Anti-deviation positioning device of transplanting mechanism

Abstract

The invention belongs to the technical field of automatic control, and discloses an anti-deviation positioning device for a transplanting mechanism, which comprises a data acquisition unit, a deviation prediction module, a self-adaptive PID control module and a deviation early warning module, the problems that in existing transplanting mechanism positioning control, PID parameter self-adaptive capacity is poor, neural network error modeling is insufficient, deviation risk prejudgment lags behind and the like are solved, and high-precision dynamic positioning and deviation risk active early warning in the transplanting process are achieved. According to the device, a PID-timing attention fusion control framework is innovatively introduced, the real-time adjustment characteristic of a traditional PID is combined with the multi-dimensional error prediction capability of a timing attention neural network, and real-time deviation correction and pre-judgment of positioning deviation are achieved through a double-branch cooperation mechanism; on the deviation prediction level, a multi-scale time sequence convolution-gating cycle unit (TCN-GRU) fusion model is constructed, multi-dimensional time sequence feature extraction is carried out on motion parameters and environmental perception data of the transplanting mechanism, and positioning deviation trends under different working conditions are accurately predicted.

Description

Technical Field

[0001] This invention relates to the field of automation control technology, and in particular to a positioning device for preventing deviation in transplanting mechanisms. Background Technology

[0002] In the field of intelligent manufacturing and automated production lines, the transfer mechanism is the core execution unit for material transfer and workpiece assembly. Its positioning accuracy directly determines the processing quality and operating efficiency of the production line. Existing transfer mechanisms (such as gantry robots, linear module transfer platforms, etc.) generally adopt a closed-loop control scheme of servo motor + grating ruler, and use traditional PID algorithm to adjust motion parameters to achieve positioning correction.

[0003] However, several technical challenges still need to be overcome in actual working conditions:

[0004] Firstly, the movement of the transfer mechanism is affected by multiple factors such as load changes (workpiece weight differences), dynamic fluctuations in the guide rail friction coefficient, and air pressure disturbances of pneumatic components. The fixed parameters of the traditional PID algorithm are difficult to adapt to the dynamic response requirements under complex working conditions, and are prone to overshoot, oscillation or steady-state error, resulting in the positioning offset exceeding the allowable range of the process.

[0005] Secondly, although some studies have attempted to introduce a single neural network model to compensate for positioning errors, the positioning errors of transplanting mechanisms are highly temporal (highly correlated with motion speed, acceleration, and dwell time) and multi-dimensionally coupled (mutual interference of X / Y / Z axis motion). Traditional fully connected neural networks or shallow CNN models have limitations in temporal feature extraction and multi-dimensional error coupling relationship modeling, and cannot achieve accurate prediction and early compensation for positioning deviations. Thirdly, existing anti-offset control strategies only passively correct the offset that has already occurred, and lack an offset trend prediction mechanism based on multi-source sensing data. When the mechanism has hidden faults such as increased mechanical backlash or degraded servo drive performance, it is impossible to identify offset risks in advance, which can easily lead to poor processing of batch workpieces and increase rework costs on the production line. Summary of the Invention

[0006] The purpose of this invention is to provide a positioning device for transplanting mechanisms to prevent deviation, thereby solving problems such as poor adaptive capability of PID parameters, insufficient neural network error modeling, and delayed prediction of deviation risks in existing positioning control of transplanting mechanisms. This device achieves high-precision dynamic positioning and proactive early warning of deviation risks during the transplanting process. The device innovatively introduces a PID-temporal attention fusion control architecture, combining the real-time adjustment characteristics of traditional PID with the multi-dimensional error prediction capability of temporal attention neural networks. Through a dual-branch collaborative mechanism, it achieves "real-time correction + early prediction" of positioning deviations. At the deviation prediction level, a multi-scale temporal convolutional-gated recurrent unit (TCN-GRU) fusion model is constructed to extract multi-dimensional temporal features from the motion parameters and environmental perception data of the transplanting mechanism, accurately predicting positioning deviation trends under different working conditions. At the parameter adjustment level, based on the coupling characteristics of predicted deviation and real-time deviation, a reinforcement learning-driven PID parameter adaptive update mechanism is adopted to dynamically optimize the proportional, integral, and derivative coefficients, significantly improving the robustness of positioning control. This solution not only improves the positioning accuracy and anti-interference capability of the transplanting mechanism under complex working conditions but also enhances the system's ability to identify and respond to implicit deviation risks.

