 1 The detection module 410 includes an air flow sensor 150 (which can also be a volume flow sensor) and a pressure sensor 160. The air flow sensor 150 and the pressure sensor 160 are disposed on the trachea between the solenoid valve 130 and the actuator body 140. Thus, the air flow sensor 150 can measure the gas flow rate injected into the actuator body 140 in real time. Since the trachea and the actuator body 140 are in a communicating state, the pressure sensor 160 can measure the gas pressure in the actuator body 140.

[0079] Furthermore, it can further include a computer device (which can be a server or a server cluster) communicatively connected to the controller 110 of the pneumatic soft actuator 10, which is used to train an actuation estimation model based on the experimental data of the pneumatic soft actuator 10 and deploy the actuation estimation model to the controller 110.

[0080] It should be noted that the above-mentioned controller 110 can be, but is not limited to, a single-chip microcomputer, a programmable logic controller 110, a field programmable gate array, a personal computer, etc. The material of the airbag 142 of the actuator body 140 can be TPU-coated nylon or any other material with lightweight, high strength and tear resistance characteristics, and the flexible integrated board 141 can be a cloth of any material.

[0081] Please refer to Figure 3 , which is a block diagram of the electronic device 20. The electronic device 20 can be Figure 1 the controller 110 in the pneumatic soft actuator 10 shown. The electronic device 20 includes a memory 210, a processor 220 and a communication module 230. Each element of the memory 210, the processor 220 and the communication module 230 is electrically connected directly or indirectly to each other to achieve data transmission or interaction. For example, these elements can be electrically connected to each other through one or more communication buses or signal lines.

[0082] Among them, the memory 210 is used to store programs or data. The memory 210 can be, but is not limited to, a random access memory, a read-only memory, a programmable read-only memory, an erasable read-only memory, an electrically erasable read-only memory, etc.

[0083] The processor 220 is used to read / write the data or programs stored in the memory 210 and perform corresponding functions. For example, Figure 1 in the pneumatic soft actuator 10 shown, the processor 220 of the controller 110 reads the computer program stored in the memory 210 to execute the sensorless control method of the pneumatic soft actuator 10 provided by the embodiments of the present invention.

[0084] The communication module 230 is used to establish a communication connection between the electronic device 20 and other communication terminals by means of wired or wireless, etc., and is used to receive and transmit data through the network. For example, Figure 1 in the pneumatic soft actuator 10 shown, the controller 110 communicates with the air pump 120, the air flow sensor 150, the pressure sensor 160 and the solenoid valve 130 through the communication module 230 to receive and transmit data.

[0085] It should be understood that Figure 3 the structure shown is only a schematic diagram of the structure of the electronic device 20, and the electronic device 20 may further include more or fewer components than Figure 3 shown, or have a different configuration from Figure 3 shown. Figure 3 Each component shown in

[0086] Please refer to Figure 4, an embodiment of the present invention provides a sensorless control method for a pneumatic soft actuator 10, including steps 11 to 17. Thus, it is made that: Figure 1 In the pneumatic soft actuator 10 shown, the controller 110 Figure 2 In the structure shown, when the processor 220 executes the computer program stored in the memory 210, steps 11 to 17 are implemented.

[0087] Step 11, obtain the current measurement value of the detection module.

[0088] Among them, the measurement value includes the injected air quality and pressure.

[0089] Step 13, input the injected air quality and pressure into a preset actuation estimation model, and use the actuation estimation model to obtain the current actuation estimation quantity of the actuator body.

[0090] Among them, the actuation estimation model is a model obtained by training with the experimental data of the actuator body, which characterizes the relationship between the input quantity and the actuation quantity of the actuator body.

[0091] Step 15, obtain an error value according to the actuation estimation quantity and the target control quantity.

[0092] Step 17, input the error value into the controller, so that the controller adjusts the gas injection quantity of the air pump module based on the error value to adjust the actuation quantity of the actuator body.

[0093] Here, the target control quantity can be the target torque or the target angle.

[0094] Exemplarily, in combination with Figure 1 the pneumatic soft actuator 10 shown, after the actuator body 140 is worn on parts such as the knee joint or ankle joint of the human body and enters the working mode, the air flow sensor 150 measures the current gas flow rate in real time, and the pressure sensor 160 measures the current gas pressure in real time. The controller 110 samples the air flow sensor 150 and the pressure sensor 160 to obtain the real-time injected air quality and pressure.

[0095] The controller 110 inputs the current injected air quality and pressure into a preset actuation estimation model, uses the actuation estimation model to obtain the current actuation estimation quantity of the actuator body 140, and obtains an error value according to the actuation estimation quantity and the target control quantity. Thus, the controller 110 adopts the PID control principle and outputs the control voltage corresponding to the current error value to the solenoid valve 130 based on the current error value, adjusts the opening degree of the solenoid valve 130 in the air pump module to adjust the gas flow rate and / or pressure (i.e., the gas injection quantity) injected into the actuator body 140, and further adjusts the actuation quantity (i.e., the angle or the output torque) of the actuator body 140, so that the actuation quantity of the actuator body 140 is closer to the target actuation quantity.

