Pneumatic actuators with embedded displacement sensing units and their intelligent position control systems

Abstract

The pneumatic actuator with an embedded displacement sensing unit and its intelligent position control system integrate a fixed resistive element module and a sliding contact module inside the pneumatic actuator. Through the distributed contact between the annular brush of the sliding contact module and the circuit board of the fixed resistive element module, the linear motion of the piston is continuously converted into an electrical signal, achieving compact and highly reliable displacement measurement resistant to axial rotation interference. The annular brush adopts a wave structure with elastic deformation capability, maintaining sliding contact with the circuit board. At the same time, an improved neural network is used to tune the linear active disturbance rejection controller online. The total system disturbance is estimated and compensated in real time through a linearly extended state observer. The exponential variable step size learning strategy and the online pruning mechanism based on connection saliency are used to keep the computation lightweight, achieving adaptive online adjustment of controller bandwidth, observer bandwidth and control gain, and maintaining optimal control performance under load or operating condition changes.

Description

Technical Field

[0001] This invention belongs to the field of pneumatic technology, specifically relating to a pneumatic actuator with an embedded displacement sensing unit and its intelligent position control system. Background Technology

[0002] Against the backdrop of the in-depth advancement of intelligent manufacturing and industrial automation, pneumatic actuators, as the core power unit in linear drive scenarios, have become a key factor restricting the precise positioning, intelligent control, and operational reliability of equipment due to the real-time and accurate displacement feedback. Currently, although external displacement detection schemes based on gratings, magnetic gratings, or Hall effect principles are widely used, their "sensing-actuation" separation architecture has caused inherent contradictions in spatial integration, environmental tolerance, and signal links, which are difficult to eradicate through local improvements.

[0003] External sensors must be attached to the outside of the cylinder using brackets, reading heads, and cables, resulting in unnecessary expansion of the actuator's external dimensions, directly conflicting with the trend towards compact equipment. The sensing interface is exposed to dust, moisture, and vibration, making it difficult to balance protection and accuracy.

[0004] More importantly, the irregular circumferential rotation that accompanies the reciprocating motion of the pneumatic actuator piston rod will disrupt the precise relative posture required by the external sensor. In non-contact solutions, air gap deviation will introduce measurement errors, while in contact solutions, trajectory offset will exacerbate local wear.

[0005] Furthermore, the signal path from the moving end to the controller traverses multiple mechanical connections and exposed cables, continuously introducing impedance changes and noise, which impairs the long-term stability of the feedback. For example, Chinese Patent No. CN 222760024 U discloses a pneumatic actuator control system and related pneumatic equipment. Although this system improves response performance through high-frequency valve groups, it still relies on external position sensors at the feedback level. It fails to fundamentally overcome inherent limitations such as space redundancy and high environmental sensitivity, and there is still room for improvement in closed-loop control accuracy and system robustness under complex working conditions. Summary of the Invention

[0006] To address the problems of existing external displacement sensors being susceptible to environmental interference and exhibiting poor detection stability, as well as the difficulty in achieving high-precision position tracking using traditional control methods due to the strong nonlinearity of pneumatic systems, this application proposes a pneumatic actuator with an embedded displacement sensing unit and its intelligent position control system. This system integrates displacement detection functionality within the pneumatic actuator, achieving compact and interference-resistant displacement measurement through a distributed contact method between a ring brush and a fixed resistive element module, allowing circumferential rotation of the piston rod at any position. Simultaneously, an improved neural network-based online-tuned linear active disturbance rejection controller is employed. This controller uses a linearly extended state observer to estimate and compensate for the total system disturbance in real time, enabling adaptive online adjustment of controller parameters and maintaining optimal control performance under varying loads or operating conditions. The specific technical solution is as follows:

[0007] A pneumatic actuator with an embedded displacement sensing unit includes a pneumatic actuator module, a fixed resistive element module, a sliding contact module, and a signal processing module.

[0008] The pneumatic actuator module includes a cylinder, a pneumatic actuator front cover, and a pneumatic actuator rear cover connected in sequence. A piston rod is slidably mounted through the pneumatic actuator front cover, and a piston is slidably mounted inside the cylinder. The piston is fixedly mounted on the piston rod and divides the inside of the cylinder into a rod chamber and a rodless chamber. The pneumatic actuator front cover has a vent hole for the rod chamber that communicates with the rod chamber, and the pneumatic actuator rear cover has a vent hole for the rodless chamber that communicates with the rodless chamber. A central chamber is provided inside the piston rod.

[0009] The signal processing module is installed on the rear end cover of the pneumatic actuator;

[0010] The fixed resistor module includes a guide rod fixedly connected to the rear end cover of the pneumatic actuator, and circuit boards A and B mounted on the guide rod; the surface of circuit board A is provided with a resistive film, and the surface of circuit board B is provided with a conductive slider.