[0007] This invention provides a positioning device for preventing displacement of a transplanting mechanism, the device comprising: a data acquisition unit, a deviation prediction module, an adaptive PID control module, and a displacement early warning module;

[0008] The data acquisition unit collects operational status data of the transplanting mechanism;

[0009] The deviation prediction module receives multi-dimensional time-series data output by the data acquisition unit, processes the data using a multi-scale TCN-GRU fusion neural network model, and generates a predicted value of the positioning offset of the transplanting mechanism.

[0010] The adaptive PID control module receives the offset prediction value output by the deviation prediction module and the real-time positioning offset output by the data acquisition unit, and generates motion control commands using a reinforcement learning-driven adaptive PID algorithm.

[0011] The offset warning module generates an offset warning signal based on the comparison between the offset prediction value output by the deviation prediction module and the preset threshold, combined with the offset risk level output by the neural network.

[0012] Furthermore, the process of implementing anti-deviation positioning of the transplanting mechanism using a PID-time attention fusion control architecture specifically includes the following steps:

[0013] Step S1 (Multi-dimensional Time Series Data Acquisition and Dataset Construction): The data acquisition unit collects multi-dimensional operational data of the transplanting mechanism in real time and constructs a time series feature dataset; the time series feature dataset includes motion state features and environmental disturbance features;

[0014] Step S2 (Predicting Positioning Offset using Multi-Scale TCN-GRU Fusion Model): The deviation prediction module uses a multi-scale TCN-GRU fusion neural network model to process the temporal feature dataset and generate predicted positioning offset values.

[0015] Step S3 (Reinforcement learning driven adaptive PID parameter optimization and control command generation): The adaptive PID control module takes the real-time positioning offset as the feedback quantity and the offset prediction value as the feedforward compensation quantity, inputs them into the reinforcement learning driven adaptive PID algorithm, dynamically updates the PID parameters and generates motion control commands, and sends them to the servo drive unit of the transplanting mechanism.

[0016] Step S4 (Offset Trend Analysis and Risk Warning Signal Output): Based on the changing trend of the offset prediction value and the risk level determination rules, the offset warning module outputs an offset warning signal to achieve proactive warning of offset risk.

[0017] Furthermore, step S2 specifically includes the following steps:

[0018] Step S21 (Time-series data preprocessing and multi-dimensional input matrix construction): Preprocess the time-series feature dataset, including data standardization, outlier removal, and time step alignment, and construct a multi-dimensional time-series input matrix; the dimensions of the input matrix include: X-axis motion velocity / acceleration, Y-axis motion velocity / acceleration, Z-axis motion velocity / acceleration, load weight, guide rail temperature, air pressure value, and real-time positioning deviation value, with a time step of 10ms;

[0019] Step S22 (Multi-scale local feature and long-term time-dependent feature extraction): For the multi-dimensional temporal input matrix, a multi-scale TCN branch is used to extract local features at different time scales (10ms, 50ms, 100ms). The receptive field is expanded through dilated convolution mechanism to capture error features at different motion stages. At the same time, a GRU branch is used to extract long-term time-dependent features to capture the error evolution law of key stages such as start-up, acceleration and deceleration of the transplanting mechanism.

[0020] Step S23 (Temporal Attention Weighted Fusion and Offset Prediction Output): A temporal attention mechanism is introduced to weight and fuse the output features of the TCN branch and the GRU branch, highlighting the feature dimensions and time stages that have the greatest impact on the localization offset, and generating a fused feature vector; the fused feature vector is mapped to the localization offset prediction value (unit: μm) through a fully connected layer, and the prediction time window is the maximum offset within the next 500ms.