[0096] Compared with the existing control method of the pneumatic soft actuator 10, in the above sensorless control method, the precise control of the actuator body 140 can be achieved only based on the measured values (i.e., the injected air quality and pressure) of the detection module 410 arranged on the air pipe, without installing sensors for measuring torque, angle, etc. on the actuator body 140, greatly reducing the volume and gravity of the actuator body 140, thereby improving the flexibility, comfort and portability of the actuator body 140. At the same time, the limitation of the sensor weight and size on the force output of the soft actuator is reduced or even eliminated, and the risk of mechanical and electronic failures is reduced.

[0097] Among them, when the detection module 410 includes an air flow sensor 150 and a pressure sensor 160, the air flow sensor 150 obtains the current gas flow rate injected into the actuator body 140 in real time. At this time, the method of obtaining the measured values (i.e., the injected air quality and pressure) can be flexibly selected. For example, the injected air quality can be estimated by an estimation model, or the injected air quality can be calculated by a pre-designed formula, and its implementation method is not limited.

[0098] In order to obtain more accurate measured values, the idea of integrating the gas flow rate is introduced in the process of obtaining the measured values in step 11 to obtain the injected air quality of the actuator body 140 at the current time. Refer to Figure 5 , in step 11, the measured values are obtained through the following steps 111 to 113.

[0099] Step 111, sample the gas flow sensor and the pressure sensor to obtain the real-time gas flow rate and pressure.

[0100] Step 113, take the start time to the current time as the target period, and combine the average molar mass of the gas output by the air pump to integrate the gas flow rates in the target period to obtain the injected air quality at the current time.

[0101] Here, the start time refers to the time when the air pump module starts to work, and the injected air quality at the current time refers to the total mass of air injected into the actuator body when reaching the current time.

[0102] Specifically, according to the atmospheric pressure and the pressure measured by the pressure sensor, the average air pressure in each sampling period is obtained, and then the gas flow rates in the target period are integrated in combination with the average air pressure and the average molar mass of the gas output by the air pump to obtain the injected air quality at the current time.

[0103] The process of obtaining the above injected air quality can be expressed by the formula: Wherein, M represents the injected air quality, σ represents the gas flow rate at the nth sampling moment, μ represents the average molar mass of air, R represents the molar gas constant, T represents the room temperature, P̅ represents the average air pressure during the sampling period, n represents that there are n sampling moments before the current moment, and Δt represents the sampling period (i.e., the time interval between two adjacent sampling moments). P d represents the atmospheric pressure, and P n represents the pressure measured by the pressure sensor 160 at the nth sampling moment.

[0104] It should be noted that the average molar mass of air, the room temperature, and the atmospheric pressure are all known set values.

[0105] Through the above steps 111 to 113, the air quality injected into the actuator body 140 is obtained by means of integration, and the influence of the pressure measured by the pressure sensor 160, the room temperature, the atmospheric pressure, and the gas flow rate during each sampling period is considered during the integration process, making the calculated injected air quality more accurate.

[0106] After calculating the injected air quality in the above manner, step 13 is executed, and the input air quality and pressure are input into the preset actuator estimation model, and the current actuator estimation quantity of the actuator body 140 can be obtained.

[0107] Wherein, the actuator estimation quantity can be a first angular quantity and / or a first torque quantity.

[0108] The actuator estimation model characterizes the relationship between the input quantities (i.e., the injected air quality and pressure) of the actuator body 140 and the output quantities (i.e., the angle and / or torque) of the actuator. The actuator estimation model can be a model obtained by regression of experimental data, or a model obtained by training a neural network model using experimental data, and its acquisition method is not limited.

[0109] In order to make the error of the actuator estimation model smaller and the accuracy higher, and at the same time improve the training speed and calculation speed of the actuator estimation model, the double-layer feedforward neural network is iteratively trained using experimental data to obtain the actuator estimation model. Refer to Figure 6 , which includes steps 21 to 25.

[0110] Step 21, obtain multiple experimental data of the actuator body.

[0111] Wherein, the experimental data includes input quantities and output quantities, the output quantities include angles and / or torques, and the input quantities include injected air quality and pressure.

[0112] Here, a simulation experiment can be carried out on the actuator body 140 in the pneumatic soft actuator 10 shown in Figure 1 to obtain experimental data.

[0113] For example, a measurement experiment system 30 as shown in Figure 7 can be constructed. The measurement experiment system 30 includes a support platform 310, a lever 320, a moving column 330, a tension pulley 340, and a flexible rope 350. One end of the lever 320 is hinged to the support platform 310, and the actuator body 140 is disposed between the support platform 310 and the lever 320. A slide rail is provided on the moving column 330, the tension pulley 340 is disposed on the moving column 330, and the tension pulley 340 is slidably engaged with the slide rail. One end of the flexible rope 350 is wound around the tension pulley 340, and the end of the flexible rope 350 away from the tension pulley 340 is disposed on the lever 320. Thus, by adjusting the position of the moving column 330 and moving and fixing the tension pulley 340 at any position on the slide rail, the lever 320 can be fixed at any angle, and thus the actuator body 140 can be restricted at any angle.