[0011] The sliding contact module includes a brush core tube and an annular brush. The brush core tube is disposed in the central cavity of the piston rod and sleeved on the outside of the guide rod. The two ends of the brush core tube are connected to the piston rod and the annular brush, respectively. The annular brush adopts a wave structure, has elastic deformation capability, is sleeved on the outside of the guide rod, and maintains sliding contact with circuit boards A and B. The wave structure of the annular brush has at least one contact point with each of circuit boards A and B to form a circuit.

[0012] Furthermore, a piston rod bushing disposed inside the cylinder is fixedly installed on the rear end cover of the pneumatic actuator, the piston rod is slidably fitted on the piston rod bushing, and the brush core tube is slidably installed inside the piston rod bushing.

[0013] Furthermore, the piston rod wall is provided with a piston rod vent hole that connects to the outside atmosphere and the central chamber; the brush core tube wall is provided with a brush core tube vent hole that connects the internal chamber of the brush core tube and the central chamber of the piston rod.

[0014] Furthermore, two mounting grooves are respectively provided on the end faces of the guide rod, which are symmetrically arranged about the center of the rod axis; the circuit board A and the circuit board B are respectively arranged on both radial sides of the guide rod, and the two ends of each circuit board are respectively embedded and fixed in the corresponding mounting grooves at both ends of the guide rod.

[0015] Furthermore, the brush core tube and the piston rod are connected by a ball joint.

[0016] Furthermore, the signal processing module includes a signal processing circuit board, which comprises a power supply module and a microprocessor module; the power supply module provides operating voltage to each functional unit on the signal processing circuit board; the microprocessor module integrates a signal conditioning unit, an analog-to-digital converter (ADC), and a storage unit.

[0017] The signal conditioning unit performs high-impedance buffering and filtering on the analog signal input from the displacement feedback module; the analog-to-digital converter (ADC) is connected to the signal conditioning unit and is used to convert the conditioned analog voltage signal into the original displacement digital quantity; the storage unit pre-stores calibration parameters for the linearity deviation and assembly error of the resistive film; the microprocessor module calls the calibration parameters in the storage unit to perform real-time linearization correction on the original displacement digital quantity, generates displacement feedback data, and outputs it.

[0018] Furthermore, the signal processing circuit board is fixedly installed inside the rear end cover of the pneumatic actuator by fastening bolts, and a protective end cover is provided on the outer side of the signal processing circuit board;

[0019] The guide rod has a through hole along its axial direction at its center; the pneumatic actuator has a positive terminal hole, a negative terminal hole, a signal terminal hole, and a signal output hole on its rear end cover; the potential input terminal at the right end of circuit board A is electrically connected to the positive terminal of the power supply module through the positive terminal hole, and the potential input terminal at its left end passes through the guide rod through hole and is connected to the negative terminal of the power supply module through the negative terminal hole; circuit board B is electrically connected to the signal processing module through the signal terminal hole.

[0020] The position intelligent control system includes an air supply module, a motion control module, and a pneumatic actuator with an embedded displacement sensing unit;

[0021] The air supply module includes an air source, a pressure reducing valve, a pressure stabilizing air tank, and a proportional directional control valve; the input end of the pressure reducing valve is connected to the air source, and the output end is connected to the pressure stabilizing air tank; the output end of the pressure stabilizing air tank is connected to the air inlet of the proportional directional control valve; the two working ports of the proportional directional control valve are respectively connected to the rod-side chamber and rodless chamber of the pneumatic actuator with the embedded displacement sensing unit through air pipelines.

[0022] The motion control module is signal-connected to the pneumatic actuator of the embedded displacement sensing unit and the air supply module, and performs the following control steps:

[0023] S1. State Space Reconstruction: Real-time velocity and acceleration are extracted using a tracking differentiator to construct a third-order kinematic state space;

[0024] S2. Nonlinear feature vector extraction: Calculate the real-time position error and error change rate, and combine them with the real-time effective volume of the two chambers of the pneumatic actuator to construct a feature vector containing position error, effective volume of the two chambers and acceleration.

[0025] S3. Online Tuning and Control: The feature vector is input into the improved neural network controller to tune the three parameters of the linear active disturbance rejection controller in real time: controller bandwidth, observer bandwidth, and control gain. The linear active disturbance rejection controller uses the three parameters to estimate the total system disturbance in real time through the linear extended state observer, realizes feedforward compensation of aerodynamic disturbance, and generates control quantity to drive the proportional directional control valve.

[0026] Furthermore, in the control steps executed by the motion control module, the improved neural network controller employs an exponential variable step size learning strategy and online pruning topology optimization based on connection saliency, specifically including:

[0027] Learning rate update formula:

[0028]

[0029] in, for Learning rate at any given moment The initial learning rate, To minimize the learning rate, The attenuation coefficient is... This represents the current iteration number. As the base time, The decay index;

[0030] The controller monitors the saliency indicators of each neuron connection in real time. :

[0031]

[0032] in, For the first The input neuron and the first The connection weights between the output neurons The loss function;

[0033] When significance index If the value remains below the preset pruning threshold, the connection is determined to be a redundant connection and a pruning operation is performed, which means that the weight of the connection is forcibly cleared to zero through a mask matrix.

