Air pressure control method and air pressure control system applied to air pressure valve control equipment

Abstract

The embodiment of the invention discloses an air pressure control method and an air pressure control system applied to air pressure valve control equipment. A specific embodiment of the method comprises the following steps: acquiring target pose information for a target software driver; inputting the target pose information into an air pressure information generation model to obtain predicted air pressure information; the method comprises the following steps: performing multi-dimensional data acquisition on a target software driver through an attitude sensor and a laser radar to obtain attitude information and three-dimensional point cloud information; generating error pose information according to the target pose information, the pose information and the three-dimensional point cloud information; generating compensation air pressure information according to the error pose information; determining target air pressure information according to the compensated air pressure information and the predicted air pressure information; and according to the target air pressure information and the compensation air pressure information, an air pressure control instruction is generated, and the air pressure control instruction is sent to air pressure valve control equipment to control the target software driver to execute displacement operation. According to the embodiment, the error displacement operation of the target software driver is reduced.

Description

Technical Field

[0001] The embodiments disclosed herein relate to the field of fluid drive technology, and more specifically to a pneumatic control method and a pneumatic control system applied to pneumatic valve control equipment. Background Technology

[0002] Pneumatic systems are widely used across various industries. Simultaneously, with the advancement of artificial intelligence technology, the robotics industry has experienced rapid development. The technology for controlling and driving robots using pneumatic systems (e.g., regulating air pressure in the air circuit using a filter pressure reducing valve) has become increasingly mature. In particular, the robotics industry requires pneumatic controllers to drive soft actuators to complete various target actions. Currently, when operating soft actuators using pneumatic controllers, the common approach is to: construct a complex mechanical model based on numerous assumptions, predict the control air pressure using open-loop control, and then perform displacement operations on the soft actuator based on the predicted air pressure. Alternatively, pressure sensors can be installed in the air circuit at the front end of the soft actuator to detect the actual air pressure, and the difference between the predicted and detected air pressures can be used to perform closed-loop corrections to the control air pressure of the soft actuator for displacement operations.

[0003] However, when using the above method, the following technical problems often arise: Because the methods for constructing mechanical models are based on numerous assumptions, the process inevitably simplifies or ignores the influence of many factors, significantly reducing the effectiveness of theoretical models in practical applications (for example, making structural assumptions about a soft manipulator will cause the model to completely fail when the manipulator faces external disturbances). Furthermore, although pressure sensors measure the air pressure within the air chamber, they cannot determine whether this air pressure has generated the desired pose (due to the nonlinearity, hysteresis, creep, and environmental load variations of soft actuators, the same air pressure value can lead to completely different poses at different times and under different conditions). This results in numerous displacement operation errors by the soft actuators.

[0004] The information disclosed in this background section is only intended to enhance the understanding of the background of the inventive concept, and therefore may contain information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] The summary portion of this disclosure is intended to provide a brief overview of the concepts, which will be described in detail in the detailed description portion. This summary portion is not intended to identify key or essential features of the claimed technical solutions, nor is it intended to limit the scope of the claimed technical solutions.

[0006] Some embodiments of this disclosure provide a pneumatic control method, pneumatic control system, apparatus, electronic device, and computer-readable medium for use in pneumatic valve control equipment to solve the technical problems mentioned in the background section above.

[0007] In a first aspect, some embodiments of this disclosure provide a pneumatic control method applied to a pneumatic valve control device. The method includes: acquiring target pose information for a target soft actuator, wherein the target pose information represents the segmented displacement path of the target soft actuator; inputting the target pose information into a pre-trained pneumatic information generation model to obtain predicted pneumatic information, wherein the pneumatic information generation model represents the mapping relationship between pose information and pneumatic information; and performing multi-dimensional data acquisition on the target soft actuator using an attitude sensor and a lidar disposed at the end of the target soft actuator to obtain attitude information and... The system generates three-dimensional point cloud information, wherein the attitude sensor and the lidar data are time-synchronized; based on the target pose information, the attitude information, and the three-dimensional point cloud information, error pose information corresponding to the target software actuator is generated; based on the error pose information, compensation air pressure information corresponding to the target software actuator is generated; based on the compensation air pressure information and the predicted air pressure information, the target air pressure information is determined; based on the target air pressure information and the compensation air pressure information, an air pressure control command is generated, and the air pressure control command is sent to the air pressure valve control device to control the target software actuator to perform displacement operations.

[0008] Secondly, some embodiments of this disclosure provide a pneumatic control system applied to the method described in the first aspect. The system includes: a controller, configured to: run a control algorithm to generate pneumatic control commands based on target pose information and feedback actual pose information, to control the target software actuator to perform displacement operations; a pose sensor, configured to: acquire the pose information of the target software actuator; a lidar, configured to: acquire the three-dimensional spatial position information of the target software actuator; and a pneumatic valve control device, configured to: control the inflation and deflation of the target software actuator according to the pneumatic control commands from the controller. The pneumatic valve control device includes: a motor drive module, a positive pressure pump, a negative pressure pump, a digital-to-analog converter module, a positive pressure proportional valve, a negative pressure proportional valve, a pulse width modulation module, a positive pressure three-way solenoid valve, and a negative pressure three-way solenoid valve. The motor drive module is configured to: receive motor drive commands from the controller and control and drive the positive pressure pump and... The negative pressure pump operates as follows: the positive pressure pump is configured to output positive pressure gas to the target software driver; the negative pressure pump is configured to output negative pressure gas to the target software driver; the analog-to-digital converter is configured to convert the digital voltage signal from the controller into an analog voltage signal; the positive pressure proportional valve is configured to receive the analog voltage signal and adjust the positive pressure value output to the target software driver; and the negative pressure proportional valve is configured to receive the analog voltage signal and adjust the output to the target software driver. The negative air pressure value of the driver, the pulse width modulation module is configured to: receive the pulse width modulation parameters of the controller, and generate a square wave electrical signal with a corresponding duty cycle and frequency according to the pulse width modulation parameters, to drive the opening and closing of the positive air pressure three-way solenoid valve and the negative air pressure three-way solenoid valve, the positive air pressure three-way solenoid valve is configured to: control the opening and closing to switch the connection between the air circuit of the target software driver and the positive air pressure pump, the negative air pressure three-way solenoid valve is configured to: control the opening and closing to switch the connection between the air circuit of the target software driver and the negative air pressure pump.

[0009] Thirdly, some embodiments of this disclosure provide a pneumatic control device applied to a pneumatic valve control equipment. The device includes: an acquisition unit configured to acquire target pose information for a target software actuator, wherein the target pose information represents the segmented displacement path of the target software actuator; an input unit configured to input the target pose information into a pre-trained pneumatic information generation model to obtain predicted pneumatic information, wherein the pneumatic information generation model represents the mapping relationship between pose information and pneumatic information; and a data acquisition unit configured to perform multi-dimensional data acquisition on the target software actuator using an attitude sensor and a lidar disposed at the end of the target software actuator to obtain attitude information and three-dimensional point cloud information. The attitude sensor and the lidar acquire data in time. The first generation unit is configured to generate error pose information corresponding to the target software driver based on the target pose information, the attitude information, and the three-dimensional point cloud information. The second generation unit is configured to generate compensation air pressure information corresponding to the target software driver based on the error pose information. The determination unit is configured to determine the target air pressure information based on the compensation air pressure information and the predicted air pressure information. The third generation unit is configured to generate an air pressure control command based on the target air pressure information and the compensation air pressure information, and send the air pressure control command to the air pressure valve control device to control the target software driver to perform displacement operations.

[0010] Fourthly, some embodiments of this disclosure provide an electronic device, including: one or more processors; and a storage device having one or more programs stored thereon, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method described in any implementation of the first aspect above.

[0011] Fifthly, some embodiments of this disclosure provide a computer-readable medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the method described in any of the implementations of the first aspect above.

