Intelligent pneumatic manipulator control system

Abstract

The intelligent pneumatic manipulator control system of the application comprises a host computer, a microcontroller, a compressed air source, an electrical proportional valve, an electromagnetic valve group, an air pressure sensor array and a pneumatic soft actuator. The microcontroller obtains the control state of the electrical proportional valve and each electromagnetic valve according to the opening and closing state of the detected local channel enable switch to control the air supply and cut-off of the corresponding chamber. The control method of the application comprises: a) parameter setting; b) system initialization; c) pressure sampling and filtering; d) main path pressure supply; e) manual control; f) automatic control; g) manual takeover control; h) air path coupling feedforward control; i) branch control and output protection; j) data uploading and receiving. The pneumatic manipulator control system and method of the application improve the safety of on-site operation, the stability of pressure, and the service life of the actuator through local manual priority takeover, manual / automatic non-disturbance switching, cascaded anti-noise filtering, steady-state dead zone maintenance and actuator service life statistics.

Description

[0001] This invention relates to a robotic arm control system and method, and more specifically, to a pneumatically driven robotic arm control system and method. Background Technology

[0002] Generally, there are two types of driving methods for robotic arms: electromagnetic motor driving and fluid driving. Electromagnetic motor driving uses a motor to drive the extension and contraction of tendons and ligaments to move the corresponding joints and complete the corresponding actions of the robotic arm. Fluid driving is divided into hydraulic driving and pneumatic driving. Pneumatic driving achieves mechanical movement by controlling the gas pressure in the corresponding pipeline. Due to the compressibility of air, pneumatically driven robotic arms have better compliance, making them suitable for applications where compliance with contact forces is high, such as massage robotic arms.

[0003] Pneumatic manipulators (also known as pneumatic soft actuators) utilize compressed air to drive flexible cavities to expand, bend, elongate, or change stiffness, featuring compliant structure, light weight, and high human-machine interface safety. To achieve multi-degree-of-freedom movements, practical systems typically employ multiple independent air chambers, with each chamber's inflation and deflation controlled separately by multi-way valves. When multiple branches share a common air source, the instantaneous flow demand of the common main pipe increases significantly when multiple branches open at similar times, easily causing fluctuations in supply pressure and branch flow. If each branch only relies on conventional PID control for passive correction after pressure drop, slow recovery, overshoot, and inter-branch interference can easily occur.

[0004] When the branch-on / off solenoid valve is driven by PWM, periodic current changes and electromagnetic interference will occur, which will affect the ADC sampling of the analog pressure sensor and the local potentiometer. Single sampling or simple averaging is difficult to simultaneously achieve both pulse spike suppression and dynamic response.

[0005] In laboratory, rehabilitation, and soft robotics applications, systems often require both remote setup via a host computer and manual adjustment on-site. Without physical authorization, mode prioritization, and seamless switching mechanisms, issues such as remote commands overriding on-site operations, sudden changes in duty cycle during switching, and pressure shocks may occur. Furthermore, existing systems often fragment control, display, and data recording, making it difficult to establish a complete experimental traceability chain. Summary of the Invention

[0006] In order to overcome the shortcomings of the above-mentioned technical problems, the present invention provides an intelligent pneumatic manipulator control system and method.