[0034] Furthermore, the motion control module controls the opening degree of the proportional directional control valve through the following logic:

[0035] The total system disturbance term is estimated in real time based on the linear extended state observer. Control gain and the linear error combination term consisting of position and velocity errors. Calculate the theoretical control quantity ;then the theoretical control quantity Mapped to a nonlinear flow output function with dead zone compensation, the output is zero when the absolute value of the theoretical control quantity is less than the dead zone threshold; when the absolute value of the theoretical control quantity is greater than the dead zone threshold, the output is linearly related to the theoretical control quantity; the final generated control voltage drives the valve core of the proportional directional control valve to move, thereby adjusting the air mass flow rate entering the rod chamber and the rodless chamber in real time by changing the effective flow area of ​​the valve port.

[0036] Because the present invention adopts the above-described technical solution, compared with the prior art, the present invention has the following beneficial effects:

[0037] 1. The pneumatic actuator with an embedded displacement unit provided by the present invention is based on a distributed contact displacement detection method using a built-in annular brush and a fixed resistive element module. The displacement sensing unit is integrated inside the pneumatic actuator. The annular brush uniformly covers the fixed resistive element module circumferentially and maintains sliding contact with the circuit board. The irregular circumferential rotation of the piston rod during reciprocating motion does not change the effective contact state between the annular brush and the circuit board, fundamentally eliminating measurement errors and contact wear caused by attitude deviation. It effectively eliminates the interference of piston circumferential rotation on the detection signal. The system has strong anti-environmental interference capability, compact structure, and significantly improved detection accuracy and long-term stability.

[0038] 2. The position intelligent control system provided by this invention, when performing position closed-loop control, is based on an improved neural network-tuned linear active disturbance rejection controller. It can extract nonlinear features such as displacement, velocity, acceleration, and effective volume of the two chambers of the pneumatic actuator in real time, adaptively adjust the controller parameters, and perform feedforward compensation for the total system disturbance. This enables high-precision continuous control of the proportional directional control valve and avoids the problem of control performance degradation of traditional pneumatic systems when the load or operating conditions change. Attached Figure Description

[0039] Figure 1 This is a diagram of the pneumatic position control system based on an embedded displacement sensor and a linear pneumatic actuator, according to the present invention.

[0040] Figure 2 This is a cross-sectional structural diagram of the pneumatic actuator with an embedded displacement sensing unit according to the present invention.

[0041] Figure 3 This is a schematic cross-sectional view of the fixed resistor module of the present invention.

[0042] Figure 4 This is a schematic diagram of the guide rod in the fixed resistor module of the present invention.

[0043] Figure 5 This is a cross-sectional structural diagram of the sliding contact module of the present invention.

[0044] Figure 6 This is a schematic diagram of the annular brush structure in the sliding contact module of the present invention.

[0045] Figure 7 This is a schematic cross-sectional view of the assembly of the piston rod, fixed resistor module, and sliding contact module of the present invention.

[0046] Figure 8 This is a schematic cross-sectional view of the assembly of the fixed resistor module, the sliding contact module, and the pneumatic actuator rear end cover of the present invention.

[0047] Figure 9 This is a schematic diagram of the internal circuit connection of the pneumatic actuator with embedded displacement sensing unit of the present invention.

[0048] Figure 10 This is the logic block diagram of the intelligent position control system based on improved neural network-assisted linear active disturbance rejection control of the present invention.

[0049] Icon labels:

[0050] 1- Pneumatic actuator with embedded displacement sensing unit; 11- Pneumatic actuator module; 111- Cylinder; 112- Front cover of pneumatic actuator; 1121- Vent hole of rod chamber; 113- Rear cover of pneumatic actuator; 1131- Vent hole of rodless chamber; 1132- Signal output hole; 1133- Sealing ring; 1134- Positive terminal hole; 1135- Negative terminal hole; 1136- Signal terminal hole; 114- Piston; 115- Piston rod; 1151- Filter screen; 1152- Piston rod vent hole; 116- Rear cover of piston rod; 117- Piston rod bushing; 12 - Signal processing module; 121- Signal processing circuit board; 1211- Power supply module; 1212- Microprocessor module; 12121- Signal conditioning unit; 12122- Analog-to-digital converter (ADC); 12123- Storage unit; 122- Protective end cap; 123- Fastening bolt; 13- Fixed resistor module; 131- Circuit board A; 132- Circuit board B; 133- Guide rod; 1331- Guide rod through hole; 1332- Mounting slot; 14- Sliding contact module; 141- Brush core tube; 1411- Brush core tube vent hole; 142- Ring brush;

[0051] 2-Gas supply module; 21-Gas source; 22-Pressure reducing valve; 23-Pressure stabilizing gas tank; 24-Proportional directional control valve;

[0052] 3-Motion control module; 31-Industrial computer; 32-Data acquisition card. Detailed Implementation

[0053] The technical solution of the present invention will be further described in detail below through embodiments and in conjunction with the accompanying drawings. Many specific details are set forth in the following description to provide a thorough understanding of the present invention; however, the present invention can be practiced in many other ways different from those described herein, and those skilled in the art can make similar modifications without departing from the spirit of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0054] In the description of this invention, it should be noted that the terms "upper", "lower", "in", "out", "front", "rear", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship in which the product of this invention is usually placed when in use. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limiting this invention.