[0012] The various embodiments of this disclosure have the following beneficial effects: the pneumatic control method applied to a pneumatic valve control device according to some embodiments of this disclosure can reduce erroneous displacement operations of the soft actuator. Specifically, the reason for the large number of displacement operation errors of the soft actuator is that the method of constructing the mechanical model is based on a large number of assumptions. In the process of making assumptions, it is unavoidable to simplify or ignore the influence of many factors, which greatly reduces the effectiveness of the theoretical model in practical applications (for example, if structural assumptions are made about the soft actuator arm, the model will completely fail when the soft actuator arm faces external interference). Furthermore, although the pressure sensor measures the air pressure in the air chamber, it cannot know whether this air pressure has produced the desired pose (due to the nonlinearity, hysteresis, creep, and changes in environmental load of the soft actuator, the same air pressure value may lead to completely different poses at different times and under different states). This results in a large number of displacement operation errors of the soft actuator. Based on this, the pneumatic control method applied to a pneumatic valve control device according to some embodiments of this disclosure first obtains the target pose information for the target soft actuator. The target pose information represents the segmented displacement path of the target soft actuator. Therefore, target pose information representing the desired path of the target soft actuator can be obtained, and the corresponding segmented target action can be completed with the desired path as a reference. Then, the target pose information is input into a pre-trained air pressure information generation model to obtain predicted air pressure information. This air pressure information generation model represents the mapping relationship between pose information and air pressure information. Thus, based on the network model, predicted air pressure information corresponding to the displacement action of the target soft actuator can be obtained, and based on the predicted air pressure information, the target soft actuator can be subjected to extension and retraction displacement operations. Next, multi-dimensional data acquisition of the target soft actuator is performed using an attitude sensor and a lidar located at the end of the target soft actuator, obtaining attitude information and 3D point cloud information. The data acquired by the attitude sensor and the lidar are time-synchronized. Thus, attitude information representing the actual attitude of the target soft actuator and 3D point cloud information representing its actual position can be obtained. Next, based on the target pose information, the attitude information, and the 3D point cloud information, error pose information corresponding to the target soft actuator is generated. Thus, error pose information representing the error between the desired action and the actual action of the target soft actuator can be obtained. Then, based on the aforementioned error pose information, compensation pressure information corresponding to the target software actuator is generated. Thus, the error between the expected and actual actions of the target software actuator can be converted into compensation pressure information that can compensate for the action error, thereby correcting the target software actuator's actions. Afterwards, the target pressure information is determined based on the aforementioned compensation pressure information and the aforementioned predicted pressure information. Thus, the target pressure information characterizing the action error correction can be obtained.Finally, based on the target air pressure information and the compensated air pressure information, an air pressure control command is generated and sent to the air pressure valve control device to control the target software actuator to perform a displacement operation. Thus, the target software actuator can perform a displacement operation according to the desired path based on the correct air pressure control command, completing the target action. Furthermore, because air pressure prediction is performed using a pre-trained network model, the problem of traditional air pressure prediction models relying on numerous assumptions and resulting in low effectiveness in practical applications can be avoided. Moreover, because the pose information representing the desired path of the target software actuator and the actual pose information fed back from the attitude sensor and lidar can be used to determine and correct the pose error, erroneous displacement operations of the software actuator can be reduced. Attached Figure Description

[0013] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and elements are not necessarily drawn to scale.

[0014] Figure 1 This is a schematic diagram illustrating an application scenario of a pneumatic control method for a pneumatic valve control device, according to some embodiments of this disclosure. Figure 2 This is a flowchart of some embodiments of a pneumatic control method applied to a pneumatic valve control device according to the present disclosure; Figure 3 This is a schematic diagram of the rapid positive pressure regulation mode of the pneumatic valve control device disclosed herein; Figure 4 This is a schematic diagram of the rapid negative pressure adjustment mode of the pneumatic valve control device disclosed herein; Figure 5 This is a schematic diagram of the rapid positive and negative pressure mixed regulation mode of the pneumatic valve control device disclosed herein; Figure 6 This is a schematic diagram of the precise positive pressure regulation mode of the pneumatic valve control device disclosed herein; Figure 7 This is a schematic diagram of the precise negative pressure regulation mode of the pneumatic valve control device disclosed herein; Figure 8 This is a schematic diagram of the precise positive and negative pressure mixed regulation mode of the pneumatic valve control device disclosed herein; Figure 9 This is a schematic diagram of the structure of some embodiments of a pneumatic control device applied to a pneumatic valve control equipment according to the present disclosure; Figure 10This is a schematic diagram of the structure of an electronic device suitable for implementing some embodiments of the present disclosure. Detailed Implementation

[0015] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0016] It should also be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings. Unless otherwise specified, the embodiments and features described in this disclosure can be combined with each other.

[0017] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are used only to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependencies.

[0018] It should be noted that the terms "a" and "a plurality of" used in this disclosure are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0019] The names of messages or information exchanged between multiple devices in the embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

[0020] This disclosure will now be described in detail with reference to the accompanying drawings and embodiments.

[0021] Figure 1 This is a schematic diagram illustrating an application scenario of a pneumatic control method for a pneumatic valve control device, according to some embodiments of this disclosure.

[0022] exist Figure 1In the application scenario, firstly, the computing device 101 can acquire target pose information 102 for the target soft actuator. Then, the computing device 101 can input the target pose information 102 into a pre-trained air pressure information generation model 103 to obtain predicted air pressure information 104. Afterwards, the computing device 101 can perform multi-dimensional data acquisition on the target soft actuator using an attitude sensor (not shown in the figure) and a lidar (not shown in the figure) located at the end of the target soft actuator, obtaining attitude information 105 and three-dimensional point cloud information 106. Next, the computing device 101 can generate error pose information 107 corresponding to the target soft actuator based on the target pose information 102, the attitude information 105, and the three-dimensional point cloud information 106. Then, the computing device 101 can generate compensated air pressure information 108 corresponding to the target soft actuator based on the error pose information 107. Finally, the computing device 101 can determine the target air pressure information 109 based on the compensated air pressure information 108 and the predicted air pressure information 104. Finally, the computing device 101 can generate a pressure control command 110 based on the target pressure information 109 and the compensation pressure information 108, and send the pressure control command 110 to the pressure valve control device (not shown in the figure) to control the target software driver to perform displacement operation.

[0023] It should be noted that the aforementioned computing device 101 can be either hardware or software. When the computing device is hardware, it can be implemented as a distributed cluster composed of multiple servers or terminal devices, or as a single server or a single terminal device. When the computing device is software, it can be installed in the hardware devices listed above. It can be implemented as, for example, multiple software programs or software modules used to provide distributed services, or as a single software program or software module. No specific limitations are made here. It should be understood that... Figure 1 The number of computing devices in the system can be arbitrary, depending on the implementation requirements.

[0024] refer to Figure 2 The flowchart 200 illustrates some embodiments of a pneumatic control method for a pneumatic valve control device according to the present disclosure. The pneumatic control method for a pneumatic valve control device includes the following steps: Step 201: Obtain target pose information for the target software driver.

[0025] In some embodiments, the execution subject (e.g., computing device 101) of the pneumatic control method applied to a pneumatic valve control device can obtain target pose information for a target software driver from a local pose information database via a wired or wireless connection. The target software driver can be a software driver capable of sorting target items. The target items are not specifically limited here. For example, the target items may include, but are not limited to, electronic products and food. The target pose information can characterize the segmented displacement path of the target software driver. The segmented displacement path can be understood as the displacement path corresponding to a single action performed by the target software driver. Performing a single action can be understood as an action of extending or contracting. For example, the target pose information can be the pose information of the target software driver performing a clamping operation on the target item. The target pose information can include position information and attitude information. The position information can be the coordinate position information of the three-dimensional coordinates corresponding to the end of the target software driver. The attitude information can characterize the rotational orientation of the target software driver (or end). The local pose information database can be a database storing the target pose information. The aforementioned local pose information database is not specifically limited here. It should be noted that the aforementioned wireless connection methods may include, but are not limited to, 3G / 4G connections, WiFi connections, Bluetooth connections, WiMAX connections, Zigbee connections, UWB (ultra wideband) connections, and other currently known or future wireless connection methods.

[0026] As an example, the aforementioned execution entity can also receive target pose information generated in real time by a host computer via a wireless network. The user of the host computer can then generate the target pose information in real time by operating the corresponding gamepad, keyboard, mouse, and graphical interface. Here, the host computer can be a PC client.