[0114] The measurement experiment system 30 further includes a first angle meter 360, a second angle meter 370, a tensiometer 380, an air flow sensor 150, a pressure sensor 160, a barometer 390, and a thermometer 391. The first angle meter 360 and the tensiometer 380 are disposed on the flexible rope 350, and the second angle meter 370 is disposed on the lever 320. The air flow sensor 150 and the pressure sensor 160 are disposed on the air pipe and are located between the actuator body 140 and the solenoid valve 130. The controller 110 is communicatively connected to the air pump 120, the solenoid valve 130, the air flow sensor 150, the pressure sensor 160, the barometer 390, and the thermometer 391 respectively.

[0115] The torque of the actuator body 140 is measured and recorded by the tensiometer 380, the angle of the actuator body is measured and recorded by the second angle meter 370, the measurement angle of the first angle meter 360 is used to determine whether the pulling direction of the flexible rope 350 is perpendicular to the lever 320, the current atmospheric pressure is measured by the barometer 390, the current room temperature is measured by the thermometer 391, the gas flow rate injected into the actuator body 140 is measured by the air flow sensor 150, and the gas pressure of the pneumatic module 1 is measured by the pressure sensor 160.

[0116] Based on the above structure, experiments on the actuator body 140 can be quickly carried out to obtain multiple groups of experimental data. Each group of experimental data includes at least the angle, torque, injected air quality, and pressure of the actuator body 140.

[0117] Step 23: Label each piece of experimental data with the output quantity to obtain a sample set.

[0118] Step 25: Use the sample set and Bayesian regularization to iteratively train a two-layer feedforward neural network to obtain an actuator estimation model.

[0119] Among them, the double-layer feedforward neural network includes a hidden layer with 7 neurons with sigmoid transfer functions and an output layer with 1 linear transfer function connected in sequence.

[0120] When the training objective is: actuating the estimation model to take the injected air quality and pressure as inputs and output torque and angle, then the label of each sample in the sample set is torque + angle. When the training objective is: actuating the estimation model to take the injected air quality and pressure as inputs and output torque, then the label of each sample in the sample set is torque. When the training objective is: actuating the estimation model to take the injected air quality and pressure as inputs and output angle, then the label of each sample in the sample set is angle.

[0121] Furthermore, in order to make the estimation of the actuating estimation model more accurate, in step 25, an angle estimation model trained separately with the injected air quality and pressure as inputs and angle as output, and a torque estimation model with the injected air quality and pressure as inputs and torque as output are introduced.

[0122] At this time, the sample set obtained in step 23 includes a first sample set with angle as the label and a second sample set with torque as the label. The process of training the torque estimation model in step 25 is: using Bayesian regularization, iteratively training the double-layer feedforward neural network with the first sample set to obtain the angle estimation model; using Bayesian regularization, iteratively training the double-layer feedforward neural network with the second sample set to obtain the torque estimation model.

[0123] For example, when training the angle estimation model, first define a prior distribution for the model parameters of the double-layer feedforward neural network. This prior distribution can be a Gaussian distribution (corresponding to L2 regularization) or a Laplace distribution (corresponding to L1 regularization). When choosing the Gaussian distribution as the prior distribution, its probability density function can be expressed as: α represents the regularization parameter, E w represents the sum of the squares of the model weights, Z w (α) represents the normalization factor, x i represents the i-th parameter (such as weight or bias) of the double-layer feedforward neural network.

[0124] According to the first sample set, use Bayes' theorem to update the prior distribution to obtain the posterior distribution. The posterior distribution can be expressed as: P(D|x,β,N) represents the likelihood function (i.e., the probability that the model observes data D given parameters β and sample x), P(x|α,N) represents the prior distribution, P(D|α,β,N) represents the normalization factor, N represents a two-layer feedforward neural network (including the number of layers of the network and the number of neurons in each layer), and α, β represent hyperparameters (i.e., regularization parameters).

[0125] Furthermore, randomly select batch samples from the first sample set, input the injected air quality and pressure in the batch samples into the two-layer feedforward neural network, calculate the loss value (i.e., the error value) based on the output of the two-layer feedforward neural network and the labels of the batch samples, and determine the data error term in the batch samples. Then, maximize the posterior probability to estimate the model parameters, that is, minimize the cost function: F(x) = βE D +E w α, E D represents the data error term, and adjust the hyperparameters according to this cost function. After determining the hyperparameters, use gradient descent or other optimization algorithms to minimize the cost function F(x) to complete one iteration of training. Stop until the preset number of training times is reached or the convergence condition is met to obtain the angle estimation model.

[0126] The process of training to obtain the torque estimation model is the same as the above process of training to obtain the angle estimation model, and will not be elaborated here.

[0127] The angle estimation model characterizes the relationship between the input quantities (i.e., the injected air quality and pressure) of the actuator body and the angle of the actuator, and the torque estimation model characterizes the relationship between the input quantities (i.e., the injected air quality and pressure) of the actuator body and the torque output by the actuator.