[0055] Example:

[0056] Pneumatic actuators with embedded displacement sensing units and their intelligent position control systems, such as Figure 1As shown, it includes a pneumatic actuator 1 with an embedded displacement sensing unit, an air supply module 2, and a motion control module 3.

[0057] The pneumatic actuator 1 with an embedded displacement sensing unit includes a pneumatic actuator module 11, a displacement sensing module, and a signal processing module 12.

[0058] like Figure 2 As shown, the pneumatic actuator module 11 includes a pneumatic actuator front cover 112, a cylinder 111, and a pneumatic actuator rear cover 113 connected in sequence. A piston rod 115 is slidably mounted through the pneumatic actuator front cover 112, and a piston 114 is slidably mounted inside the cylinder 111. The two ends of the cylinder 111 are sealed to the pneumatic actuator front cover 112 and the pneumatic actuator rear cover 113, respectively, forming a sealed piston movement chamber inside. The piston 114 is fixedly mounted on the piston rod 115, and together with the piston rod 115, constitutes a piston assembly. The piston 114 divides the interior of the cylinder 111 into a rod chamber and a rodless chamber. The pneumatic actuator front cover 112 has a rod chamber vent hole 1121 communicating with the rod chamber, and the pneumatic actuator rear cover 113 has a rodless chamber vent hole 1131 communicating with the rodless chamber, used to realize the normal reciprocating motion of the pneumatic actuator. The piston rod 115 has a central chamber inside.

[0059] In one specific embodiment of this invention, a piston rod bushing 117 (e.g., bolted) is fixedly mounted on the rear end cover 113 of the pneumatic actuator and is disposed inside the cylinder 111. This bushing supports the axial movement of the piston rod 115, ensuring the straightness of the piston rod 115's movement. The piston rod 115 is slidably fitted onto the piston rod bushing 117, and a piston rod rear end cover 116 and a seal are provided at the connection between the piston rod 115 and the piston rod bushing 117 to prevent compressed air leakage and ensure the working air pressure of the pneumatic actuator.

[0060] The piston rod 115 has a piston rod vent 1152 on its rod wall, which connects to the outside atmosphere and the central chamber. A filter screen 1151 is arranged at the piston rod vent 1152 to balance the internal and external air pressure of the piston rod 115 and filter air impurities entering the inner cavity of the piston rod 115, so as to prevent dust and other pollutants from adhering to the contact surface of the resistive film and the brush, and to ensure the accuracy and reliability of displacement detection.

[0061] The displacement sensing module is located inside the pneumatic actuator module 11, and the signal processing module 12 is installed on the rear cover 113 of the pneumatic actuator.

[0062] like Figures 3 to 8 As shown, the displacement sensing module includes a fixed resistive element module 13 and a sliding contact module 14.

[0063] The fixed resistor module 13 includes circuit board A131, circuit board B132, and guide rod 133. The guide rod 133 is fixedly connected to the rear end cover 113 of the pneumatic actuator (e.g., via a threaded connection). It is an axially extending cylindrical rod with two mounting grooves 1332 symmetrically arranged about the center of the rod's axis on its two end faces. Circuit boards A131 and B132 are respectively arranged on both radial sides of the guide rod 133, with each circuit board's ends embedded and fixed in the corresponding mounting grooves 1332 at both ends of the guide rod 133. A guide rod through hole 1331 is formed at the center of the guide rod 133 along its axial direction for threading electrical connection wires. A resistive film is provided on the surface of circuit board A131, and a conductive slider is provided on the surface of circuit board B132.

[0064] The sliding contact module 14 includes a brush core tube 141 and an annular brush 142. The brush core tube 141 is made of polyoxymethylene insulating material and is a hollow tubular structure extending axially. It is disposed in the central cavity of the piston rod 115 and slidably installed in the piston rod bushing 117. It is also sleeved outside the guide rod 133 of the fixed resistor module 13 and arranged coaxially with the guide rod 133. The two ends of the brush core tube 141 are respectively connected to the piston rod 115 and the annular brush 142. Specifically, one end of the brush core tube 141 is connected to the piston rod 115 through a ball joint, and the other end is connected to the annular brush 142 through a thread.