[0007] The intelligent pneumatic manipulator control system of the present invention includes a host computer, a microcontroller, a compressed air source, an electro-proportional valve, a solenoid valve group, a pressure sensor array, and a pneumatic soft actuator. The microcontroller communicates with the host computer via a USART serial port. The solenoid valve group and the pressure sensor array are each composed of several solenoid valves and several pressure sensors. The air supply port of the compressed air source is connected to the air inlet of the electro-proportional valve, and the air outlet of the electro-proportional valve is connected to the common air inlet of the solenoid valve group. The air outlets of several solenoid valves are connected to different air supply chambers of the pneumatic soft actuator via pipelines. The electro-proportional valve is used to control the main circuit pressure supply, and the solenoid valves control the branch circuit pressure supply. The microcontroller's signal input terminal is connected to multiple local channel enable switches and multiple local rotary potentiometers. The microcontroller's signal output terminal is connected to a MOSFET switching module. The output terminal of the MOSFET switching module is connected to the control terminal of the electro-proportional control valve and each solenoid valve in the solenoid valve group. The pressure detection port of the pressure sensor is connected to the air outlet of the solenoid valve via a pipeline, and the signal output terminal of the pressure sensor is connected to the signal input terminal of the microcontroller. The microcontroller obtains the control status of the electro-proportional valve and each solenoid valve based on the detected open / closed state of the local channel enable switch, thereby enabling the corresponding solenoid valve to control the air supply and cut-off of the corresponding chamber in the pneumatic soft actuator. In manual mode, the microcontroller obtains the target pressure value of the corresponding solenoid valve outlet based on the acquired value of the local rotary potentiometer. In automatic mode, the microcontroller controls the pressure of the corresponding solenoid valve outlet based on the target pressure value sent by the host computer. The microcontroller outputs a PWM signal indicating the on / off ratio between the electro-proportional valve and the solenoid valve through the MOSFET switching module. The air pressure sensors in the air pressure sensor array are used to provide feedback on the real-time pressure of the corresponding solenoid valve outlet. The microcontroller sends the target pressure, real-time pressure, control status, and PWM signal to the host computer via serial communication, and receives the target air pressure value and control commands sent by the host computer.

[0008] The intelligent pneumatic manipulator control system of the present invention includes a solenoid valve group comprising 10 three-port two-position solenoid valves, a pressure sensor array comprising 9 pressure sensors, a local channel enable switch comprising 10 switches, and a local rotary potentiometer comprising 10 potentiometers.

[0009] The intelligent pneumatic manipulator control system of the present invention uses an STM32F103ZET6 microcontroller. The microcontroller calls two analog-to-digital converter peripherals, ADC1 and ADC3, and configures DMA1 channel 1 and DMA2 channel 5 respectively to form two independent ADC and DMA continuous acquisition links. The analog-to-digital converter peripheral ADC1 operates in scanning and continuous conversion mode. Its regular conversion sequence includes 15 analog input channels, of which the first 9 conversion sequences are used to acquire the analog output signals of 9 pressure sensors, and the last 6 conversion sequences are used to acquire the analog output signals of the first 6 local rotary potentiometers. The analog-to-digital converter peripheral ADC3 conversion sequence includes 4 analog input channels, which are used to acquire the analog output signals of the last 4 local rotary potentiometers 6.

[0010] The intelligent pneumatic manipulator control system of the present invention uses a microcontroller to call five timers TIM8, TIM1, TIM2, TIM3 and TIM4. TIM8 is used to generate a 1kHz PWM control signal required by the main electro-proportional valve. The multiple PWM output channels of TIM1, TIM2 and TIM4 together form 10 branch PWM outputs, which are used to drive 10 three-port two-position solenoid valves in the solenoid valve group through MOSFET switching modules. TIM3 is used to generate a 1ms system time base, providing a unified time reference for pressure sampling, control cycle timing, mode switching, anti-jitter judgment and valve life statistics.