[0128] The angle estimation model can be expressed as: θ = f θ (M,P), and the torque estimation model can be expressed as: τ = f τ (M,P), where θ represents the angle, τ represents the torque, P represents the pressure, and M represents the injected air quality.

[0129] During the training process of the model, the mean squared error loss function, cross-entropy loss function, or any loss function can be used to calculate the difference between the label (i.e., the true value) of the sample and the estimated value of the model. It is also possible to use a combination of several loss functions to calculate the difference between the label (i.e., the true value) of the sample and the estimated value of the model.

[0130] Through the above steps 21 to 25, prior knowledge is introduced in model training through Bayesian regularization to prevent overfitting, improve the generalization ability of the model, and improve the accuracy of the finally obtained angle estimation model and torque estimation model. At the same time, a simple two-layer feedforward neural network is used as the basic model, which improves the convergence speed of model training and improves the calculation speed of the model by reducing the calculation amount of the model, so as to facilitate quickly obtaining the first angle quantity or the first torque quantity using the model.

[0131] After obtaining the actuation estimation model (including the angle estimation model and the torque estimation model) in the above manner, in step 13, an appropriate model can be selected according to the target control quantity to generate the actuation estimation quantity.

[0132] For example, when the target control quantity is an angle, the angle estimation model is selected to generate the current first angle quantity of the actuator body 140, and then, in steps 15 and 17, the gas injection amount of the air pump module (i.e., the opening degree of the solenoid valve 130) is adjusted according to the error value between the target control quantity and the first angle quantity. At this time, the control system architecture diagram can be as Figure 8 shown.

[0133] When the target control quantity is a torque, the torque estimation model is selected to generate the current first angle quantity of the actuator body 140, and then, in steps 15 and 17, the gas injection amount of the air pump module (i.e., the opening degree of the solenoid valve 130) is adjusted according to the error value between the target control quantity and the first torque quantity. At this time, the control system architecture diagram can be as Figure 9 shown.

[0134] In order to improve the accuracy of the angle value and thus improve the control precision, the idea of introducing and obtaining the estimation quantities of the angle and torque at the current moment (i.e., the first angle quantity and the first torque quantity), converting one of the estimation quantities and then fusing it with the other estimation quantity, and adjusting the opening degree of the solenoid valve 130 according to the error between the fused value and the target control quantity is introduced.

[0135] Among them, the method of converting one of the estimation quantities and then fusing it with the other estimation quantity can be set flexibly. For example, it can be to convert one of the estimation quantities using a pre-trained conversion model and then fuse it, or to convert one of the estimation quantities according to a preset rule and then fuse it. And the above methods are only examples, and their implementation methods are not limited.

[0136] When the target control quantity is an angle, in order to ensure that there is no sensor on the actuator body and improve the control precision at the same time, in step 15, the idea of introducing an angle prediction model that uses torque and pressure as inputs and predicts the angle, converting the first torque quantity into an angle, and then fusing it with the first angle quantity to calculate the error is introduced. Refer to Figure 10, the process of obtaining the error value in step 15 includes steps 151A to 155A.

[0137] Step 151A, input the first torque amount and the pressure into a preset angle prediction model, and use the angle prediction model to generate a second angle amount.

[0138] Among them, the angle prediction model is a model obtained by training the model with the experimental data of the actuator body. The method of obtaining the angle prediction model is the same as the method of obtaining the angle estimation model in the above text. That is, using the training samples obtained after annotating the experimental data (at this time, the training samples use the angle in a set of experimental data as the label, and the pressure and torque in this set of experimental data as the input), iteratively training the double-layer feedforward neural network to obtain the angle prediction model, and this angle prediction model can be expressed as θ = f θ (τ, P).

[0139] Step 153A, fuse the first angle amount and the second angle amount to obtain the current angle value.

[0140] Here, the fusion method can be set flexibly. For example, the average value of the second angle amount and the first angle amount can be used as the current angle value, or the second angle amount and the first angle amount can be weighted and fused to obtain the current angle value. And the above fusion methods are only examples, and the implementation methods are not limited.

[0141] Step 155A, use the difference between the current angle value and the target control amount as the error value.

[0142] Similarly to the above steps 151A to 155A, when the target control amount is a torque, in order to ensure that there is no sensor on the actuator body and improve the control accuracy at the same time, in step 15, a torque prediction model that predicts the angle with the angle and pressure as the input and is pre-trained is introduced, and the first angle amount is converted into a torque, and then fused with the first torque amount to calculate the error idea. Refer to Figure 11 , the process of obtaining the error value in step 15 includes steps 151B to 155B.

[0143] Step 151B, input the first angle amount and the pressure into a preset torque prediction model, and use the torque prediction model to generate a second torque amount.

[0144] Among them, the torque prediction model is a model obtained by training the model with the experimental data of the actuator body. The method for obtaining the torque prediction model is the same as the method for obtaining the torque estimation model in the above text. That is, using the training samples obtained after annotating the experimental data (at this time, the training samples use the torque in a set of experimental data as the label, and the pressure and angle in this set of experimental data as the input variables), iteratively training the double-layer feedforward neural network to obtain the torque prediction model, which can be expressed as τ = f τ (θ, P).