[0027] Step 202: Input the target pose information into the pre-trained air pressure information generation model to obtain the predicted air pressure information.

[0028] In some embodiments, the execution entity can input the target pose information into a pre-trained air pressure information generation model to obtain predicted air pressure information. The air pressure information generation model represents the mapping relationship between pose information and air pressure information. The air pressure information generation model can be a network model that takes the target pose information as input and the predicted air pressure information as output. For example, the air pressure information generation model can be a GA-BP neural network model (GeneticAlgorithm Optimized Backpropagation Neural Network). The air pressure information generation model can include an input layer, hidden layers, and an output layer. The input layer can be a network layer that receives input data. The hidden layer can include shallow hidden layers and deep hidden layers. The shallow hidden layer can be a network layer that performs simple feature extraction on the input data. Here, simple features can be understood as basic, local motion features (such as instantaneous velocity and acceleration). For example, the shallow hidden layer can include a one-dimensional convolutional layer, an activation layer (ReLU activation function), a global average pooling layer, and a fully connected layer. The aforementioned deep network layers can be network layers that perform complex feature extraction on the input data. Here, complex features can be understood as intent policy features containing complete semantic context (combining continuous instantaneous velocity and acceleration features based on historical context to infer low-speed attitude adjustment at motion inflection points). For example, the aforementioned deep network layers may include: GRU (Gated Recurrent Unit) layers, fully connected layers, activation layers, and fully connected layers. The aforementioned output layer can be a network layer that outputs data after complex feature extraction. Specifically, the aforementioned input layer can receive target pose information as input. The aforementioned hidden layers can be network layers that perform simple and complex feature extraction on the input target pose information in a temporal sequence. The aforementioned output layer can be a network layer that outputs predicted air pressure information corresponding to the input target pose information. The aforementioned predicted air pressure information can be air pressure values ​​that enable the target software actuator to reach the desired path corresponding to the target pose information.

[0029] Optionally, before inputting the target pose information into the pre-trained air pressure information generation model to obtain the predicted air pressure information, the execution entity may also perform the following steps: The first step is to acquire a set of air pressure samples. These samples include sample pose information and corresponding air pressure information. The air pressure sample set is obtained by applying different air pressure combinations to the target software actuator to continuously change its shape, and by using training data measured by LiDAR and attitude sensors. The air pressure information can be the sample label corresponding to the sample pose information. For example, the sample pose information (input features) can be: [x=0.1m, y=0.02m, z=0.15m, roll=5°, pitch=10°, yaw=0°]. x can represent the horizontal coordinate. y can represent the vertical coordinate. z can represent the vertical coordinate. roll can represent rotation around the x-axis. pitch can represent rotation around the y-axis. yaw can represent rotation around the z-axis. The sample air pressure information (label) can be: [air chamber 1=30kPa, air chamber 2=15kPa]. It should be noted that the entity that performs the training of the above-mentioned air pressure information generation model can be the aforementioned entity or other computing devices.

[0030] The second step involves performing the following training steps based on the air pressure sample set: The first sub-step involves inputting the pose information of at least one pressure sample from the pressure sample set into an initial pressure information generation model to obtain the pressure information corresponding to each pressure sample in the at least one pressure sample set. The initial pressure information generation model can be an initial neural network capable of obtaining pressure information based on pose information. This initial neural network can be a neural network to be trained. Specifically, the initial neural network can be a neural network including an input layer, hidden layers, and an output layer.

[0031] The second sub-step involves comparing the pressure information corresponding to each pressure sample in the at least one pressure sample with the corresponding sample pressure information. Here, the comparison can be a comparison of the magnitude of the pressure information corresponding to each pressure sample in the at least one pressure sample with the magnitude of the corresponding sample pressure information.

[0032] The third sub-step involves determining, based on the comparison results, whether the initial air pressure information generation model has achieved the preset optimization objective. Here, the optimization objective can refer to the loss function value of the initial air pressure information generation model being trained being less than or equal to a preset threshold. This preset threshold can be 0.001. The loss function corresponding to the aforementioned loss function value can be the mean squared error loss function.

[0033] The fourth sub-step is to determine the initial air pressure information generation model as the trained air pressure information generation model in response to the determination that the initial air pressure information generation model has achieved the above optimization objective.

[0034] Optionally, the steps for training the above-mentioned air pressure information generation model may further include: The fifth sub-step, in response to the determination that the initial air pressure information generation model has not achieved the above optimization objective, adjusts the network parameters of the initial air pressure information generation model, and uses unused air pressure samples to form an air pressure sample set. The adjusted initial air pressure information generation model is then used as the new initial air pressure information generation model, and the above training steps are executed again. As an example, the back propagation algorithm (BP algorithm) and gradient descent methods (such as mini-batch gradient descent) can be used to adjust the network parameters of the initial air pressure information generation model.

[0035] Step 203: Using the attitude sensor and lidar installed at the end of the target software driver, multi-dimensional data is collected from the target software driver to obtain attitude information and three-dimensional point cloud information.

[0036] In some embodiments, the aforementioned execution entity can acquire multi-dimensional data from the target software driver using an attitude sensor and a lidar mounted at the end of the target software driver, obtaining attitude information and three-dimensional point cloud information. The attitude sensor can be used to acquire the attitude information. The lidar can be used to acquire the three-dimensional point cloud information. The attitude information represents the rotational orientation of the end of the target software driver. The three-dimensional point cloud information represents the coordinate position of the end of the target software driver in three-dimensional space. The time of data acquisition by the attitude sensor and the lidar can be synchronized. Specifically, synchronization can be achieved through hardware or software. Here, the three-dimensional space can be understood as the specific operating environment of the target software driver in the real physical world, which can be detected and quantified by the lidar.

[0037] In practice, as an example, an attitude sensor or lidar can be designated as the master device (with an internal trigger generator, such as a dedicated clock chip or specific circuitry), and the other as the slave device. The master device generates periodic trigger pulse signals. After sensing and receiving the trigger pulse signals, the attitude sensor (master or slave device) rotates the target software driver to acquire attitude information. The lidar (slave or master device) acquires 3D position data from the target software driver, obtaining 3D point cloud information. As another example, the aforementioned execution entity can first acquire attitude information using an attitude sensor located at the end of the target software driver. Then, it can scan using a lidar located at the end of the target software driver to obtain 3D point cloud information. Finally, data acquisition time synchronization is achieved through a timestamp alignment algorithm (e.g., linear interpolation).

[0038] Step 204: Generate the error pose information of the corresponding target software actuator based on the target pose information, attitude information and 3D point cloud information.

[0039] In some embodiments, the execution entity can generate error pose information corresponding to the target software driver based on the target pose information, the attitude information, and the three-dimensional point cloud information. The error pose information can characterize the displacement operation error between the expected target action and the actual target action of the target software driver.

[0040] In practice, the aforementioned execution entity can invoke a pre-trained GRU (Gated Recurrent Unit) network model with temporal characteristics, and use the aforementioned target pose information, attitude information, and 3D point cloud information as input data to predict the model and obtain error pose information. Here, the GRU network model can represent the mapping relationship between the input data (target pose information, attitude information, and 3D point cloud information) and the output data (error pose information).

[0041] In some optional implementations of certain embodiments, based on the target pose information, the attitude information, and the three-dimensional point cloud information, the execution entity can generate error pose information corresponding to the target software driver through the following steps: The first step is to generate actual pose information based on the above attitude information and the above 3D point cloud information.

[0042] In practice, based on the aforementioned posture information and 3D point cloud information, the executing entity can generate actual pose information through the following steps: The first generation step involves performing zero-bias correction processing on the aforementioned attitude information to obtain corrected attitude information. In practice, the executing entity can determine the corrected attitude information as the difference between the aforementioned attitude information and a preset zero-bias estimation threshold. The preset zero-bias estimation threshold is the average value of various attitude information collected over a pre-defined time period.

[0043] The second generation step involves filtering the obtained corrected attitude information to obtain filtered attitude information. In practice, the aforementioned execution entity can use a low-pass filter to filter the obtained corrected attitude information to obtain filtered attitude information.