[0011] The control method of the intelligent pneumatic manipulator control system of the present invention is characterized by being implemented through the following steps: a) Parameter setting; according to the required movement mode of the pneumatic soft actuator, close the corresponding switch in the local channel enable switch to enable the corresponding proportional valve in the electro-proportional valve; according to the control requirements, turn each potentiometer in the local rotary potentiometer to the corresponding position so as to determine the target pressure input from the solenoid valve to each chamber of the pneumatic soft actuator in manual control mode; in automatic mode, determine the control parameters sent from the host computer to the microcontroller. b) System initialization; the microcontroller initializes the USART, GPIO, ADC, DMA, PWM timer and control parameters, and sets the outputs of the electric proportional valve and solenoid valve group to zero during the start-up protection time; c) Pressure sampling and filtering: The microcontroller cyclically acquires pressure signals from 9 barometric pressure sensors and 10 potentiometer signals, and writes the 9 effective pressure ADC samples into the corresponding sliding windows at fixed intervals; then, it performs extreme value elimination averaging on the sliding window data of the pressure signals from the 9 barometric pressure sensors, performs exponential low-pass filtering, and converts it into pressure values ​​through independent channel calibration parameters; d) Main line pressure supply: The microcontroller calculates the main line target air pressure value with pressure margin based on the user's main line setting and the maximum value of the target pressure of all enabled branches, and maps the target air pressure value to the high-frequency PWM duty cycle of the electro-proportional valve to realize the pressure supply to the common air inlet; e) Manual control; read the status of the 10 local channel enable switches. When the switch is closed, the corresponding channel gains control authority. When the switch is open, the output of the corresponding branch is cleared. The ADC values ​​of the 10 local rotary potentiometers are converted into the branch PWM duty cycle through dead-time processing and linear mapping. After limiting, the corresponding branch solenoid valve is driven to open and close to control the pressure value of the corresponding branch to reach the target air pressure value, thereby driving the pneumatic soft actuator to perform the corresponding action. f). Automatic control; When the actual pressure measured by the air pressure sensor array in manual control mode deviates from the pressure corresponding to the local rotary potentiometer within the threshold for a preset time, or when a target pressure command is received from the host computer, the target pressure is locked and the automatic closed loop is entered. The actual valve duty cycle before switching is preloaded into the PID integral term for automatic control. g) Manual takeover control: In automatic control mode, compare the current duty cycle of the local rotary potentiometer with the reference duty cycle when entering automatic mode. When the deviation exceeds the takeover threshold and is established for multiple consecutive control cycles, switch back to local manual mode and clear or reset the PID integrator. h). Gas path coupling feedforward control; For 9 branches, when the target gas pressure value of a branch changes, the predicted pressure drop compensation caused by the shared gas source is calculated based on the actual PWM duty cycle of the other enabled branches. This compensation is then combined with the target pressure and used in PID calculation. The pressure feedforward compensation amount is calculated based on the actual PWM output state of the other enabled branches to achieve coupled feedforward compensation control of the main road and the enabled branches. i) Branch control and output protection; PID pressure closed-loop control is used to regulate the air pressure of 9 branches. When the error between the actual air pressure value and the target air pressure value enters the dead zone, the duty cycle of the previous cycle is maintained. When the error exceeds the dead zone, the minimum effective duty cycle, upper and lower limit saturation and unit cycle change rate limit are implemented. j) Data upload and reception: The host computer sends the target air pressure values ​​of each branch, the target air pressure value of the main branch, and control commands to the microcontroller. The microcontroller returns the actual air pressure values, target air pressure values, manual or automatic control status, PWM duty cycle, valve action count, and cumulative power-on time of each branch to the host computer control terminal.

[0012] The beneficial effects of the present invention are as follows: The intelligent pneumatic manipulator control system and control method of the present invention mainly have the following beneficial effects: (1) The intelligent pneumatic manipulator control system and control method of the present invention addresses the problem of common air supply disturbance caused by multiple branches using air simultaneously when 10 pneumatic branches share the main air path. The microcontroller calculates the air path coupling feedforward compensation amount of the current closed-loop branch based on the actual PWM duty cycle of other enabled branches, and adds it to the target pressure of the current branch before participating in PID control. Branches 1 to 9 are used as pressure closed-loop objects, and the actual air usage state of the 10th branch still participates in the disturbance compensation of other closed-loop branches.

[0013] (2) The main air circuit adopts the ITV2050-312L electric proportional valve. The STM32 generates a 1 kHz PWM signal and generates the proportional valve control voltage through the level conversion and PWM to 0~10 V module. The 10 branches adopt the SMC VQ100 series three-port two-position normally closed solenoid valve, and use 40 Hz PWM for time proportional supply and exhaust. Among them, branches 1 to 9 are equipped with end pressure feedback, and branch 10 is an open loop or reserved branch without pressure feedback.

[0014] (3) The local channel enable switch serves as the physical authorization condition for branch output. After the channel is enabled, it defaults to manual potentiometer control; after branches 1 to 9 reach the manual stability condition, the current target pressure is locked and switched to automatic pressure closed loop through integral preloading. In automatic mode, when the potentiometer offset relative to the reference position exceeds the control threshold for multiple consecutive control cycles, the channel exits the current automatic mode and returns to local manual mode.

[0015] (4) The system uses two sets of ADC-DMA continuous acquisition links to acquire signals from 9 effective pressure sensors and 10 potentiometers, and reserves 1 pressure acquisition interface. The pressure signal is sequentially passed through sliding window extreme value elimination averaging, exponential low-pass filtering and independent channel calibration before entering the branch pressure closed loop, so as to reduce the impact of PWM electromagnetic interference and individual sensor differences on control stability. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the intelligent pneumatic manipulator control system of the present invention.