[0145] Step 153B: Fuse the first torque quantity and the second torque quantity to obtain the current torque value.

[0146] Here, the fusion method can be set flexibly. For example, the average value of the first torque quantity and the second torque quantity can be used as the current torque value, or the first torque quantity and the second torque quantity can be weighted and fused to obtain the current torque value. And the above fusion methods are only examples, and the implementation methods are not limited.

[0147] Step 155B: Use the difference between the current torque value and the target control quantity as the error value.

[0148] Through the above steps 151A to 155A and steps 151B to 155B, after converting one of the estimated quantities obtained by the actuation estimation model into a value of the same type as the other estimated quantity and then fusing them to calculate the error, the error caused by the performance of a single model itself can be eliminated, thereby improving the overall control accuracy of the pneumatic soft actuator by improving the accuracy of the error. At the same time, there is no need to add additional sensors to the actuator body, ensuring the high-precision control of the pneumatic soft actuator while ensuring the small volume, flexibility, comfort, lightness and low failure rate of the actuator body.

[0149] To improve the accuracy, in step 15, it is also possible to introduce the position, center of gravity line and negative weight of the application object of the actuator body 140, convert the estimated torque value into an angular quantity, and then fuse it with the first angular quantity, and obtain the error value with the target control quantity using the fused angle.

[0150] For example, first, according to the center of gravity line of the application object and the installation position of the actuator body on the application object, obtain the distance between the center of gravity line and the installation position.

[0151] Here, the center of gravity line refers to the line perpendicular to the ground passing through the center of gravity point. The application object can be the human body, and the installation position of the actuator body 140 can be the knee joint, ankle joint, etc. of the human body. Stress sensors can be set on the application object (such as the human body), and inertial sensors and pressure sensors can be set at the installation position, so as to measure the real-time center of gravity point based on the stress sensors, and then update the real-time center of gravity line. Based on the measurement values of the inertial sensors, the installation position (coordinates) of the actuator body 140 can be obtained in real time, and the real-time negative weight can be obtained based on the pressure sensors. Furthermore, by using the projection method, the distance (vertical distance) between the center of gravity line and the installation position can be calculated.

[0152] Next, based on the negative weight and the distance, the moment of inertia and the load amount are obtained.

[0153] Here, multiplying the gravitational acceleration, the negative weight, and the distance can obtain the load amount. Multiplying the negative weight and the square of the distance gives the moment of inertia. The acquisition of the load amount can be expressed by the formula: Load = 9.8WL, and the acquisition of the moment of inertia can be expressed by the formula: J = WL 2 , where L represents the distance between the center of gravity line and the installation position, and W represents the negative weight.

[0154] Continuing, based on the torque-angle relationship of the pneumatic soft actuator, as well as the first torque amount, the moment of inertia, and the load amount, the conversion angle amount corresponding to the first torque amount is obtained.

[0155] Among them, the acquisition of the conversion angle amount can be expressed by the formula: τ e represents the first torque amount, θ c represents the conversion angle amount, c and k are both constant parameters, and s represents the variable in the frequency domain of the transformation. The above conversion formula is obtained by performing the Laplace transform on the dynamic Euler-Lagrange equation of the pneumatic soft actuator 10.

[0156] Thus, by fusing the conversion angle amount and the first angle amount, the current angle value is obtained.

[0157] The average value of the conversion angle amount and the first angle amount can be used as the current angle value, or the angle amount and the first angle amount can be weighted and fused to obtain the current angle value. And the above ways of obtaining the current angle value are only examples, and the implementation methods are not limited.

[0158] Furthermore, the difference between the current angle value and the target control amount is used as the error value.

[0159] In the above manner, according to the relevant influencing quantities of the application object and the installation position of the actuator body 140, the first torque quantity is converted into an angle, and then the converted angle quantity and the estimated angle quantity are fused to further reduce the angle error. Furthermore, based on the currently obtained angle quantity and the target control quantity (i.e., the target angle) after fusion, the error value of the current control is calculated, making the error value more accurate and helping to improve the control precision.

[0160] To verify the accuracy of the above sensorless control method, the actuator body 140 of the pneumatic soft actuator 10 is respectively installed at the knee joint and ankle joint of the human body for simulation verification.

[0161] When the pneumatic soft actuator 10 is applied to the ankle joint of the human body, the dynamic simulation model of the pneumatic soft actuator 10 is as Figure 12 shown. The angle is estimated using the angle estimation model, and the torque is estimated using the torque estimation model. The PID controller simulates the controller 110 to control the injected air quality (i.e., output the control quantity for controlling the opening of the solenoid valve 130 based on the PID principle), and the load module is used to simulate the load at the ankle joint. The air quality output by the PID controller and the output pressure of the actuator simulation model are fed into the angle estimation model to estimate the angle. Then, this estimated angle is fed back to the PID controller to adjust the air quality injected into the actuator, thereby controlling the bending angle of the actuator. The load module is used to simulate the load (i.e., load torque) during the bending process of the actuator. This module takes the angle as the input, outputs 0 load when the angle is less than 0°, and outputs a load of 21.2 N·m when the angle exceeds 60°. Two oscilloscope modules are used to display the angle and torque during the entire simulation process.