[0044] The third generation step involves pre-integrating the filtered attitude information to obtain pre-integrated attitude information. Here, pre-integration can be understood as integrating the attitude information collected by the attitude sensor between two consecutive LiDAR frames. Since the LiDAR data acquisition frequency is low, while the attitude sensor data acquisition frequency is high, integrating all the attitude sensor data from scratch in each optimization would be computationally intensive. Therefore, using pre-integrated segments can reduce the computational workload of subsequent optimizations.

[0045] The fourth generation step involves performing distortion correction processing on the aforementioned 3D point cloud information to obtain distorted point cloud information. In practice, the aforementioned execution entity can obtain distorted point cloud information through the following steps: The first step is to perform spherical linear interpolation on the pre-integrated attitude information for each three-dimensional point in the above three-dimensional point cloud information using the timestamp corresponding to the three-dimensional point information, so as to generate the radar pose corresponding to the above three-dimensional point information and obtain the radar pose set.

[0046] The second step is to determine the target 3D point information as the 3D point information with the latest timestamp in the above 3D point cloud information.

[0047] The third step is to determine the radar pose corresponding to the target's three-dimensional point information in the radar pose set as the reference pose.

[0048] The fourth step involves performing the following processing steps for each radar pose in the aforementioned radar pose set: The first sub-step involves generating a relative pose based on the aforementioned reference pose and radar pose. The relative pose can be generated using the following formula: .

[0049] Among them, the above It can represent the radar pose. (The above...) This can represent the inverse of the reference pose. (The above...) It can represent relative pose.

[0050] The second sub-step involves updating the three-dimensional point information corresponding to the radar pose based on the relative pose described above, to generate updated three-dimensional point information as distortion-free point information. Here, the product of the relative pose and the three-dimensional point information can be determined as the distortion-free point information.

[0051] The point cloud composed of the obtained distortion-free point information is defined as the distortion-free point cloud information.

[0052] The fifth generation step involves performing inter-frame registration processing on the aforementioned distorted point cloud information to obtain inter-frame registered point cloud information. In practice, the executing entity can use a preset matching algorithm to match the aforementioned distorted point cloud information with the distorted point cloud information of the previous frame corresponding to the aforementioned distorted point cloud information to obtain inter-frame registered point cloud information. Here, the preset matching algorithm can be the ICP (Iterative Closest Point) algorithm. The inter-frame registered point cloud information can characterize the pose transformation of the current frame point cloud information of the lidar relative to the previous frame point cloud information.

[0053] The sixth generation step involves fusing and optimizing the inter-frame registration point cloud information and the pre-integrated pose information to obtain optimized fused pose information. In practice, the executing entity can utilize a preset optimizer (Gauss-Newton) to fuse and optimize the inter-frame registration point cloud information and the pre-integrated pose information to obtain optimized fused pose information.

[0054] The seventh generation step involves performing loop closure detection processing on the optimized and fused pose information to obtain loop closure detected pose information as the actual pose information. In practice, the execution entity can perform loop closure detection processing on the optimized and fused pose information using the following steps to obtain loop closure detected pose information as the actual pose information: The first step is to extract the pose features corresponding to the optimized and fused pose information. In practice, the executing entity can use a feature extraction model to extract features from the optimized and fused pose information to obtain pose features. Here, the feature extraction model can be a VGG (Visual Geometry Group) network model (excluding connection layers).

[0055] The second step involves generating a pose distance sequence based on the aforementioned pose features and the preset pose feature sequence. In practice, for each preset pose feature in the preset pose feature sequence, the executing entity can determine the pose distance as the Euclidean distance between the preset pose feature and the aforementioned pose feature, thus obtaining the pose distance sequence. The aforementioned preset pose feature sequence can be a sequence of pose features corresponding to each pose information of a preset complete target action executed by a pre-defined target software driver. The aforementioned preset complete target action can be a pre-defined action to complete a preset operation task. The aforementioned preset operation task can be a pre-defined operation task. The aforementioned preset operation task is not specifically limited here. For example, the aforementioned preset operation task can be the target software driver performing an operation task of sorting electronic products (e.g., electronic bracelets).

[0056] The third step involves determining the pose distances in the aforementioned pose distance sequence that satisfy a preset distance condition as a candidate pose distance set. The preset distance condition can be that the pose distance is greater than a preset distance threshold. This preset distance threshold can be a pre-defined distance threshold.

[0057] The fourth step is to determine each preset pose feature corresponding to the above candidate pose distance set as a candidate preset pose feature set.

[0058] The fifth step involves using a preset matching algorithm to perform feature matching processing on the aforementioned pose features and each candidate preset pose feature included in the aforementioned candidate preset pose feature set, resulting in a matching result set. The matching results in the aforementioned matching result set include: relative pose and fitting score. The relative pose characterizes the transformation from the pose feature to the candidate preset pose feature. The fitting score characterizes the matching probability between the aforementioned pose feature and the aforementioned candidate preset pose feature. Here, the preset matching algorithm can be a pre-defined matching algorithm. For example, the aforementioned preset matching algorithm can be the G-ICP (Generalized Iterative Closest Point) algorithm.

[0059] The sixth step involves determining the target fitting score based on the fact that at least one fitting score in the matching result set is less than a preset fitting threshold. This target fitting threshold can be a pre-defined fitting threshold, for example, 0.01.

[0060] The seventh step is to determine the relative pose corresponding to the above fitting scores as the closure constraint information.

[0061] The eighth step involves adding the aforementioned loop constraint information to a preset pose graph, and then using a preset optimizer to optimize the pose of the preset pose graph, obtaining the optimized fused pose information as the actual pose information. The preset pose graph can be a pose graph constructed based on the pose information of each pose of the target software driving the execution of a preset complete target action. Each pose information corresponds one-to-one with a preset pose feature in the preset pose feature sequence. Nodes in the preset pose graph can be pose information. Edges in the preset pose graph can represent the constraint relationships between nodes. The preset optimizer can be a pre-defined optimizer, for example, the aforementioned preset optimizer can be a G2o (General Graph Optimization) optimizer. The aforementioned constraint relationships can represent the relative geometric relationship between two connected pose nodes (describing how the coordinate system should transform from one pose node to another).

[0062] In the ninth step, in response to the fact that all fitting scores in the above matching result set are greater than or equal to the above preset fitting threshold, the optimized fused pose information is determined as the actual pose information.

[0063] The second step is to generate error pose information corresponding to the target software driver based on the target pose information and the actual pose information.

[0064] In practice, based on the target pose information and the actual pose information, the execution entity can generate the error pose information corresponding to the target software driver through the following steps: The first step is to extract the target position information and target attitude information from the target pose information mentioned above. This extraction can be understood as splitting the data according to the index corresponding to the target position information in the target pose information.

[0065] The second step is to extract the actual position information and actual posture information from the above actual pose information.

[0066] The third step involves determining the position error vector based on the target position information extracted from the target pose information and the actual position information extracted from the actual pose information. In practice, the executing entity can determine the position error vector as the vector corresponding to the difference between the target position information and the actual position information.

[0067] The fourth step involves determining the attitude error vector based on the target attitude information extracted from the target pose information and the actual attitude information extracted from the actual pose information. This attitude error vector represents the minimum axis-angle rotation required to rotate from the actual pose to the target pose. Here, both the target attitude information and the actual attitude information are represented using quaternions. In practice, the execution entity can determine the attitude error vector using the following steps: The first sub-step involves generating difference rotation information based on the actual attitude information and the target attitude information described above. In practice, difference rotation information can be generated using the following formula: .

[0068] Among them, the above It can represent target attitude information. (The above) This can represent difference rotation information. (The above...) This can represent actual attitude information. (The above...) This can represent the conjugate of actual attitude information. (The above...) It can represent multiplication.

[0069] The second sub-step involves performing vector rotation processing on the aforementioned difference rotation information to obtain the attitude error vector. In practice, the execution entity can obtain the attitude error vector by calling a preset conversion function interface and using the aforementioned difference rotation information as input parameters to the preset conversion function interface. The preset conversion function interface can be a pre-encapsulated function that converts the difference rotation information into an attitude error vector. For example, the preset conversion function interface can be a conversion function built into the scipy library.