[0017] In the diagram: 1. Host computer, 2. Microcontroller, 3. Control algorithm module, 4. MOSFET switch module, 5. Local channel enable switch, 6. Local rotary potentiometer, 7. Compressed air source, 8. Electro-proportional valve, 9. Solenoid valve group, 10. Pneumatic soft actuator, 11. Air pressure sensor array. Detailed Implementation

[0018] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0019] The purpose of this invention is to provide a multi-channel intelligent pneumatic manipulator control system and its control method. On the basis of ensuring independent pressure closed loop of each branch, the system reduces pressure fluctuations when multiple channels operate simultaneously by means of high-frequency pressure supply to the main channel, low-frequency drive to the branches, shared air source feedforward compensation, local manual priority takeover, manual / automatic seamless switching, cascaded noise reduction filtering, steady-state dead zone maintenance, and actuator life statistics. This improves on-site operation safety, pressure stabilization accuracy, actuator life, and data traceability.

[0020] like Figure 1 The diagram shows the schematic of the intelligent pneumatic manipulator control system of the present invention. It consists of a host computer 1, a microcontroller 2, a compressed air source 7, an electro-proportional valve 8, a solenoid valve group 9, a pneumatic soft actuator 10, a pressure sensor array 11, a MOSFET switching module 4, a local channel enable switch 5, a local rotary potentiometer 6, and the pneumatic soft actuator 10. The host computer 1 and the microcontroller 2 are connected via a USART serial communication interface to achieve data transmission between them. The compressed air source 7 provides compressed gas. The air supply port of the compressed air source 7 is connected to the air inlet of the electro-proportional valve 8, and the air outlet of the electro-proportional valve 8 is connected to the common air inlet of the solenoid valve group 9. The pipeline between the electro-proportional valve 8 and the solenoid valve group 9 is the main pipeline, and the main pipeline pressure can be achieved under the control of the electro-proportional valve 8. The solenoid valve group 9 consists of 10 three-port two-position solenoid valves. The air inlet of each solenoid valve is the same as the main line, and the air outlet is connected to different air supply chambers of the pneumatic soft actuator through pipelines to form a branch line. The solenoid valve controls the air supply of the branch line.

[0021] The electro-proportional valve 8 is used to adjust the basic air supply pressure of the common main circuit according to the electrical input command. When the microcontroller 2 uses the TIM8 to generate a 1kHz PWM signal as the intermediate control quantity of the main circuit, the analog command required by the electro-proportional valve 8 is formed through the PWM signal to voltage conversion module. The 10 branch solenoid valves correspond to 10 pneumatic branches. Each branch solenoid valve 10 is a three-port two-position direct-acting solenoid valve. Each branch solenoid valve 10 includes an inlet port P, a working port A, and an exhaust port R. The inlet port P is connected to the common air supply channel of the 10-position solenoid valve integrated manifold 9, and the exhaust port R is connected to the common exhaust channel of the solenoid valve.

[0022] Pressure sensor 11 is positioned between the inlet of the corresponding branch solenoid valve and the pneumatic soft actuator to collect the end pressure between the output of the branch solenoid valve 10 and the cavity of the pneumatic soft actuator in real time. Compared with placing the pressure sensor at the inlet of the main air path or manifold, this arrangement reduces the influence of pipeline pressure drop, gas flow resistance, and aerodynamic coupling between branches on the pressure measurement results, making the collected pressure closer to the actual pressure acting on the internal cavity of the pneumatic soft actuator.

[0023] The branch solenoid valve adopts a normally closed SMC VQ100 series three-port two-position solenoid valve. When the solenoid valve coil is energized, the inlet port P is connected to the working port A, and the compressed gas enters the corresponding pneumatic soft actuator through the working solenoid valve integrated manifold and the branch solenoid valve. When the solenoid valve coil is de-energized, the working port A is connected to the exhaust port R, allowing the gas in the connected pneumatic soft actuator to be discharged through the common exhaust channel. The lower-level computer adjusts the effective air supply time and exhaust time of the branch solenoid valve in a unit control cycle by changing the duty cycle of the PWM control signal applied to the branch solenoid valve 10, thereby changing the equivalent gas flow rate of the corresponding branch and realizing closed-loop regulation of the cavity pressure of the pneumatic soft actuator.