[0162] The simulation parameters of the control system are as follows: load = 21.2 N·m; moment of inertia = 0.1037 kg·m2; damping coefficient c = 1, stiffness coefficient k = 0.05. The bending angle range is 60 - 86°, the time delay of air diffusion in the actuator simulation model is 0.003 s, and the simulation time is 0 - 20 seconds. The simulation results are as Figure 13 shown.

[0163] At this time, the simulation is divided into four stages. In the first stage (0 - 3.7 seconds), the actuator is not subject to load, the output torque is zero, and the internal pressure of the actuator is zero. The bending angle begins to increase. When the injected gas mass is 1.2 g, the angle reaches 60 degrees at 3.7 seconds. In the second stage (3.7 - 5 seconds), when the actuator contacts the outer skin of the tibia, it encounters resistance. As the injected air mass increases to 2.1 g, the pressure rises to 23 kPa and the torque increases to 21.2 N·m, and the bending angle stabilizes at 60°. In the third stage (5 - 10 seconds), as the air mass and pressure continue to increase until the torque of the actuator exceeds the load torque, the acceleration generated by this torque difference drives the knee joint angle to 86°. In the final stage (after 10 seconds), the torque of the actuator is equal to the load torque of 21.2 N·m, the bending angle remains at 86°, the air mass is 2.8 g, and the pressure is 27 kPa. The air mass flow rate during the simulation is less than 1 g / sec; the volumetric air flow rate is less than 30 L / min.

[0164] When the pneumatic soft actuator 10 is applied to the knee joint of the human body, the simulation system is the same as the system applied to the ankle joint above, except that the load module is used to simulate the load at the ankle joint. At this time, the computer simulation parameters of the control system are: the load variation range is 0 - 54.7 N.m; the moment of inertia variation range is 0 - 0.69 kg.m2; the damping coefficient c = 2.5, and the stiffness coefficient k = 0.02. The bending angle range is 40 - 157.1°, the time delay of air diffusion in the actuator model is = 0.003 s, and the simulation time is: 0 - 20 seconds. The simulation results are as Figure 14 shown.

[0165] At this time, the simulation is also divided into four stages. In the first stage (0 - 2.5 seconds), the load on the actuator is zero, the output torque is zero, and the bending angle begins to increase. The angle reaches 40° at 2.5 seconds, and the injected air mass is 0.6 g. In the second stage (2.5 - 3.7 seconds), when the actuator contacts the outer skin of the tibia, it encounters resistance, the torque of the actuator increases, and the bending angle stabilizes at 40°. When the injected air mass increases to 2.2 g, the pressure rises from 0 to 48 kPa, and the torque increases from 0 to 54.7 N·m. In the third stage (3.7 - 9 seconds), the further increase in air mass and pressure causes the torque of the actuator to exceed the load torque, and the acceleration generated by this torque difference drives the knee joint angle to 157.1°. In the final stage (after 9 seconds), the bending angle remains stable at 157.1°, the air mass is 2.7 g, the pressure is 4.2 kPa, and the torque is 0 N·m. During the entire simulation process, the air mass flow rate varies between 1.6 g / sec and -1.5 g / sec, where the negative flow rate indicates that air is pumped out of the actuator. The volumetric air flow rate ranges between 62 L / min and -43 L / min.

[0166] As can be seen from the above simulation results, when the above sensorless control method is adopted, the pneumatic soft actuator 10 without an angle sensor can effectively support an individual to transition from a squatting posture to a standing posture, even when bearing an additional weight of 90 kg (such as a barbell). The actuator body 140 operates at an internal gas pressure below 50 kPa, minimizing the discomfort of the wearer while improving the overall usability and comfort.

[0167] Based on the same concept as the sensorless control method of the above pneumatic soft actuator 10, referring to Figure 15 , an embodiment of the present invention further provides a sensorless control device 40, which can be applied to Figure 1 the controller 110 of the pneumatic soft actuator 10 shown in the figure, and includes a detection module 410, an estimation module 420, an error acquisition module 430, and a control adjustment module 440.

[0168] The detection module 410 is configured to obtain the current measurement value of the detection module 410.

[0169] Among them, the measurement value includes the injected air quality and pressure.

[0170] The estimation module 420 is configured to input the injected air quality and pressure into a preset actuator estimation model, and use the actuator estimation model to obtain the current actuator estimation amount of the actuator body 140.

[0171] Among them, the actuator estimation model is a model obtained by training with the experimental data of the actuator body 140, and represents the relationship between the input amount and the actuation amount of the actuator body 140.

[0172] The error acquisition module 430 is configured to obtain an error value according to the actuator estimation amount and the target control amount.

[0173] The control adjustment module 440 is configured to input the error value into the controller 110, so that the controller 110 adjusts the gas injection amount of the air pump module based on the error value to adjust the actuation amount of the actuator body 140.

[0174] In addition, the sensorless control device 40 further includes a model acquisition module, and the model acquisition module can be deployed on a computer device communicatively connected to the controller 110 of the pneumatic soft actuator 10.