[0070] The fifth step involves combining the aforementioned position error vector and attitude error vector to generate error pose information. This combination can be achieved through concatenation. For example, the position error vector can be vector A, the attitude error vector can be vector B, and the error pose vector can be [A, B].

[0071] Step 205: Generate the compensation air pressure information for the corresponding target software actuator based on the error pose information.

[0072] In some embodiments, the execution entity may generate compensation air pressure information corresponding to the target software driver based on the error pose information.

[0073] In practice, the aforementioned execution entity can use a pre-trained LSTM (Long Short-Term Memory Network) model with temporal features, along with the aforementioned error pose information as input data, to predict and obtain the compensated air pressure information. Here, the LSTM network model can represent the mapping relationship between the input data (error pose information) and the output data (compensated air pressure information).

[0074] In some optional implementations of certain embodiments, based on the aforementioned error pose information, the executing entity can generate compensated air pressure information through the following steps: The first step is to filter the aforementioned error pose information to obtain filtered error pose information. The filtering methods can include, but are not limited to, low-pass filtering, moving average filtering, and Kalman filtering. Here, Kalman filtering is chosen as the preferred method. In practice, the executing entity can perform Kalman filtering on the error pose information to obtain filtered error pose information. This reduces the impact of high-frequency noise in the data on the output air pressure.

[0075] The second step involves generating original compensation air pressure information based on the aforementioned filtered error pose information. In practice, the aforementioned actuator can use a proportional-integral-derivative control algorithm to predict the filtered error pose information and obtain the original compensation air pressure information. This original compensation air pressure information can be air pressure information capable of correcting the error between the desired and actual paths of the target software actuator.

[0076] The third step is to perform amplitude limiting processing on the above-mentioned original compensation pressure information to obtain amplitude-limited original compensation pressure information. In practice, the above-mentioned implementing entity can perform amplitude limiting processing on the original compensation pressure information using the following formula to obtain amplitude-limited original compensation pressure information: .

[0077] Among them, the above This can represent the original compensation air pressure information for the amplitude limiting. (The above...) This can represent a preset upper limit air pressure threshold. (The above...) This can represent a preset lower limit air pressure threshold. Here, both the preset lower limit air pressure threshold and the preset upper limit air pressure threshold are pre-set air pressure thresholds, and the preset lower limit air pressure threshold is less than the preset upper limit air pressure threshold. This can represent the original compensated air pressure information. (The above...) This means that when the original compensated air pressure information is less than a preset lower limit air pressure threshold, the preset lower limit air pressure threshold is used. The above... This means that when the original compensation pressure information is greater than a preset upper limit pressure threshold, the preset upper limit pressure threshold will be used. This ensures that the transmitted pressure is within a safe and effective range.

[0078] The fourth step involves performing integral anti-saturation processing on the aforementioned original limited-amplitude compensation pressure information to obtain the compensation pressure information. Here, the integral anti-saturation processing methods can include, but are not limited to, integral limiting processing and integral separation processing. In practice, the aforementioned executing entity can perform integral separation processing on the aforementioned original limited-amplitude compensation pressure information to obtain the compensation pressure information.

[0079] Step 206: Determine the target air pressure information based on the compensated air pressure information and the predicted air pressure information.

[0080] In some embodiments, the executing entity may determine the target air pressure information based on the compensated air pressure information and the predicted air pressure information. In practice, the executing entity may determine the target air pressure information as the sum of the compensated air pressure information and the predicted air pressure information.

[0081] Step 207: Based on the target air pressure information and the compensation air pressure information, generate an air pressure control command and send the air pressure control command to the air pressure valve control device to control the target software driver to perform displacement operation.

[0082] In some embodiments, the execution entity may generate a pressure control command based on the target pressure information and the compensation pressure information, and send the pressure control command to the pressure valve control device to control the target software driver to perform a displacement operation.

[0083] The aforementioned pneumatic valve control equipment may include: a motor drive module, a positive pressure pump, a negative pressure pump, a digital-to-analog converter module, a positive pressure proportional valve, a negative pressure proportional valve, a pulse width modulation module, a positive pressure three-way solenoid valve, and a negative pressure three-way solenoid valve.

[0084] In practice, the aforementioned executing entity can determine the air pressure control command corresponding to the target air pressure information by querying a preset air pressure command mapping table. This preset air pressure command mapping table represents the correspondence between the target air pressure information and the control quantities (e.g., duty cycle) identifiable by the air pressure valve (e.g., a three-way solenoid valve). Here, an open-loop approach is used to generate the air pressure control command; that is, the process of generating the air pressure control command does not involve compensating for air pressure information.

[0085] In some optional implementations of certain embodiments, the aforementioned execution entity may generate a pressure control command based on the target pressure information and the compensation pressure information through the following steps, and send the pressure control command to the pressure valve control device to control the target software driver to perform a displacement operation: The first step involves generating air pressure control commands based on the target air pressure information, the compensated air pressure information, and the preset command generation function interface. These air pressure control commands may include: motor drive commands, digital voltage signals, and pulse width modulation (PWM) parameters. The motor drive commands can be commands issued by the controller to drive and control the operation of the positive and negative air pressure pumps. The digital voltage signals can be digital voltage signals issued by the controller to drive and control the positive and negative proportional valves for air pressure regulation. The PWM parameters can be parameters issued by the controller that cause the PWM module to generate square wave electrical signals (corresponding to square wave electrical signals with corresponding duty cycles and frequencies) to drive the opening and closing of the positive and negative three-way solenoid valves. The preset command generation function interface can be a pre-packaged function that can generate air pressure control commands based on the target air pressure information and the compensated air pressure information. In practice, the executing entity can generate air pressure control commands by calling the preset command generation function interface and using the target air pressure information and the compensated air pressure information as input parameters.

[0086] The second step involves performing the following rapid pressure adjustment steps in response to the absolute value of the aforementioned compensated air pressure information being greater than the first preset air pressure threshold: In the first sub-step, in response to the target air pressure information being greater than a third preset air pressure threshold, a motor drive command indicating the activation of the positive pressure pump is sent to the motor drive module to activate the positive pressure pump, and the pulse width modulation parameters are sent to the pulse width modulation module to drive the positive pressure three-way solenoid valve to open. The first preset air pressure threshold can be a pre-set air pressure threshold. The third preset air pressure threshold can be a positive number.

[0087] As an example, you can refer to Figure 3 When the absolute value of the compensated air pressure information is greater than the first preset air pressure threshold, and the target air pressure information is greater than the third preset air pressure threshold, the controller 301 can send a motor drive command to the motor drive module 302 to start the positive pressure pump 303. Simultaneously, the controller 301 sends pulse width modulation parameters to the pulse width modulation module 307 to open the positive pressure three-way solenoid valve 308 for air pressure regulation. The signal corresponding to the red line is the control signal, and the signal corresponding to the black line is the air path signal. Therefore, when the error between the expected path and the actual path of the aforementioned target software driver is large, and the air pressure value corresponding to the target air pressure information is large and positive, the positive pressure three-way solenoid valve is opened, entering a rapid positive pressure regulation mode.

[0088] In the second sub-step, in response to the target air pressure information being less than the opposite of the third preset air pressure threshold, a motor drive command representing the activation of the negative pressure air pump is sent to the motor drive module to activate the negative pressure air pump, and the pulse width modulation parameter is sent to the pulse width modulation module to drive the negative pressure three-way solenoid valve to open.

[0089] As an example, you can refer to Figure 4 When the absolute value of the compensated air pressure information is greater than the first preset air pressure threshold, and the target air pressure information is less than the negative of the third preset air pressure threshold, the controller 301 can send a motor drive command to the motor drive module 302 to start the negative air pressure pump 304. Simultaneously, the controller 301 sends pulse width modulation parameters to the pulse width modulation module 307 to activate the negative air pressure three-way solenoid valve 309 for air pressure regulation. Therefore, when the error between the expected path and the actual path representing the aforementioned target software driver is large, and the air pressure value corresponding to the target air pressure information is large and negative, the negative air pressure three-way solenoid valve is activated, entering a rapid negative air pressure regulation mode.