[0024] Branch solenoid valve 10 is a valve assembly of an SMC VQ100 series three-way solenoid valve. When the target pressure is zero, the corresponding branch enters the venting or shut-off state. The main proportional valve is used to regulate the base supply pressure of the common main pipe, and the branch solenoid valves are used to regulate the local pressure of each actuator cavity.

[0025] When the target pressure setting value of the corresponding branch is greater than zero, the lower computer calculates the PWM duty cycle of the branch solenoid valve 10 based on the deviation between the target pressure and the actual pressure collected by the pressure sensor 11, and drives the corresponding solenoid valve to act through the MOSFET drive matrix; when the target pressure setting value is zero, the lower computer stops the air supply control of the corresponding branch, so that the branch solenoid valve 10 enters the exhaust state with the working port A connected to the exhaust port R, until the cavity pressure of the corresponding pneumatic soft actuator drops to the preset safe pressure range.

[0026] The main circuit's electro-proportional valve 8 is used to regulate the base air supply pressure provided by the compressed air source 7 to the main circuit, providing a common pressure source for the 10 pneumatic branch circuits. The 10 branch circuit solenoid valves are used to regulate the local air supply, exhaust, and chamber pressure of the corresponding pneumatic soft actuators. By combining the main circuit's proportional pressure regulation with the branch circuit's PWM pressure regulation, a hierarchical pneumatic control structure is formed, combining main circuit pressure coordination with independent closed-loop control of the branch circuits. This allows multiple pneumatic soft actuators to achieve independent pressure setting and dynamic control while sharing a compressed air source.

[0027] Microcontroller 2 calls two analog-to-digital converter peripherals, ADC1 and ADC3, and configures DMA1 channel 1 and DMA2 channel respectively to form two independent ADC-DMA continuous acquisition links to continuously and cyclically acquire the analog signals output by 9 pressure sensors 11 and 10 local rotary potentiometers 6; reads the on / off status of 10 local channel enable switches 5 through GPIO input ports; outputs main and branch PWM signals through 5 hardware timers; and exchanges target pressure, real-time pressure, control status, valve duty cycle and life statistics with the host computer control terminal 1 through serial port.

[0028] In the intelligent pneumatic manipulator control system, microcontroller 2 uses an STM32F103ZET6, calling two analog-to-digital converter peripherals, ADC1 and ADC3, and configuring DMA1 channel 1 and DMA2 channel respectively to form two independent ADC and DMA continuous acquisition links. ADC1 operates in scanning and continuous conversion mode, and its regular conversion sequence includes 15 analog input channels. The first 9 conversion sequences are used to acquire the analog output signals of 9 pressure sensors 11, and the last 6 conversion sequences are used to acquire the analog output signals of the first 6 local rotary potentiometers 6. The conversion results of ADC1 are automatically stored in the first sampling buffer through DMA1 channel 1 in a cyclic transmission manner.

[0029] ADC3 also operates in scanning and continuous conversion modes. Its regular conversion sequence includes four analog input channels for acquiring the analog output signals of the four local rotary potentiometers 6. The conversion results of ADC3 are automatically stored in the second sampling buffer via DMA2 channel 5 in a cyclic transfer manner. Thus, microcontroller 2 can complete the continuous acquisition of 19 analog signals, including 9 pressure sensor signals and 10 potentiometer signals, with relatively low processor resource consumption.

[0030] Microcontroller 2 reads the status of each of the 10 local channel enable switches 5 through the 10 digital input ports of GPIOG and GPIOE. Each local channel enable switch 5 corresponds to one branch and serves as a physical authorization condition for allowing the branch output.

[0031] Microcontroller 2 calls five timers: TIM8, TIM1, TIM2, TIM3, and TIM4. TIM8 generates the 1 kHz PWM control signal required by the main electro-proportional valve 8. Multiple PWM output channels of TIM1, TIM2, and TIM4 together form 10 branch PWM outputs, which drive 10 SMC VQ100 series three-port two-position solenoid valves respectively via MOSFET drive matrix 4. TIM3 generates a 1 ms system time base, providing a unified time reference for pressure sampling, control cycle timing, mode switching, anti-jitter judgment, and valve life statistics.