[0175] The model acquisition module is configured to: obtain a plurality of experimental data of the actuator body 140; among them, the experimental data includes an input amount and an output amount, the output amount includes an angle and / or a torque, and the input amount includes the injected air quality and pressure; label each experimental data with the output amount to obtain a sample set; and use the sample set and Bayesian regularization to iteratively train a two-layer feedforward neural network to obtain an actuator estimation model.

[0176] The sensorless control device 40 of the pneumatic soft actuator 10 mentioned above, under the synergistic effect of the detection module 410, the estimation module 420, the error acquisition module 430, and the control adjustment module 440, can achieve precise control of the actuator body 140 only based on the measured values (i.e., the injected air quality and pressure) of the detection module 410 arranged on the air pipe, without installing sensors for measuring torque, angle, etc. on the actuator body 140, greatly reducing the volume and gravity of the actuator body 140, thereby enhancing the flexibility, comfort, and portability of the actuator body 140. At the same time, it reduces or even eliminates the limitation of the sensor weight and size on the force output of the soft actuator, and reduces the risk of mechanical and electronic failures.

[0177] Regarding the specific implementation and effects of the sensorless control device 40, reference can be made to the description of the implementation of the sensorless control method in the above text. For example, for the specific implementation and effects of the detection module 410, reference can be made to the description of the relevant content in step 11 above; for the specific implementation and effects of the estimation module 420, reference can be made to the description of the relevant content in step 13 above; for the specific implementation and effects of the error acquisition module 430, reference can be made to the description of the relevant content in step 15 above; for the specific implementation and effects of the control adjustment module 440, reference can be made to the description of the relevant content in step 17 above, which will not be elaborated here.

[0178] In addition, each module of the above sensorless control device 40 can be implemented in whole or in part by software, hardware, and their combination. The above modules can be embedded in or independent of the processor 220 in the electronic device 20 in the form of hardware, or stored in the memory 210 of the electronic device 20 in the form of software, so that the processor 220 can call and execute the operations corresponding to each of the above modules to implement the sensorless control method of the pneumatic soft actuator 10 provided in the above text.

[0179] The embodiment of the present invention also provides an electronic device 20, including a processor 220 and a memory 210. The memory 210 stores a computer program that can be executed by the processor 220, and the processor 220 can execute the computer program to implement the sensorless control method of the pneumatic soft actuator 10 proposed in the embodiment of the present invention.

[0180] The embodiment of the present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by the processor 220, it implements the sensorless control method of the pneumatic soft actuator 10 proposed in the embodiment of the present invention.

[0181] To sum up, the sensorless control method, device, pneumatic soft actuator, and electronic device provided by the embodiment of the present invention have at least the following beneficial effects:

[0182] (1) There is no need to install a large-volume sensor on the actuator body. Precise control of the actuator body can be achieved only based on the injected air quality and pressure, greatly reducing the volume and weight of the actuator body, thereby improving the flexibility, comfort, and portability of the actuator body.

[0183] (2) It improves the limitation of the weight and size of the sensor on the force output of the soft actuator. The actuator without a sensor is not afraid of collision and moisture, and can even be used in water, reducing the risk of mechanical and electronic failures.

[0184] (3) The real-time first angular quantity and / or first torque quantity of the actuator body are obtained through an actuation estimation model, and combined with the PID control principle for control, realizing high-precision and high-response control.

[0185] In several embodiments provided by the present invention, it should be understood that the disclosed devices and methods can also be implemented in other ways. The device embodiments described above are only illustrative. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architectures, functions, and operations of the devices, methods, and computer program products according to multiple embodiments of the present invention. In this regard, each block in the flowchart or block diagram may represent a module, a program segment, or a part of code, and the module, program segment, or part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order from that marked in the accompanying drawings. For example, two consecutive blocks can actually be executed substantially in parallel, and they can sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, as well as the combination of blocks in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system for performing the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0186] In addition, the functional modules in each embodiment of the present invention can be integrated together to form an independent part, or each module can exist alone, or two or more modules can be integrated to form an independent part.

[0187] When the above-mentioned functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories, random access memories, magnetic disks, or optical discs that can store program codes.

[0188] The foregoing are only the preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A sensorless control method for a pneumatic soft actuator, characterized in that, The pneumatic soft actuator includes a controller, a gas pump module, a detection module, and an actuator body. There is no need to install a large-volume sensor on the actuator body. The gas pump module is connected to the actuator body through a gas pipe. The detection module is arranged on the gas pipe. The method includes: Obtain the current measured value of the detection module; wherein, the measured value includes the injected air quality and pressure; Input the injected air quality and the pressure into a preset actuation estimation model, and use the actuation estimation model to obtain the current actuation estimation amount of the actuator body; wherein, the actuation estimation model is a model obtained by training with the experimental data of the actuator body, which characterizes the relationship between the input amount and the actuation amount of the actuator body; Obtain an error value according to the actuation estimation amount and the target control amount; Input the error value into the controller, so that the controller adjusts the gas injection amount of the gas pump module based on the error value to adjust the actuation amount of the actuator body.