[0065] As a specific embodiment of the ball joint connection, a ball joint seat is provided inside the left end of the piston rod 115, and a ball head is embedded and fixed inside the ball joint seat. The ball head is fixedly connected to the left end of the brush core tube 141. The ball joint seat and the ball head form a spherical pair connection, which allows the brush core tube 141 to undergo multi-degree-of-freedom adaptive angular deflection relative to the piston rod 115. When the piston rod 115 rotates axially or wobbles radially, the ball joint structure, through the adaptive rotation of the spherical pair, ensures that the brush core tube 141 always maintains coaxiality with the guide rod 133, thereby ensuring that the contact pressure between the annular brush 142 and the circuit boards A131 and B132 is uniform and stable.

[0066] The brush core tube 141 has a vent hole 1411 on its wall, which connects the internal chamber of the brush core tube 141 and the central chamber of the piston rod 115, and is used to balance the air pressure inside and outside the brush core tube 141.

[0067] The annular brush 142 adopts a wave structure, which has elastic deformation capability. It is preferably made of an elastic conductive material and is sleeved on the outside of the guide rod 133, maintaining sliding contact with circuit boards A131 and B132. The wave structure can generate elastic deformation when subjected to radial pressure, so that the annular brush 142 maintains a stable contact pressure with circuit boards A131 and B132, forming multiple independent elastic contact points.

[0068] The annular brush 142 forms a circuit through a distributed contact method, continuously converting the motion of the piston 114 into an electrical signal. The distributed contact method means that there is at least one contact point on the wave structure of the annular brush 142 with both circuit boards A131 and B132 to form a circuit. Even if the piston rod 115 rotates axially, the wave structure can still adaptively adjust the positional relationship of each contact point through elastic deformation, ensuring the continuous conduction of the signal detection circuit and ensuring the reliability of displacement detection.

[0069] When piston 114 drives piston rod 115 to move axially, brush core tube 141 adaptively adjusts its angle through ball joint, ensuring that annular brush 142 maintains stable multi-point sliding contact with circuit boards A131 and B132. Even if piston rod 115 rotates slightly axially, annular brush 142 can still maintain reliable contact with circuit boards A131 and B132 due to the elastic deformation of its wave structure, thereby eliminating interference from piston axial rotation on displacement detection signals.

[0070] like Figures 8-10 As shown, the signal processing module 12 includes a signal processing circuit board 121, a protective end cover 122, and fastening bolts 123. The signal processing circuit board 121 is fixedly installed inside the rear end cover 113 of the pneumatic actuator by the fastening bolts 123. The protective end cover 122 covers the outside of the signal processing circuit board 121, isolating the signal processing circuit board 121 from the outside world, reducing the impact of external dust, moisture, etc. on the signal processing circuit board 121, and ensuring the long-term stable operation of the signal processing module 12.

[0071] The pneumatic actuator's rear end cover 113 has a positive terminal hole 1134, a negative terminal hole 1135, and a signal terminal hole 1136. The signal processing circuit board 121 is electrically connected to the displacement sensing module through these terminals. Each terminal hole is filled with a polymer insulating matrix, and a dense encapsulation layer that fits tightly against the hole wall is formed through an overall potting process. This ensures electrical insulation performance and enhances the structural strength and sealing reliability of the terminal holes.

[0072] The pneumatic actuator rear end cover 113 is also provided with a signal output hole 1132, and a sealing ring 1133 is provided at the signal output hole 1132 to ensure the airtightness of the pneumatic actuator cavity.

[0073] The signal processing circuit board 121 includes a power supply module 1211 and a microprocessor module 1212. The power supply module 1211 provides operating voltage to the various functional units on the signal processing circuit board 121. The microprocessor module 1212 integrates a signal conditioning unit 12121, an analog-to-digital converter (ADC) 12122, and a storage unit 12123. Specifically, the signal conditioning unit 12121 performs high-impedance buffering and filtering on the analog signal input from the displacement feedback module to eliminate load effects and suppress high-frequency noise interference, ensuring the fidelity of the original displacement signal. The ADC 12122 is connected to the signal conditioning unit 12121 and converts the conditioned analog voltage signal into the original digital displacement value. The storage unit 12123 pre-stores calibration parameters for resistive film linearity deviation and assembly errors. The microprocessor module 1212 calls the calibration parameters in the storage unit 12123 to perform real-time linearization correction on the original digital displacement value, generating high-precision displacement feedback data and outputting it to the motion control module 3.

[0074] Reference Figure 9 As shown, the circuit connection between the displacement sensing module and the signal processing circuit board 121 is as follows: The potential input terminal at the right end of circuit board A131 is electrically connected to the positive terminal of the power supply module 1211 from point A1 through the positive terminal connection hole 1134. The potential input terminal at the left end passes through the guide rod through hole 1331 from point A3 and is connected to the negative terminal of the power supply module 1211 through the negative terminal connection hole 1135, thereby forming a uniform potential distribution on the resistive film of circuit board A131. Circuit board B132 is electrically connected to the signal processing module 12 from point B1 through the signal connection hole 1136. When the piston 114 moves, it drives the annular brush 142 to slide along circuit board A131 and circuit board B132. The annular brush 142 connects the sampling point A2 on circuit board A131 and the acquisition point B2 on circuit board B132 in real time, introducing the position potential into circuit board B132.