[0021] Furthermore, in step S3, the process of generating motion control commands using a reinforcement learning-driven adaptive PID algorithm specifically includes the following steps:

[0022] Step B1 (Definition of PID Core Control Formula): Define the core control formula of the PID algorithm:

[0023]

[0024] Where u(t) is the control quantity output by the PID (servo motor speed adjustment value), K p K is the proportionality coefficient. i K is the integral coefficient. d Here are the differential coefficients, e(t) is the real-time positioning offset (feedback value), and u is the value of the differential coefficient. comp (t) represents the feedforward compensation amount based on the offset prediction value;

[0025] Step B2 (Construction of Reinforcement Learning Agent): Construct a reinforcement learning agent with "minimizing localization bias + no overshoot during adjustment" as the reward function. The state space includes the current K... p / K i / K d Parameters, real-time offset, predicted offset, motion velocity, motion space is K p / K i / K d The adjustment step size is (±0.01~±0.1);

[0026] Step B3 (Markov Decision Process-Driven Iterative Optimization of PID Parameters): Iteratively update PID parameters through Markov Decision Process (MDP): In each movement cycle of the transplanting mechanism, the agent selects parameters to adjust actions based on the current state, evaluates the adjustment effect through the reward function (reward value = 1 / (steady-state error + overshoot + settling time)), and updates the action value function using the Q-Learning algorithm to achieve adaptive optimization of PID parameters;

[0027] Step B4 (Substituting Optimized PID Parameters and Generating Control Commands): Substitute the optimized PID parameters into the control formula to generate a servo motor speed adjustment command, driving the transplanting mechanism to perform positioning and correction actions. Simultaneously, the feedforward compensation amount u... comp (t) is superimposed on the control quantity to offset the predicted offset in advance.

[0028] Furthermore, the working process of the offset warning module specifically includes the following steps:

[0029] Step Y1 (Offset Warning Threshold Level Preset): Preset offset threshold levels: Level 1 warning (offset predicted value ≥ 50μm), Level 2 warning (30μm ≤ offset predicted value < 50μm), Level 3 warning (10μm ≤ offset predicted value < 30μm);

[0030] Step Y2 (Risk Level Classification Neural Network Prediction): Input the offset prediction value output by the deviation prediction module into the pre-trained risk level classification neural network (Softmax classifier), and combine it with static features such as the cumulative running time of the transplanting mechanism and the wear of the guide rail to output the offset risk level;

[0031] Step Y3 (Early Warning Signal Output and Linkage Control Trigger): When the offset risk level reaches Level 1 / Level 2, the offset early warning module sends an audible and visual early warning signal to the production line central control system and triggers the emergency parameter optimization process of the adaptive PID control module; when the Level 3 early warning is reached, only the offset trend data is recorded for subsequent maintenance decisions.

[0032] By adopting the above solution, the beneficial effects achieved by the present invention are as follows:

[0033] This invention introduces a PID-time-attention fusion control architecture, combining for the first time the real-time closed-loop regulation of traditional PID with the multi-dimensional error prediction of a time-attention neural network. While retaining the advantages of fast response and strong stability of the PID algorithm, it compensates for its insufficient parameter adaptive capability. Compared with existing single PID or single neural network control methods, this architecture effectively solves the positioning offset problem of the transplanting mechanism under load changes and environmental disturbances, realizing an upgrade from "passive correction" to "active prediction + real-time adjustment", significantly improving positioning accuracy (steady-state error ≤ ±5μm) and dynamic response speed (adjustment time ≤ 200ms), laying the core foundation for the high-precision positioning control of this invention.

[0034] Furthermore, the reinforcement learning-driven adaptive PID parameter update mechanism proposed in this invention incorporates positioning accuracy, adjustment stability, and response speed into the parameter optimization objectives through a reward function. This allows the PID parameters to be dynamically adjusted according to the real-time working conditions (load, speed, environment) of the transfer mechanism, significantly enhancing the robustness of the control system under complex working conditions. Compared with traditional manual tuning or segmented PID parameters, this invention demonstrates adaptability and versatility to different workpiece transfer scenarios in actual deployment. It can accurately adjust the speed and acceleration parameters of the servo motor, significantly reducing the overshoot rate and steady-state error of the positioning offset.