2. The sensorless control method of the pneumatic soft actuator according to claim 1, characterized in that The method for obtaining the actuation estimation model includes: Obtain a plurality of experimental data of the actuator body; wherein, the experimental data includes an input amount and an output amount, the output amount includes an angle and / or a torque, and the input amount includes the injected air quality and pressure; Label each piece of experimental data with the output amount to obtain a sample set; Use the sample set and Bayesian regularization to perform iterative training on a two-layer feedforward neural network to obtain an actuation estimation model.

3. The sensorless control method of the pneumatic soft actuator according to claim 2, characterized in that The sample set includes a first sample set labeled with an angle and a second sample set labeled with a torque; The step of using the sample set and Bayesian regularization to perform iterative training on a two-layer feedforward neural network to obtain an actuation estimation model includes: Use Bayesian regularization to perform iterative training on the two-layer feedforward neural network with the first sample set to obtain an angle estimation model; Use Bayesian regularization to perform iterative training on the two-layer feedforward neural network with the second sample set to obtain a torque estimation model.

4. The sensorless control method of the pneumatic soft actuator according to any one of claims 1 to 3, characterized in that, The detection module includes a gas flow sensor and a pressure sensor; The step of obtaining the current measured value of the detection module includes: Sample the gas flow sensor and the pressure sensor to obtain the real-time gas flow rate and pressure; Use the time period from the starting moment to the current moment as the target time period, and integrate the gas flow rates in the target time period in combination with the average molar mass of the gas output by the gas pump to obtain the injected air quality at the current moment.

5. The sensorless control method of the pneumatic soft actuator according to any one of claims 1 to 3, characterized in that, The actuation estimation amount includes a first angle amount and a first torque amount; When the target control amount is an angle, the step of obtaining an error value according to the actuation estimation amount and the target control amount includes: Input the first torque amount and the pressure into a preset angle prediction model, and use the angle prediction model to generate a second angle amount; wherein, the angle prediction model is a model obtained by training with the experimental data of the actuator body; Fuse the first angle amount and the second angle amount to obtain the current angle value; Take the difference between the current angle value and the target control amount as the error value.

6. The sensorless control method of the pneumatic soft actuator according to any one of claims 1 to 3, characterized in that The actuation estimation amount includes a first angle amount and a first torque amount; When the target control amount is torque, the step of obtaining the error value according to the actuation estimation amount and the target control amount includes: Input the first angle amount and the pressure into a preset torque prediction model, and use the torque prediction model to generate a second torque amount; wherein, the torque prediction model is a model obtained by training with experimental data of the actuator body; Fuse the first torque amount and the second torque amount to obtain the current torque value; Take the difference between the current torque value and the target control amount as the error value.

7. The sensorless control method of the pneumatic soft actuator according to any one of claims 1 to 3, characterized in that, The actuator body includes a flexible integrated board and a plurality of airbags arranged on the flexible integrated board; One end of the airbag is arranged on the flexible integrated board, and the distance between the arrangement positions of every two adjacent airbags is equal to a preset value; Each airbag is provided with a pore, and the airbag is communicated with the trachea through the pore.

8. A sensorless control device for a pneumatic soft actuator, characterized in that, The pneumatic soft actuator includes a controller, a gas pump module, a detection module and an actuator body. No large-volume sensor needs to be installed on the actuator body. The gas pump is communicated with the actuator body through a trachea. The detection module is arranged on the trachea. The device includes a detection module, an estimation module, an error acquisition module and a control adjustment module; The detection module is used to obtain the current measurement value of the detection module; wherein, the measurement value includes the injected air quality and pressure; The estimation module is used to input the injected air quality and the pressure into a preset actuation estimation model, and use the actuation estimation model to obtain the current actuation estimation amount of the actuator body; wherein, the actuation estimation model is a model obtained by training with experimental data of the actuator body, representing the relationship between the input amount and the actuation amount of the actuator body; The error acquisition module is used to obtain the error value according to the actuation estimation amount and the target control amount; The control adjustment module is used to input the error value into the controller, so that the controller adjusts the gas injection amount of the gas pump module based on the error value to adjust the actuation amount of the actuator body.

9. A pneumatic soft actuator, characterized in that, It includes a controller, a gas pump module, a detection module and an actuator body. The actuator body includes a flexible integrated board and a plurality of airbags arranged on the flexible integrated board. No large-volume sensor needs to be installed on the actuator body; One end of the airbag is arranged on the flexible integrated board, and the distance between the arrangement positions of every two adjacent airbags is equal to a preset value. Each airbag is provided with a pore; The gas pump module is communicated with the pore of the airbag through a trachea, and the detection module is arranged on the trachea; The detection module is used to continuously detect the gas injected by the gas pump module into the actuator body to obtain a real-time measurement value; The controller is used to implement the sensorless control method of the pneumatic soft actuator according to any one of claims 1 to 7.

10. An electronic device, characterized in that, It includes a processor and a memory, and the memory stores a computer program that can be executed by the processor. The processor can execute the computer program to implement the sensorless control method of the pneumatic soft actuator as described in any one of claims 1 to 7.

Images