[0075] like Figure 1 As shown, the air supply module 2 includes an air source 21, a pressure reducing valve 22, a pressure stabilizing gas tank 23, and a proportional directional control valve 24. The input end of the pressure reducing valve 22 is connected to the air source 21, and the output end is connected to the pressure stabilizing gas tank 23; the output end of the pressure stabilizing gas tank 23 is connected to the air inlet of the proportional directional control valve 24; the two working ports of the proportional directional control valve 24 are respectively connected to the rod-side and rodless-side chambers of the pneumatic actuator 1 with an embedded displacement sensing unit through air pipelines. The pressure reducing valve 22 is used to coarsely adjust the system working pressure, and the pressure stabilizing gas tank 23 is used to buffer air source pressure fluctuations to ensure stable air pressure entering the proportional directional control valve 24.

[0076] The motion control module 3 includes an industrial computer 31 and a data acquisition card 32. The industrial computer 31 acquires the displacement signal output by the pneumatic actuator 1 with the embedded displacement sensing unit through the data acquisition card 32, and directly outputs an analog voltage signal through the output channel of the data acquisition card 32 to control the opening degree of the proportional directional control valve 24 according to the control algorithm, thereby realizing closed-loop control of the position of the pneumatic actuator 1 with the embedded displacement sensing unit.

[0077] like Figure 10 As shown, motion control module 3 executes the following control steps:

[0078] S1, State-space reconstruction: Reconstructing the displacement electrical signal Mapped to physical displacement And use a tracking differentiator to extract real-time velocity. and acceleration Construct a third-order kinematic state space:

[0079]

[0080] in, These correspond to displacement, velocity, and acceleration, respectively. To control the quantity, This is an external disturbance.

[0081] S2. Nonlinear Feature Vector Extraction: Calculating Real-Time Position Error and error change rate And combined with the real-time effective volume of the rod-side and rodless chambers of the pneumatic actuator and Construct feature vectors ,in, For the target location, , These are the effective piston areas of the rod chamber and the rodless chamber, respectively. , These are the equivalent lengths corresponding to the initial volumes of the rod-shaped cavity and the rodless cavity, respectively. This represents the maximum stroke of the piston.

[0082] S3. Online Tuning and Control: The vector... Inputting an improved neural network controller, the three parameters of the linear active disturbance rejection controller (LADRC) are tuned in real time online. :

[0083]

[0084] in, These are the controller bandwidth, observer bandwidth, and control gain, respectively. For activation function, These are the connection weights from the input layer to the hidden layer. For feature vectors The One portion, This is the controller bandwidth output bias; This is the ratio of the observer bandwidth to the controller bandwidth. The connection weights from the hidden layer to the control gain output layer are... To control the output bias of the gain.

[0085] In the above steps, the tracking differentiator effectively suppresses the influence of measurement noise on speed and acceleration estimation through nonlinear tracking and differential extraction; the real-time effective volume of the two chambers changes continuously with the piston position, and its introduction enables the feature vector to characterize the pressure-volume dynamic coupling characteristics of the pneumatic actuator; the improved neural network controller adopts online learning and pruning strategies to maintain computational lightweight while tuning LADRC parameters, thereby achieving high-frequency, high-precision real-time adjustment of the proportional directional control valve 24.

[0086] Specifically, in this embodiment, the improved neural network controller employs an exponential variable step size learning strategy and online pruning topology optimization based on connection saliency, specifically including:

[0087] (1) Learning rate update formula:

[0088]

[0089] in, for Learning rate at any given moment The initial learning rate, To minimize the learning rate, The attenuation coefficient is... This represents the current iteration number. As the base time, This is the decay index.

[0090] (2) The controller monitors the saliency indicators of each neuron connection in real time. :

[0091]

[0092] in, For the first The input neuron and the first The connection weights between the output neurons This is the loss function.

[0093] The learning rate update formula follows an exponential decay form, gradually decreasing from its initial value to its minimum value, with the decay rate determined by a decay coefficient. The controller monitors the saliency index of each neuron connection in real time. This index is defined as the sum of the products of the absolute value of the connection weight and the gradient of its corresponding input signal. When the saliency index consistently falls below a preset pruning threshold, the connection is identified as redundant and pruning is performed, forcibly clearing the connection weight to zero using a mask matrix. The pruned network structure is sparse, reducing computational complexity and thus shortening the control cycle of the motion control module.