[0090] In the third sub-step, in response to the absolute value of the target air pressure information being less than or equal to the third preset air pressure threshold, a motor drive command representing the activation of the positive pressure pump and the negative pressure pump is sent to the motor drive module to activate the positive pressure pump and the negative pressure pump, and the pulse width modulation parameter is sent to the pulse width modulation module to drive the positive pressure three-way solenoid valve and the negative pressure three-way solenoid valve to open.

[0091] As an example, you can refer to Figure 5 When the absolute value of the compensated air pressure information is greater than the first preset air pressure threshold, and the absolute value of the target air pressure information is less than or equal to the third preset air pressure threshold, the controller 301 can send a motor drive command to the motor drive module 302 to activate the positive air pressure pump 303 and the negative air pressure pump 304. Simultaneously, the controller 301 sends pulse width modulation parameters to the pulse width modulation module 307 to activate the positive air pressure three-way solenoid valve 308 and the negative air pressure three-way solenoid valve 309 for mixed air pressure regulation. Therefore, when the error between the expected path and the actual path representing the target software driver is large, and the absolute value of the target air pressure information is small, the positive and negative air pressure three-way solenoid valves are activated, entering a rapid mixed regulation mode for positive and negative air pressure.

[0092] Third, in response to the absolute value of the compensated air pressure information being less than the second preset air pressure threshold, the following precise air pressure adjustment steps are performed: In the first sub-step, in response to the target air pressure information being greater than the third preset air pressure threshold, a motor drive command indicating the activation of the positive pressure pump is sent to the motor drive module to activate the positive pressure pump, and a digital voltage signal is sent to the digital-to-analog converter module to control the positive pressure proportional valve to regulate air pressure. The first preset air pressure threshold is greater than the second preset air pressure threshold. Both the first and second preset air pressure thresholds are pre-set air pressure thresholds.

[0093] As an example, you can refer to Figure 6 When the absolute value of the compensated air pressure information is less than the second preset air pressure threshold, and the target air pressure information is greater than the third preset air pressure threshold, the controller 301 can send a motor drive command to the motor drive module 302 to start the positive pressure pump 303. Simultaneously, the controller 301 can send a digital voltage signal to the digital-to-analog converter module 310 to drive and control the positive pressure proportional valve 305 for air pressure regulation. Therefore, when the error between the desired path and the actual path representing the target software driver is small, and the air pressure value corresponding to the target air pressure information is large and positive, the positive pressure proportional valve is selected to drive, entering the precise positive pressure regulation mode.

[0094] In the second sub-step, in response to the target air pressure information being less than the opposite of the third preset air pressure threshold, a motor drive command representing the activation of the negative air pressure pump is sent to the motor drive module to activate the negative air pressure pump, and the digital voltage signal is sent to the digital-to-analog converter module to control the negative pressure proportional valve to regulate the air pressure.

[0095] As an example, you can refer to Figure 7 When the absolute value of the compensated air pressure information is less than the second preset air pressure threshold, and the target air pressure information is less than the negative of the third preset air pressure threshold, the controller 301 can send a motor drive command to the motor drive module 302 to start the negative air pressure pump 304. Simultaneously, the controller 301 can send a digital voltage signal to the digital-to-analog converter module 310 to drive and control the negative pressure proportional valve 306 for air pressure regulation. Therefore, when the error between the desired path and the actual path representing the target software driver is small, and the air pressure value corresponding to the target air pressure information is large and negative, the negative pressure proportional valve is selected to drive, entering a precise negative air pressure regulation mode.

[0096] In the third sub-step, in response to the absolute value of the target air pressure information being less than or equal to the third preset air pressure threshold, a motor drive command representing the activation of the positive air pressure pump and the negative air pressure pump is sent to the motor drive module to activate the positive air pressure pump and the negative air pressure pump, and the digital voltage signal is sent to the digital-to-analog conversion module to control the positive pressure proportional valve and the negative pressure proportional valve to perform mixed air pressure regulation.

[0097] As an example, you can refer to Figure 8 When the absolute value of the compensated air pressure information is less than the second preset air pressure threshold, and the absolute value of the target air pressure information is less than or equal to the third preset air pressure threshold, the controller 301 can send a motor drive command to the motor drive module 302 to activate the positive air pressure pump 303 and the negative air pressure pump 304. Simultaneously, the controller 301 can send a digital voltage signal to the digital-to-analog converter module 310 to drive and control the positive pressure proportional valve 305 and the negative pressure proportional valve 306 for air pressure regulation. Therefore, when the error between the desired path and the actual path representing the target software driver is small, and the absolute value of the target air pressure information is small, the positive pressure proportional valve and the negative pressure proportional valve are selected to drive, entering a precise positive and negative air pressure mixed regulation mode.

[0098] The various embodiments of this disclosure have the following beneficial effects: the pneumatic control method applied to a pneumatic valve control device according to some embodiments of this disclosure can reduce erroneous displacement operations of the soft actuator. Specifically, the reason for the large number of displacement operation errors of the soft actuator is that the method of constructing the mechanical model is based on a large number of assumptions. In the process of making assumptions, it is unavoidable to simplify or ignore the influence of many factors, which greatly reduces the effectiveness of the theoretical model in practical applications (for example, if structural assumptions are made about the soft actuator arm, the model will completely fail when the soft actuator arm faces external interference). Furthermore, although the pressure sensor measures the air pressure in the air chamber, it cannot know whether this air pressure has produced the desired pose (due to the nonlinearity, hysteresis, creep, and changes in environmental load of the soft actuator, the same air pressure value may lead to completely different poses at different times and under different states). This results in a large number of displacement operation errors of the soft actuator. Based on this, the pneumatic control method applied to a pneumatic valve control device according to some embodiments of this disclosure first obtains the target pose information for the target soft actuator. The target pose information represents the segmented displacement path of the target soft actuator. Therefore, target pose information representing the desired path of the target soft actuator can be obtained, and the corresponding segmented target action can be completed with the desired path as a reference. Then, the target pose information is input into a pre-trained air pressure information generation model to obtain predicted air pressure information. This air pressure information generation model represents the mapping relationship between pose information and air pressure information. Thus, based on the network model, predicted air pressure information corresponding to the displacement action of the target soft actuator can be obtained, and based on the predicted air pressure information, the target soft actuator can be subjected to extension and retraction displacement operations. Next, multi-dimensional data acquisition of the target soft actuator is performed using an attitude sensor and a lidar located at the end of the target soft actuator, obtaining attitude information and 3D point cloud information. The data acquired by the attitude sensor and the lidar are time-synchronized. Thus, attitude information representing the actual attitude of the target soft actuator and 3D point cloud information representing its actual position can be obtained. Next, based on the target pose information, the attitude information, and the 3D point cloud information, error pose information corresponding to the target soft actuator is generated. Thus, error pose information representing the error between the desired action and the actual action of the target soft actuator can be obtained. Then, based on the aforementioned error pose information, compensation pressure information corresponding to the target software actuator is generated. Thus, the error between the expected and actual actions of the target software actuator can be converted into compensation pressure information that can compensate for the action error, thereby correcting the target software actuator's actions. Afterwards, the target pressure information is determined based on the aforementioned compensation pressure information and the aforementioned predicted pressure information. Thus, the target pressure information characterizing the action error correction can be obtained.Finally, based on the target air pressure information and the compensated air pressure information, an air pressure control command is generated and sent to the air pressure valve control device to control the target software actuator to perform a displacement operation. Thus, the target software actuator can perform a displacement operation according to the desired path based on the correct air pressure control command, completing the target action. Furthermore, because air pressure prediction is performed using a pre-trained network model, the problem of traditional air pressure prediction models relying on numerous assumptions and resulting in low effectiveness in practical applications can be avoided. Moreover, because the pose information representing the desired path of the target software actuator and the actual pose information fed back from the attitude sensor and lidar can be used to determine and correct the pose error, erroneous displacement operations of the software actuator can be reduced.