[0032] Microcontroller 2 communicates bidirectionally with host computer control terminal 1 via USART1. The serial communication parameters are set to 115200 bit / s, 8 data bits, 1 stop bit, and no parity. Host computer 1 sends the target air pressure values ​​of each branch, the target air pressure value of the main branch, and control commands to microcontroller 2. Microcontroller 2 returns to host computer 1 the actual air pressure values ​​of each branch, the target air pressure values, the manual or automatic control status, the PWM duty cycle, the number of valve actions, and the cumulative power-on time.

[0033] The main proportional valve uses a higher PWM frequency than the branch solenoid valves to improve the regulation resolution and response speed of the main supply pressure; the branch solenoid valves use a lower PWM frequency adapted to their mechanical response characteristics to achieve stable equivalent flow regulation. Preferably, the main PWM frequency and the branch PWM frequency are integer multiples of each other, and the control update cycle is an integer multiple of the branch PWM cycle, thereby reducing frequency skipping and scheduling jitter caused by the superposition of non-integer cycles.

[0034] The control method of the intelligent pneumatic manipulator control system of the present invention is achieved through the following steps: a) Parameter setting; according to the required movement mode of the pneumatic soft actuator, close the corresponding switch in the local channel enable switch to enable the corresponding proportional valve in the electro-proportional valve; according to the control requirements, turn each potentiometer in the local rotary potentiometer to the corresponding position so as to determine the target pressure input from the solenoid valve to each chamber of the pneumatic soft actuator in manual control mode; in automatic mode, determine the control parameters sent from the host computer to the microcontroller. b) System initialization; the microcontroller initializes the USART, GPIO, ADC, DMA, PWM timer and control parameters, and sets the outputs of the electric proportional valve and solenoid valve group to zero during the start-up protection time; c) Pressure sampling and filtering: The microcontroller cyclically acquires pressure signals from 9 barometric pressure sensors and 10 potentiometer signals, and writes the 9 effective pressure ADC samples into the corresponding sliding windows at fixed intervals; then, it performs extreme value elimination averaging on the sliding window data of the pressure signals from the 9 barometric pressure sensors, performs exponential low-pass filtering, and converts it into pressure values ​​through independent channel calibration parameters; d) Main line pressure supply: The microcontroller calculates the main line target air pressure value with pressure margin based on the user's main line setting and the maximum value of the target pressure of all enabled branches, and maps the target air pressure value to the high-frequency PWM duty cycle of the electro-proportional valve to realize the pressure supply to the common air inlet; e) Manual control; read the status of the 10 local channel enable switches. When the switch is closed, the corresponding channel gains control authority. When the switch is open, the output of the corresponding branch is cleared. The ADC values ​​of the 10 local rotary potentiometers are converted into the branch PWM duty cycle through dead-time processing and linear mapping. After limiting, the corresponding branch solenoid valve is driven to open and close to control the pressure value of the corresponding branch to reach the target air pressure value, thereby driving the pneumatic soft actuator to perform the corresponding action. f). Automatic control; When the actual pressure measured by the air pressure sensor array in manual control mode deviates from the pressure corresponding to the local rotary potentiometer within the threshold for a preset time, or when a target pressure command is received from the host computer, the target pressure is locked and the automatic closed loop is entered. The actual valve duty cycle before switching is preloaded into the PID integral term for automatic control. g) Manual takeover control: In automatic control mode, compare the current duty cycle of the local rotary potentiometer with the reference duty cycle when entering automatic mode. When the deviation exceeds the takeover threshold and is established for multiple consecutive control cycles, switch back to local manual mode and clear or reset the PID integrator. h). Gas path coupling feedforward control; For 9 branches, when the target gas pressure value of a branch changes, the predicted pressure drop compensation caused by the shared gas source is calculated based on the actual PWM duty cycle of the other enabled branches. This compensation is then combined with the target pressure and used in PID calculation. The pressure feedforward compensation amount is calculated based on the actual PWM output state of the other enabled branches to achieve coupled feedforward compensation control of the main road and the enabled branches. i) Branch control and output protection; PID pressure closed-loop control is used to regulate the air pressure of 9 branches. When the error between the actual air pressure value and the target air pressure value enters the dead zone, the duty cycle of the previous cycle is maintained. When the error exceeds the dead zone, the minimum effective duty cycle, upper and lower limit saturation and unit cycle change rate limit are implemented. j) Data upload and reception: The host computer sends the target air pressure values ​​of each branch, the target air pressure value of the main branch, and control commands to the microcontroller. The microcontroller returns the actual air pressure values, target air pressure values, manual or automatic control status, PWM duty cycle, valve action count, and cumulative power-on time of each branch to the host computer control terminal.