[0094] The motion control module 3 controls the opening degree of the proportional directional control valve 24 through the following logic:

[0095] The motion control module estimates the total system disturbance term in real time based on the Linear Extended State Observer (LESO). Control gain and the linear error combination term consisting of position and velocity errors. Calculate control quantity ;then the theoretical control quantity The output is mapped to a nonlinear flow rate function with dead-zone compensation. This function includes a dead-zone interval; when the absolute value of the theoretical control quantity is less than the dead-zone threshold, the output is zero to overcome the dead-zone characteristic of the proportional directional control valve 24; when the absolute value of the theoretical control quantity is greater than the dead-zone threshold, the output is linearly related to the theoretical control quantity. The resulting control voltage drives the valve core of the proportional directional control valve 24 to actuate, thereby adjusting the air mass flow rate entering the rod-side and rodless-side chambers in real time by changing the effective flow area of ​​the valve orifice.

[0096] The working process of the location intelligent control system is as follows:

[0097] During system operation, the air supply module 2 provides stable air pressure to the pneumatic actuator 1 with an embedded displacement sensing unit. The piston 114 within the pneumatic actuator 1 moves, causing the annular brush 142 to slide along the resistive film on circuit board A131 and the conductive slider on circuit board B132. The resulting displacement electrical signal is conditioned, converted from analog to digital, and linearized by the signal processing module 12 before being sent to the motion control module 3. The motion control module 3 executes control steps S1 to S3, calculating and outputting control voltage to the proportional directional control valve 24 in real time. This precisely adjusts the intake and exhaust volumes of the rod-side and rodless sides of the pneumatic actuator, driving the piston 114 to accurately track the target position. Throughout the control process, the improved neural network controller continuously performs online tuning and pruning optimization, enabling the system to adapt to disturbances such as load changes, air supply pressure fluctuations, and frictional changes, maintaining high-precision closed-loop position control at all times.

Claims

1. Pneumatic actuator (1) with embedded displacement sensing unit, characterized in that It includes a pneumatic actuator module (11), a fixed resistive element module (13), a sliding contact module (14), and a signal processing module (12). The pneumatic actuator module (11) includes a cylinder (111), a pneumatic actuator front end cover (112), and a pneumatic actuator rear end cover (113) connected in sequence. A piston rod (115) is slidably installed through the pneumatic actuator front end cover (112), and a piston (114) is slidably installed inside the cylinder (111). The piston (114) is fixedly installed on the piston rod (115) and divides the inside of the cylinder (111) into a rod chamber and a rodless chamber. The pneumatic actuator front end cover (112) is provided with a rod chamber vent (1121) communicating with the rod chamber, and the pneumatic actuator rear end cover (113) is provided with a rodless chamber vent (1131) communicating with the rodless chamber. A central chamber is provided inside the piston rod (115). The signal processing module (12) is installed on the rear end cover (113) of the pneumatic actuator; The fixed resistor module (13) includes a guide rod (133) fixedly connected to the rear end cover (113) of the pneumatic actuator, and a circuit board A (131) and a circuit board B (132) mounted on the guide rod (133); the surface of the circuit board A (131) is provided with a resistive film, and the surface of the circuit board B (132) is provided with a conductive slider. The sliding contact module (14) includes a brush core tube (141) and an annular brush (142); the brush core tube (141) is disposed in the central cavity of the piston rod (115) and sleeved on the outside of the guide rod (133); the two ends of the brush core tube (141) are respectively connected to the piston rod (115) and the annular brush (142); the annular brush (142) adopts a wave structure, has elastic deformation capability, is sleeved on the outside of the guide rod (133) and maintains sliding contact with circuit board A (131) and circuit board B (132); the wave structure of the annular brush (142) has at least one contact point with each of circuit board A (131) and circuit board B (132) to form a circuit.

2. The pneumatic actuator with embedded displacement sensing unit according to claim 1, characterized in that, The pneumatic actuator has a piston rod bushing (117) fixedly installed on the rear end cover (113) and disposed inside the cylinder (111). The piston rod (115) is slidably fitted on the piston rod bushing (117) and the brush core tube (141) is slidably installed inside the piston rod bushing (117).

3. The pneumatic actuator with embedded displacement sensing unit of claim 1, wherein, The piston rod (115) has a piston rod vent (1152) on its rod wall that connects to the outside atmosphere and the central chamber; the brush core tube (141) has a brush core tube vent (1411) on its tube wall that connects to the internal chamber of the brush core tube (141) and the central chamber of the piston rod (115).

4. The pneumatic actuator with embedded displacement sensing unit of claim 1, wherein, The guide rod (133) has two mounting slots (1332) arranged symmetrically about the center of the rod axis on its two end faces; the circuit board A (131) and the circuit board B (132) are respectively arranged on the radial sides of the guide rod (133), and the two ends of each circuit board are respectively embedded and fixed in the mounting slots (1332) at the two ends of the guide rod (133).

5. The pneumatic actuator with an embedded displacement sensing unit according to claim 1, characterized in that, The brush core tube (141) is ball-jointed with the piston rod (115).