[0099] Furthermore, this disclosure provides a pneumatic control system applied to the aforementioned pneumatic control method for pneumatic valve control equipment. The pneumatic control system includes: a controller, a posture sensor, a lidar, and a pneumatic valve control device. The pneumatic valve control device includes: a motor drive module, a positive pressure pump, a negative pressure pump, a digital-to-analog converter module, a positive pressure proportional valve, a negative pressure proportional valve, a pulse width modulation module, a positive pressure three-way solenoid valve, and a negative pressure three-way solenoid valve. The positive pressure proportional valve includes a filter pressure reducing valve and a proportional valve, and the negative pressure proportional valve includes a filter pressure reducing valve and a proportional valve. The posture sensor, lidar, motor drive module, digital-to-analog converter module, and pulse width modulation module are all communicatively connected to the controller. The positive pressure pump and negative pressure pump are both communicatively connected to the motor drive module. The positive pressure proportional valve and negative pressure proportional valve are both communicatively connected to the digital-to-analog converter module. The positive pressure three-way solenoid valve and negative pressure three-way solenoid valve are both communicatively connected to the pulse width modulation module. The outlet of the positive pressure pump is connected to the inlet of the positive pressure proportional valve via an air path. The filter pressure reducing valve is connected in series with the proportional valve via an air path. The outlet of the positive pressure proportional valve is connected to the inlet of the positive pressure three-way solenoid valve via an air path. The outlet of the positive pressure three-way solenoid valve is connected to the positive pressure chamber of the target software driver via an air path. The exhaust port of the positive pressure three-way solenoid valve is open to the atmosphere. The inlet of the negative pressure pump is connected to the outlet of the negative pressure proportional valve via an air path. The inlet of the negative pressure proportional valve is connected to the outlet of the negative pressure three-way solenoid valve via an air path. The inlet of the negative pressure three-way solenoid valve is connected to the negative pressure chamber of the target software driver. The exhaust port of the negative pressure three-way solenoid valve is open to the atmosphere.

[0100] The system includes: a controller configured to: run a control algorithm to generate pneumatic control commands based on target pose information and feedback actual pose information, thereby controlling the target software actuator to perform displacement operations; a pose sensor configured to: acquire the pose information of the target software actuator; a lidar configured to: acquire the three-dimensional spatial position information of the target software actuator; and a pneumatic valve control device configured to: control the inflation and deflation of the target software actuator according to the pneumatic control commands from the controller. The pneumatic valve control device includes: a motor drive module, a positive pressure pump, a negative pressure pump, a digital-to-analog converter module, a positive pressure proportional valve, a negative pressure proportional valve, a pulse width modulation module, a positive pressure three-way solenoid valve, and a negative pressure three-way solenoid valve. The motor drive module is configured to: receive motor drive commands from the controller and control and drive the working states of the positive and negative pressure pumps. The positive pressure pump is configured to: supply air to the target software actuator... The target software driver outputs positive pressure gas. The negative pressure pump is configured to output negative pressure gas to the target software driver. The digital-to-analog converter module is configured to convert the digital voltage signal emitted by the controller into an analog voltage signal. The positive pressure proportional valve is configured to receive the analog voltage signal and adjust the positive pressure value output to the target software driver. The negative pressure proportional valve is configured to receive the analog voltage signal and adjust the negative pressure value output to the target software driver. The pulse width modulation module is configured to receive the pulse width modulation parameters of the controller and generate a square wave electrical signal with a corresponding duty cycle and frequency according to the pulse width modulation parameters to drive the on / off switching of the positive pressure three-way solenoid valve and the negative pressure three-way solenoid valve. The positive pressure three-way solenoid valve is configured to be controlled on / off to switch the connection between the gas path of the target software driver and the positive pressure pump. The negative pressure three-way solenoid valve is configured to be controlled on / off to switch the connection between the gas path of the target software driver and the negative pressure pump.

[0101] Further reference Figure 9 As an implementation of the methods shown in the above figures, this disclosure provides some embodiments of a pneumatic control device applied to a pneumatic valve control equipment. These device embodiments are similar to... Figure 2 Corresponding to the method embodiments shown, the pneumatic control device applied to pneumatic valve control equipment can be specifically applied to various electronic devices.

[0102] like Figure 9As shown, a pneumatic control device 900 applied to a pneumatic valve control device in some embodiments includes: an acquisition unit 901, an input unit 902, a data acquisition unit 903, a first generation unit 904, a second generation unit 905, a determination unit 906, and a third generation unit 907. The acquisition unit 901 is configured to acquire target pose information for a target software actuator, wherein the target pose information represents the segmented displacement path of the target software actuator; the input unit 902 is configured to input the target pose information into a pre-trained pneumatic information generation model to obtain predicted pneumatic pressure information, wherein the pneumatic information generation model represents the mapping relationship between pose information and pneumatic pressure information; the data acquisition unit 903 is configured to perform multi-dimensional data acquisition on the target software actuator using an attitude sensor and a lidar located at the end of the target software actuator to obtain attitude information and three-dimensional point cloud information, wherein the attitude sensor and the lidar acquire data... Time synchronization; the first generation unit 904 is configured to generate error pose information corresponding to the target software driver based on the target pose information, the attitude information, and the three-dimensional point cloud information; the second generation unit 905 is configured to generate compensation air pressure information corresponding to the target software driver based on the error pose information; the determination unit 906 is configured to determine the target air pressure information based on the compensation air pressure information and the predicted air pressure information; the third generation unit 907 is configured to generate air pressure control commands based on the target air pressure information and the compensation air pressure information, and send the air pressure control commands to the air pressure valve control device to control the target software driver to perform displacement operations.

[0103] It is understood that the units described in the pneumatic control device 900 applied to pneumatic valve control equipment are similar to those in the reference device. Figure 2 The steps in the described method correspond accordingly. Therefore, the operations, features, and beneficial effects described above for the method are also applicable to the pneumatic control device 900 and the units contained therein, which are applied to pneumatic valve control equipment, and will not be repeated here.

[0104] The following is for reference. Figure 10 It shows a schematic diagram of the structure of an electronic device 1000 (e.g., a computing device) suitable for implementing some embodiments of the present disclosure. Figure 10 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of this disclosure.

[0105] like Figure 10As shown, the electronic device 1000 may include a processing unit 1001 (e.g., a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage device 1008 into a random access memory (RAM) 1003. The RAM 1003 also stores various programs and data required for the operation of the electronic device 1000. The processing unit 1001, ROM 1002, and RAM 1003 are interconnected via a bus 1004. An input / output (I / O) interface 1005 is also connected to the bus 1004.

[0106] Typically, the following devices can be connected to the I / O interface 1005: input devices 1006 including, for example, a touchscreen, touchpad, keyboard, mouse, camera, microphone, accelerometer, gyroscope, etc.; output devices 1007 including, for example, a liquid crystal display (LCD), speaker, vibrator, etc.; storage devices 1008 including, for example, magnetic tape, hard disk, etc.; and communication devices 1009. Communication device 1009 allows electronic device 1000 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 10 An electronic device 1000 with various devices is shown; however, it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed alternatively. Figure 10 Each box shown can represent a device or multiple devices as needed.

[0107] In particular, according to some embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, some embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication device 1009, or installed from storage device 1008, or installed from ROM 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of some embodiments of this disclosure.

[0108] It should be noted that, in some embodiments of this disclosure, the computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium may be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In some embodiments of this disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In some embodiments of this disclosure, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0109] In some implementations, clients and servers can communicate using any currently known or future-developed network protocol such as HTTP (Hypertext Transfer Protocol), and can interconnect with digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (“LANs”), wide area networks (“WANs”), the Internet (e.g., the Internet of Things), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future-developed networks.