[0035] In terms of offset early warning, this invention innovatively constructs a multi-scale TCN-GRU fusion model, which utilizes the multi-scale feature extraction capability of temporal convolution and the long temporal dependency capture capability of GRU, combined with a risk level classification neural network to achieve quantitative assessment of offset risk. This mechanism is applicable to the scenario of asynchronous acquisition of multi-source data and identification of implicit offset risks in the complex working conditions of the automated production line in which this invention is used. It can accurately identify and respond to early offset risks such as increased mechanical backlash and performance degradation of servo drive, providing a robust and reliable positioning safety guarantee mechanism for the production line, greatly reducing the workpiece processing defect rate caused by positioning offset, and improving the continuous operation stability and maintenance response efficiency of the transfer mechanism. Attached Figure Description

[0036] Figure 1 This is a schematic diagram of the control flow of a transplanting mechanism anti-deviation positioning device proposed in this invention;

[0037] Figure 2 This is a schematic diagram of the prediction process of the multi-scale TCN-GRU fusion model proposed in this invention;

[0038] Figure 3 This is a schematic diagram of the reinforcement learning-driven adaptive PID parameter optimization and control instruction generation process of the present invention;

[0039] Figure 4 This is a schematic diagram of the positioning deviation risk warning and linkage control process of the transplanting mechanism of the present invention. Detailed Implementation

[0040] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0041] Example 1

[0042] according to Figures 1-4 As shown, the present invention provides a positioning device for preventing displacement of a transplanting mechanism. The device includes: a data acquisition unit, a deviation prediction module, an adaptive PID control module, and a displacement early warning module.

[0043] The data acquisition unit includes a displacement sensor (grating ruler, accuracy 1μm), a speed sensor, a weight sensor, a temperature sensor, and a pressure sensor, which collects the operating status data of the transplanting mechanism. The operating status data includes: real-time displacement of the X / Y / Z axes, motion speed / acceleration, load weight (0-50kg), guide rail temperature (0-60℃), pneumatic gripper air pressure (0.4-0.8MPa), and real-time positioning offset.

[0044] The deviation prediction module receives multi-dimensional time-series data output by the data acquisition unit, processes the data using a multi-scale TCN-GRU fusion neural network model, and generates predicted values ​​for the positioning offset of the transplanting mechanism. The multi-scale TCN-GRU fusion neural network model includes: 3 multi-scale temporal convolutional layers (receptive fields of 10ms, 50ms, and 100ms respectively), 2 GRU layers (hidden layer dimension 128), 1 temporal attention layer, and 1 fully connected output layer.

[0045] The adaptive PID control module receives the predicted offset value output from the deviation prediction module and the real-time positioning offset output from the data acquisition unit. It then uses a reinforcement learning-driven adaptive PID algorithm to generate motion control commands. The reinforcement learning algorithm employs Q-Learning, with the reward function set as: Reward = 1 / (|es teady |+0.1·overshoot+0.05·t adjust ),

[0046] Where e steady The steady-state error is t, and the overshoot is the overshoot. adjust To adjust the time;

[0047] The offset warning module generates an offset warning signal based on the comparison between the offset prediction value output by the deviation prediction module and the preset threshold, combined with the offset risk level output by the neural network. The preset thresholds are: Level 1 warning (≥50μm), Level 2 warning (30-50μm), and Level 3 warning (10-30μm).

[0048] Example 2

[0049] This embodiment is based on Embodiment 1. In this embodiment, a PID-time attention fusion control architecture is used to realize the anti-deviation positioning process of the transplanting mechanism, which specifically includes the following steps:

[0050] Step S1: The data acquisition unit collects the X / Y / Z axis motion velocity (0-500mm / s) and acceleration (0-2000mm / s) of the transplanting mechanism in real time with a sampling period of 10ms. 2The following parameters were used to construct a 7-dimensional time-series feature dataset: load weight (25kg for this transfer), guide rail temperature (28℃), air pressure (0.6MPa), and real-time positioning deviation (initial deviation 18μm).

[0051] Step S2: The bias prediction module standardizes the temporal feature dataset (mean 0, variance 1) and inputs it into the multi-scale TCN-GRU fusion model; the TCN branch extracts local error features at different time scales through three layers of dilated convolution (dilation rates of 1, 2, and 4), while the GRU branch captures long-term temporal dependency features within 100ms; the temporal attention layer assigns weights to each feature dimension (velocity feature weight 0.35, load feature weight 0.25, temperature feature weight 0.15, and other features weight 0.25), and outputs a fused feature vector; the fully connected layer maps the fused feature vector to the predicted positioning offset value (22μm) within the next 500ms;