Claims

1. A control system for an intelligent pneumatic manipulator, comprising a host computer (1), a microcontroller (2), a compressed air source (7), an electro-proportional valve (8), a solenoid valve group (9), a pressure sensor array (11), and a pneumatic soft actuator (10), wherein the microcontroller communicates with the host computer via a USART serial port, and the solenoid valve group and the pressure sensor array are each composed of several solenoid valves and several pressure sensors; the air supply port of the compressed air source is connected to the air inlet of the electro-proportional valve, the air outlet of the electro-proportional valve is connected to the common air inlet of the solenoid valve group, and the air outlets of several solenoid valves are connected to different air supply chambers of the pneumatic soft actuator via pipelines; the electro-proportional valve is used to control the main circuit pressure supply, and the solenoid valve controls the branch circuit pressure supply; characterized in that: The microcontroller’s signal input terminal is connected to multiple local channel enable switches (5) and multiple local rotary potentiometers (6). The microcontroller’s signal output terminal is connected to a MOSFET switch module (4). The output terminal of the MOSFET switch module is connected to the control terminal of each solenoid valve in the electric proportional control valve and the solenoid valve group. The pressure detection port of the air pressure sensor is connected to the air outlet of the solenoid valve through a pipeline. The signal output terminal of the air pressure sensor is connected to the signal input terminal of the microcontroller. The microcontroller obtains the control status of the electro-proportional valve and each solenoid valve based on the detected open / closed state of the local channel enable switch, thereby enabling the corresponding solenoid valve to control the air supply and cut-off of the corresponding chamber in the pneumatic soft actuator. In manual mode, the microcontroller obtains the target pressure value of the corresponding solenoid valve outlet based on the acquired value of the local rotary potentiometer. In automatic mode, the microcontroller controls the pressure of the corresponding solenoid valve outlet based on the target pressure value sent by the host computer. The microcontroller outputs a PWM signal indicating the on / off ratio between the electro-proportional valve and the solenoid valve through the MOSFET switching module. The air pressure sensors in the air pressure sensor array are used to provide feedback on the real-time pressure of the corresponding solenoid valve outlet. The microcontroller sends the target pressure, real-time pressure, control status, and PWM signal to the host computer via serial communication, and receives the target air pressure value and control commands sent by the host computer.

2. The intelligent pneumatic manipulator control system according to claim 1, characterized in that: The solenoid valve assembly (9) contains 10 three-port two-position solenoid valves, the bar pressure sensor array (11) contains 9 bar pressure sensors, the local channel enable switch (5) consists of 10 switches, and the local rotary potentiometer (6) consists of 10 potentiometers.

3. The intelligent pneumatic manipulator control system according to claim 1 or 2, characterized in that: The microcontroller (2) is an STM32F103ZET6. The microcontroller calls two analog-to-digital converter peripherals, ADC1 and ADC3, and configures DMA1 channel 1 and DMA2 channel 5 respectively to form two independent ADC and DMA continuous acquisition links. The analog-to-digital converter peripheral ADC1 works in scanning and continuous conversion mode. Its regular conversion sequence includes 15 analog input channels. The first 9 conversion sequences are used to acquire the analog output signals of 9 pressure sensors, and the last 6 conversion sequences are used to acquire the analog output signals of the first 6 local rotary potentiometers (6). The analog-to-digital converter peripheral ADC3 conversion sequence includes four analog input channels, used to acquire the analog output signals of the last four local rotary potentiometers 6.