[0110] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device. The aforementioned computer-readable medium carries one or more programs that, when executed by the electronic device, cause the electronic device to: acquire target pose information for a target software driver, wherein the target pose information represents the segmented displacement path of the target software driver; input the target pose information into a pre-trained air pressure information generation model to obtain predicted air pressure information, wherein the air pressure information generation model represents the mapping relationship between pose information and air pressure information; and perform multi-dimensional data acquisition on the target software driver using an attitude sensor and a lidar located at the end of the target software driver to obtain the attitude. The system includes target pose information and 3D point cloud information, wherein the attitude sensor and the lidar acquire data in time; based on the target pose information, the attitude information, and the 3D point cloud information, error pose information corresponding to the target software driver is generated; based on the error pose information, compensation air pressure information corresponding to the target software driver is generated; based on the compensation air pressure information and the predicted air pressure information, target air pressure information is determined; based on the target air pressure information and the compensation air pressure information, an air pressure control command is generated, and the air pressure control command is sent to an air pressure valve control device to control the target software driver to perform displacement operations.

[0111] Computer program code for performing operations of some embodiments of this disclosure can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0112] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0113] The functions described above in this document can be performed at least in part by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), system-on-a-chip (SoCs), complex programmable logic devices (CPLDs), and so on.

[0114] The above description is merely a selection of preferred embodiments of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in the embodiments of this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions disclosed in the embodiments of this disclosure.

Claims

1. A pneumatic control method applied to a pneumatic valve control device, characterized in that, include: Obtain target pose information for a target software driver, wherein the target pose information represents the segmented displacement path of the target software driver; The target pose information is input into a pre-trained air pressure information generation model to obtain predicted air pressure information, wherein the air pressure information generation model represents the mapping relationship between pose information and air pressure information. By using an attitude sensor and a lidar installed at the end of the target software driver, multidimensional data is acquired from the target software driver to obtain attitude information and three-dimensional point cloud information. The attitude sensor and the lidar acquire data in time. Based on the target pose information, the attitude information, and the three-dimensional point cloud information, error pose information corresponding to the target software driver is generated; Based on the error pose information, compensation air pressure information corresponding to the target software driver is generated; The target air pressure information is determined based on the compensated air pressure information and the predicted air pressure information; Based on the target air pressure information and the compensation air pressure information, an air pressure control command is generated, and the air pressure control command is sent to the air pressure valve control device to control the target software driver to perform a displacement operation.

2. The method according to claim 1, wherein, The step of generating error pose information corresponding to the target software driver based on the target pose information, the attitude information, and the 3D point cloud information includes: Based on the attitude information and the three-dimensional point cloud information, actual pose information is generated; Based on the target pose information and the actual pose information, error pose information corresponding to the target software driver is generated.

3. The method according to claim 2, characterized in that, The step of generating error pose information corresponding to the target software driver based on the target pose information and the actual pose information includes: Extract target position information and target attitude information from the target pose information; Extract actual position information and actual posture information from the actual pose information; A position error vector is determined based on the target position information extracted from the target pose information and the actual position information extracted from the actual pose information. Based on the target pose information extracted from the target pose information and the actual pose information extracted from the actual pose information, a pose error vector is determined, wherein the pose error vector represents the minimum rotation axis angle required to rotate from the actual pose to the target pose; The position error vector and the attitude error vector are combined to generate error pose information.

4. The method according to claim 1, characterized in that, The step of generating compensation air pressure information corresponding to the target software driver based on the error pose information includes: The error pose information is filtered to obtain filtered error pose information; Based on the filtered error pose information, the original compensated air pressure information is generated; The original compensation pressure information is subjected to amplitude limiting processing to obtain amplitude-limited original compensation pressure information; The original compensation pressure information of the amplitude limit is subjected to integral anti-saturation processing to obtain the compensation pressure information.

5. The method according to claim 1, wherein, Before inputting the target pose information into a pre-trained air pressure information generation model to obtain predicted air pressure information, the method further includes: Obtain a set of air pressure samples, wherein the air pressure samples in the set of air pressure samples include sample pose information and sample air pressure information corresponding to the sample pose information; The following training steps are performed based on the air pressure sample set: The pose information of at least one air pressure sample in the air pressure sample set is input into the initial air pressure information generation model to obtain the air pressure information corresponding to each air pressure sample in the at least one air pressure sample. Compare the air pressure information corresponding to each air pressure sample in the at least one air pressure sample with the corresponding sample air pressure information; Based on the comparison results, determine whether the initial air pressure information generation model has achieved the preset optimization objective; In response to the determination that the initial air pressure information generation model has reached the optimization objective, the initial air pressure information generation model is determined as the trained air pressure information generation model.

6. The method according to claim 5, wherein, The steps for training the air pressure information generation model further include: In response to the determination that the initial air pressure information generation model has not achieved the optimization objective, the network parameters of the initial air pressure information generation model are adjusted, and an air pressure sample set is formed using unused air pressure samples. The adjusted initial air pressure information generation model is then used as the initial air pressure information generation model, and the training steps are executed again.

7. A pneumatic control system, applied to the pneumatic control method for a pneumatic valve control device as described in any one of claims 1 to 6, characterized in that, include: The system includes a controller, a posture sensor, a lidar, and a pneumatic valve control device. The pneumatic valve control device comprises: a motor drive module, a positive pressure pump, a negative pressure pump, a digital-to-analog converter module, a positive pressure proportional valve, a negative pressure proportional valve, a pulse width modulation module, a positive pressure three-way solenoid valve, and a negative pressure three-way solenoid valve. The positive pressure proportional valve includes a filter pressure reducing valve and a proportional valve. The negative pressure proportional valve also includes a filter pressure reducing valve and a proportional valve. The pose sensor, the lidar, the motor drive module, the digital-to-analog converter module, and the pulse width modulation module are all communicatively connected to the controller; Both the positive pressure pump and the negative pressure pump are communicatively connected to the motor drive module; Both the positive pressure proportional valve and the negative pressure proportional valve are communicatively connected to the digital-to-analog converter module. Both the positive pressure three-way solenoid valve and the negative pressure three-way solenoid valve are communicatively connected to the pulse width modulation module. The outlet of the positive pressure pump is connected to the inlet of the positive pressure proportional valve through an air circuit, and the filter pressure reducing valve is connected in series with the proportional valve through an air circuit. The outlet of the positive pressure proportional valve is connected to the inlet of the positive pressure three-way solenoid valve via an air passage. The outlet of the positive pressure three-way solenoid valve is connected to the positive pressure chamber of the target software driver via an air passage. The exhaust port of the positive pressure three-way solenoid valve is connected to the atmosphere; The inlet of the negative pressure pump is connected to the outlet of the negative pressure proportional valve via an air circuit. The inlet of the negative pressure proportional valve is connected to the outlet of the negative pressure three-way solenoid valve via an air passage. The inlet of the negative pressure three-way solenoid valve is connected to the negative pressure chamber of the target software driver. The exhaust port of the negative pressure three-way solenoid valve is connected to the atmosphere.

8. A pneumatic control device applied to pneumatic valve control equipment, characterized in that, include: The acquisition unit is configured to acquire target pose information for a target software driver, wherein the target pose information characterizes the segmented displacement path of the target software driver. The input unit is configured to input the target pose information into a pre-trained air pressure information generation model to obtain predicted air pressure information, wherein the air pressure information generation model represents the mapping relationship between pose information and air pressure information. The data acquisition unit is configured to acquire multi-dimensional data from the target software driver by means of an attitude sensor and a lidar located at the end of the target software driver, thereby obtaining attitude information and three-dimensional point cloud information, wherein the attitude sensor and the lidar acquire data in time synchronized. The first generation unit is configured to generate error pose information corresponding to the target software driver based on the target pose information, the attitude information and the three-dimensional point cloud information. The second generation unit is configured to generate compensation air pressure information corresponding to the target software driver based on the error pose information. The determining unit is configured to determine the target air pressure information based on the compensated air pressure information and the predicted air pressure information; The third generation unit is configured to generate a pressure control command based on the target pressure information and the compensation pressure information, and to send the pressure control command to the pressure valve control device to control the target software driver to perform a displacement operation.

9. An electronic device, characterized in that, include: One or more processors; A storage device on which one or more programs are stored; When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1 to 6 and the system as described in claim 7.

10. A computer-readable medium, characterized in that, It stores a computer program thereon, wherein the computer program, when executed by a processor, implements the method as described in any one of claims 1 to 6 and the system as described in claim 7.

Images