[0052] Step S3: The adaptive PID control module uses the real-time positioning offset (18μm) as the feedback value and the offset prediction value (22μm) as the feedforward compensation value, and inputs them into the reinforcement learning-driven adaptive PID algorithm; the initial PID parameters are K. p =0.8, K i =0.15, K d =0.08, the reinforcement learning agent adjusts the parameters to K based on the current state (offset 18μm, predicted value 22μm, motion speed 300mm / s). p =0.85, K i =0.12, K d =0.09; Substitute into the PID formula to generate the control quantity:

[0053]

[0054] (feedforward compensation amount u) comp (t)=5), generate a servo motor speed adjustment command (speed increases by 50r / min), drive the transplanting mechanism to perform correction action;

[0055] Step S4: The offset warning module inputs the offset prediction value (22μm) into the risk level classification neural network, outputs a three-level warning level, records only the offset trend data, and does not trigger an audible and visual warning.

[0056] Example 3

[0057] This embodiment is based on Embodiment 2. In this embodiment, the control effects of traditional PID and the adaptive PID of the present invention are compared:

[0058] Traditional PID parameters (manual tuning K) p =0.8, Ki =0.15, K d =0.08): Overshoot during positioning is 8μm, settling time is 300ms, and steady-state error is 5μm;

[0059] The adaptive PID parameters (K) of this invention p =0.85, K i =0.12, K d =0.09): Overshoot during positioning is 1μm, settling time is 150ms, and steady-state error is 2μm;

[0060] The results show that the adaptive PID control scheme of the present invention significantly reduces overshoot and settling time, and improves positioning accuracy.

[0061] Example 4

[0062] This embodiment is based on embodiment two. In this embodiment, the data acquisition unit collects the positioning offset data of the transplanting mechanism from 15 seconds to 15.1 seconds and generates a time-series change diagram of the positioning offset.

[0063] The horizontal axis represents the sampling time point, with a sampling period of 10ms, in the following order: 15 seconds 000 milliseconds (15s:000ms), 15 seconds 010 milliseconds (15s:010ms) ... 15 seconds 100 milliseconds (15s:100ms);

[0064] The vertical axis represents the values ​​of two key parameters: real-time positioning offset (unit: μm, blue line) and positioning offset after PID adjustment (unit: μm, green line).

[0065] The red shaded area (15s:010ms-15s:040ms) represents the traditional PID overshoot range, within which the real-time offset can reach a maximum of 25μm.

[0066] The green shaded area (15s:010ms-15s:040ms) is the adaptive PID adjustment range of this invention. Within this range, the maximum offset after adjustment is only 8μm, and it converges to 2μm after 15s:050ms, with no steady-state error.

[0067] The present invention and its embodiments have been described above. This description is not restrictive. The accompanying drawings are only one embodiment of the present invention, and the actual control logic is not limited thereto. In short, if a person skilled in the art is inspired by this description and designs a control architecture and embodiment similar to this technical solution without departing from the spirit of the present invention, such design should fall within the protection scope of the present invention.

Claims

1. A positioning device for preventing displacement of a transplanting mechanism, characterized in that: It includes a data acquisition unit, a deviation prediction module, an adaptive PID control module, and a deviation early warning module; The data acquisition unit is used to collect operating status data of the transplanting mechanism; The deviation prediction module receives multi-dimensional time-series data output by the data acquisition unit, processes the multi-dimensional time-series data using a multi-scale TCN-GRU fusion neural network model, and generates a predicted value of the positioning offset of the transplanting mechanism. The adaptive PID control module receives the offset prediction value output by the deviation prediction module and the real-time positioning offset output by the data acquisition unit, and generates motion control commands using a reinforcement learning-driven adaptive PID algorithm. The offset warning module generates an offset warning signal based on the comparison between the offset prediction value output by the deviation prediction module and a preset threshold, combined with the offset risk level output by the neural network.

2. The anti-deviation positioning device for a transplanting mechanism according to claim 1, characterized in that: A PID-time attention fusion control architecture is used to achieve anti-deviation positioning of the transplanting mechanism. This process includes the following steps: Step S1: The data acquisition unit collects multi-dimensional operational data of the transplanting mechanism in real time to construct a time-series feature dataset; the time-series feature dataset includes motion state features and environmental disturbance features; Step S2: The deviation prediction module uses a multi-scale TCN-GRU fusion neural network model to process the temporal feature dataset and generate a predicted positioning offset value. Step S3: The adaptive PID control module takes the real-time positioning offset as the feedback quantity and the offset prediction value as the feedforward compensation quantity, inputs them into the reinforcement learning-driven adaptive PID algorithm, dynamically updates the PID parameters and generates motion control commands, and sends them to the servo drive unit of the transplanting mechanism. Step S4: The offset warning module outputs an offset warning signal based on the changing trend of the offset prediction value and the risk level determination rules, thereby realizing proactive warning of offset risk.