4. The intelligent pneumatic manipulator control system according to claim 3, characterized in that: The microcontroller (2) calls five timers TIM8, TIM1, TIM2, TIM3 and TIM4. TIM8 is used to generate the 1kHz PWM control signal required by the main proportional valve (8). The multiple PWM output channels of TIM1, TIM2 and TIM4 together form 10 branch PWM outputs, which are used to drive 10 three-port two-position solenoid valves in the solenoid valve group (9) through the MOSFET switching module (4). TIM3 is used to generate a 1ms system time base, which provides a unified time reference for pressure sampling, control cycle timing, mode switching, anti-jitter judgment and valve life statistics.

5. A control method based on the intelligent pneumatic manipulator control system according to claim 4, characterized in that, This can be achieved through the following steps: a). Parameter setting; according to the required movement mode of the pneumatic soft actuator (10), close the corresponding switch in the local channel enable switch (5) to enable the corresponding proportional valve in the electro-proportional valve (8); according to the control requirements, turn each potentiometer in the local rotary potentiometer (6) to the corresponding position so as to determine the target pressure of the solenoid valve input to each chamber of the pneumatic soft actuator in manual control mode; in automatic mode, determine the control parameters sent by the host computer (1) to the microcontroller (2); b) System initialization; the microcontroller initializes the USART, GPIO, ADC, DMA, PWM timer and control parameters, and sets the outputs of the electric proportional valve and solenoid valve group to zero during the start-up protection time; c) Pressure sampling and filtering; The microcontroller cyclically acquires pressure signals from 9 barometric pressure sensors and 10 potentiometer signals, and writes the 9 effective pressure ADC samples into the corresponding sliding windows at fixed intervals; then, it performs extreme value removal averaging on the sliding window data of the pressure signals from the 9 barometric pressure sensors, performs exponential low-pass filtering, and converts it into pressure values ​​through independent channel calibration parameters; d) Main line pressure supply: The microcontroller calculates the main line target air pressure value with pressure margin based on the user's main line setting and the maximum value of the target pressure of all enabled branches, and maps the target air pressure value to the high-frequency PWM duty cycle of the electro-proportional valve to realize the pressure supply to the common air inlet; e) Manual control; read the status of the 10 local channel enable switches. When the switch is closed, the corresponding channel gains control authority. When the switch is open, the output of the corresponding branch is cleared. The ADC values ​​of the 10 local rotary potentiometers are converted into the branch PWM duty cycle through dead-time processing and linear mapping. After limiting, the corresponding branch solenoid valve is driven to open and close to control the pressure value of the corresponding branch to reach the target air pressure value, thereby driving the pneumatic soft actuator to perform the corresponding action. f). Automatic control; When the actual pressure measured by the air pressure sensor array in manual control mode deviates from the pressure corresponding to the local rotary potentiometer within the threshold for a preset time, or when a target pressure command is received from the host computer, the target pressure is locked and the automatic closed loop is entered. The actual valve duty cycle before switching is preloaded into the PID integral term for automatic control. g) Manual takeover control: In automatic control mode, compare the current duty cycle of the local rotary potentiometer with the reference duty cycle when entering automatic mode. When the deviation exceeds the takeover threshold and is established for multiple consecutive control cycles, switch back to local manual mode and clear or reset the PID integrator. h). Gas path coupling feedforward control; For 9 branches, when the target gas pressure value of a branch changes, the predicted pressure drop compensation caused by the shared gas source is calculated based on the actual PWM duty cycle of the other enabled branches. This compensation is then combined with the target pressure and used in PID calculation. The pressure feedforward compensation amount is calculated based on the actual PWM output state of the other enabled branches to achieve coupled feedforward compensation control of the main road and the enabled branches. i) Branch control and output protection; PID pressure closed-loop control is used to regulate the air pressure of 9 branches. When the error between the actual air pressure value and the target air pressure value enters the dead zone, the duty cycle of the previous cycle is maintained. When the error exceeds the dead zone, the minimum effective duty cycle, upper and lower limit saturation and unit cycle change rate limit are implemented. j) Data upload and reception: The host computer sends the target air pressure values ​​of each branch, the target air pressure value of the main branch, and control commands to the microcontroller. The microcontroller returns the actual air pressure values, target air pressure values, manual or automatic control status, PWM duty cycle, valve action count, and cumulative power-on time of each branch to the host computer control terminal.

Images