6. The pneumatic actuator with embedded displacement sensing unit of claim 1, wherein, The signal processing module (12) includes a signal processing circuit board (121), which includes a power supply module (1211) and a microprocessor module (1212). The power supply module (1211) provides operating voltage to each functional unit on the signal processing circuit board (121). The microprocessor module (1212) integrates a signal conditioning unit (12121), an analog-to-digital converter (ADC) (12122), and a storage unit (12123). The signal conditioning unit (12121) performs high-impedance buffering and filtering on the analog signal input from the displacement feedback module; the analog-to-digital converter (ADC) (12122) is connected to the signal conditioning unit (12121) and is used to convert the conditioned analog voltage signal into the original displacement digital quantity; the storage unit (12123) pre-stores calibration parameters for the linearity deviation and assembly error of the resistive film; the microprocessor module (1212) calls the calibration parameters in the storage unit (12123) to perform real-time linearization correction on the original displacement digital quantity, generate displacement feedback data and output it.

7. The pneumatic actuator with an embedded displacement sensing unit according to claim 6, characterized in that, The signal processing circuit board (121) is fixedly installed inside the rear end cover (113) of the pneumatic actuator by fastening bolts (123), and a protective end cover (122) is provided on the outside of the signal processing circuit board (121). The guide rod (133) has a guide rod through hole (1331) at its center along the axial direction; the pneumatic actuator rear end cover (113) has a positive terminal connection hole (1134), a negative terminal connection hole (1135), a signal connection hole (1136) and a signal output hole (1132); the potential input terminal at the right end of the circuit board A (131) is electrically connected to the positive terminal of the power supply module (1211) through the positive terminal connection hole (1134), and the potential input terminal at its left end passes through the guide rod through hole (1331) and is connected to the negative terminal of the power supply module (1211) through the negative terminal connection hole (1135); the circuit board B (132) is electrically connected to the signal processing module (12) through the signal connection hole (1136).

8. A position intelligence control system characterized by, It includes an air supply module (2), a motion control module (3), and a pneumatic actuator (1) with an embedded displacement sensing unit as described in claim 1. The air supply module (2) includes an air source (21), a pressure reducing valve (22), a pressure stabilizing gas tank (23), and a proportional directional control valve (24); the input end of the pressure reducing valve (22) is connected to the air source (21), and the output end is connected to the pressure stabilizing gas tank (23); the output end of the pressure stabilizing gas tank (23) is connected to the air inlet of the proportional directional control valve (24); the two working ports of the proportional directional control valve (24) are respectively connected to the rod chamber and rodless chamber of the pneumatic actuator (1) with embedded displacement sensing unit through air pipelines; The motion control module (3) is signal-connected to the pneumatic actuator (1) of the embedded displacement sensing unit and the air supply module (2), and performs the following control steps: S1. State Space Reconstruction: Real-time velocity and acceleration are extracted using a tracking differentiator to construct a third-order kinematic state space; S2. Nonlinear feature vector extraction: Calculate the real-time position error and error change rate, and combine them with the real-time effective volume of the two chambers of the pneumatic actuator to construct a feature vector containing position error, effective volume of the two chambers and acceleration. S3. Online tuning and control: The feature vector is input into the improved neural network controller to tune the three parameters of the linear active disturbance rejection controller in real time: controller bandwidth, observer bandwidth and control gain. The linear active disturbance rejection controller uses the three parameters to estimate the total disturbance of the system in real time through the linear extended state observer, realizes the feedforward compensation of aerodynamic disturbance, and generates the control quantity to drive the proportional directional control valve (24).

9. The position intelligence control system of claim 8, wherein, In the control steps executed by the motion control module (3), the improved neural network controller adopts an exponential variable step size learning strategy and online pruning topology optimization based on connection saliency, specifically including: Learning rate update formula: in, for Learning rate at any given moment The initial learning rate, To minimize the learning rate, The attenuation coefficient is... This represents the current iteration number. As the base time, The decay index; The controller monitors the significance indicator of each neuron connection in real time : in, For the first The input neuron and the first The connection weights between the output neurons The loss function; When significance index If the value remains below the preset pruning threshold, the connection is determined to be a redundant connection and a pruning operation is performed, which means that the weight of the connection is forcibly cleared to zero through a mask matrix.

10. The position intelligence control system of claim 9, wherein, The motion control module (3) controls the opening degree of the proportional directional control valve (24) through the following logic: The total system disturbance term is estimated in real time based on the linear extended state observer. Control gain and the linear error combination term consisting of position and velocity errors. Calculate the theoretical control quantity ;then the theoretical control quantity Mapped to a nonlinear flow output function with dead zone compensation, the output is zero when the absolute value of the theoretical control quantity is less than the dead zone threshold; when the absolute value of the theoretical control quantity is greater than the dead zone threshold, the output is linearly related to the theoretical control quantity; the final generated control voltage drives the valve core of the proportional directional control valve (24) to move, and adjusts the air mass flow rate entering the rod chamber and rodless chamber in real time by changing the effective flow area of ​​the valve port.

Images