3. The anti-deviation positioning device for a transplanting mechanism according to claim 2, characterized in that: Step S2 specifically includes the following steps: Step S21: Preprocess the time-series feature dataset, including data standardization, outlier removal, and time step alignment, to construct a multi-dimensional time-series input matrix; the dimensions of the multi-dimensional time-series input matrix include X-axis motion velocity / acceleration, Y-axis motion velocity / acceleration, Z-axis motion velocity / acceleration, load weight, guide rail temperature, air pressure value, and real-time positioning deviation value, with a time step of 10ms; Step S22: For the multi-dimensional temporal input matrix, a multi-scale TCN branch is used to extract local features at different time scales, the receptive field is expanded through a dilated convolution mechanism, and a GRU branch is used to extract long temporal dependency features. Step S23: Introduce a temporal attention mechanism to perform weighted fusion of the output features of the TCN branch and the GRU branch to generate a fused feature vector, and map the fused feature vector to the localization offset prediction value through a fully connected layer.

4. The anti-deviation positioning device for a transplanting mechanism according to claim 3, characterized in that: The local feature time scales extracted by the multi-scale TCN branch are 10ms, 50ms, and 100ms; the predicted positioning offset value is the maximum offset within the next 500ms, in μm.

5. The anti-deviation positioning device for a transplanting mechanism according to claim 2, characterized in that: The process of generating motion control commands using a reinforcement learning-driven adaptive PID algorithm in step S3 specifically includes the following steps: Step B1: Define the core control formula of the PID algorithm. The output of the control formula is the servo motor speed adjustment value. The input of the control formula includes the real-time positioning offset and the feedforward compensation based on the offset prediction value. Step B2: Construct a reinforcement learning agent with "minimizing localization bias + no overshoot during adjustment" as the reward function. The state space includes the current PID parameters, real-time offset, predicted offset value, and movement speed. The action space is the adjustment step size of the PID parameters. Step B3: The PID parameters are iteratively updated through a Markov decision process. In each movement cycle of the transplanting mechanism, the agent selects parameters to adjust the action according to the current state, evaluates the adjustment effect through a reward function, and updates the action value function using a Q-Learning algorithm. Step B4: Substitute the optimized PID parameters into the control formula to generate the servo motor speed adjustment command, and simultaneously add the feedforward compensation.

6. The anti-deviation positioning device for a transplanting mechanism according to claim 5, characterized in that: The adjustment step size of the PID parameters is ±0.01 to ±0.1; the reward value of the reward function is 1 / (steady-state error + overshoot + settling time).

7. The anti-deviation positioning device for a transplanting mechanism according to claim 1, characterized in that: The working process of the offset early warning module specifically includes the following steps: Step Y1: Preset offset threshold levels. Level 1 warning is when the offset prediction value is ≥50μm, Level 2 warning is when 30μm≤offset prediction value<50μm, and Level 3 warning is when 10μm≤offset prediction value<30μm. Step Y2: Input the offset prediction value output by the deviation prediction module into the pre-trained risk level classification neural network, and combine it with static features such as the cumulative running time of the transplanting mechanism and the wear of the guide rail to output the offset risk level; Step Y3: When the offset risk level reaches Level 1 / Level 2, the offset early warning module sends an audible and visual early warning signal and triggers the emergency parameter optimization process of the adaptive PID control module; when the Level 3 early warning is reached, only the offset trend data is recorded.

8. The anti-deviation positioning device for a transplanting mechanism according to claim 1, characterized in that: The data acquisition unit includes a displacement sensor, a velocity sensor, a weight sensor, a temperature sensor, and a pressure sensor. The displacement sensor is a grating ruler with an accuracy of 1μm; the operating status data includes real-time displacement of the X / Y / Z axes, motion speed / acceleration, load weight, guide rail temperature, pneumatic gripper air pressure, and real-time positioning offset.

Images