An intelligent end-pickup for photovoltaic module paving

Abstract

This invention discloses an intelligent end effector for photovoltaic module installation, comprising an end effector body, a vacuum adsorption system, a laser ranging and positioning system, a displacement detection system, and a controller. The laser ranging and positioning system is used for non-contact measurement of the relative pose between the end effector body and the photovoltaic module surface. The controller is configured to execute the following control flow: receive and process distance data; determine the current pose parameters of the photovoltaic module surface through spatial geometric calculations; generate instructions based on these parameters to drive the end effector body to move, adjusting the pose to an allowable range; control the end effector to perform a bonding action, and monitor the displacement of all suction cups in real time through the displacement detection system; when the displacement of all suction cups reaches the bonding displacement threshold and stabilizes, bonding is determined to be complete; and the vacuum adsorption system is activated to complete the gripping. This invention achieves automatic high-precision alignment and reliable bonding judgment before photovoltaic module gripping, significantly improving the gripping success rate and operational safety.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic module installation equipment technology, and more specifically to an intelligent end-feed device for photovoltaic module installation. Background Technology

[0002] Photovoltaic power generation is a crucial component of clean energy, and its large-scale development has created an urgent need for automated installation technologies. Currently, the installation of photovoltaic modules is gradually shifting from manual methods to automated methods using installation robots. Installation robots typically require end effectors to pick up and transport photovoltaic modules. However, traditional end effectors used in fixed scenarios such as palletizing have simple structures, usually only possessing fixed vacuum adsorption functions and a single repetitive motion trajectory, lacking the ability to intelligently sense and adjust the posture of the work object (photovoltaic modules) in real time. At the photovoltaic module installation site, due to potential deviations in module placement, non-perfectly level surfaces, and extremely high requirements for stability and alignment accuracy during the gripping process, traditional end effectors struggle to achieve reliable and precise automatic gripping, easily leading to module collisions, unstable adsorption, or even damage, severely hindering the improvement of automation levels and operational safety. Therefore, there is an urgent need for an end effector capable of intelligently sensing the posture of photovoltaic modules and automatically adjusting alignment to ensure a stable and reliable gripping process, in order to meet the high requirements of automated photovoltaic module installation. Summary of the Invention

[0003] In view of the above problems, the present invention provides an intelligent end-feeder for photovoltaic module installation, which aims to solve the technical problems of low installation efficiency and easy damage to photovoltaic modules caused by the lack of intelligent sensing and adaptive control capabilities of existing conventional end-feeders in photovoltaic module installation scenarios, which makes them unable to automatically and accurately align and reliably grasp photovoltaic modules.

[0004] To achieve the above objectives, in a first aspect, the present invention provides an intelligent end-feeder for photovoltaic module installation, comprising: End effector body; A vacuum adsorption system includes multiple suction cups, which are mounted on the end effector body and are used to adsorb photovoltaic modules. A laser ranging and positioning system is installed on the end effector body and is used for non-contact measurement of the relative pose between the end effector body and the surface of the photovoltaic module. A displacement detection system, installed on the end effector body, is used to detect the displacement of the suction cup relative to the end effector body in order to determine the adhesion status between the suction cup and the photovoltaic module. The controller is electrically connected to the vacuum adsorption system, the laser ranging and positioning system, and the displacement detection system, respectively. The laser ranging and positioning system includes: At least four laser ranging sensors are deployed on the end effector body to measure the distance from the laser ranging sensors to different measurement points on the surface of the photovoltaic module, thereby obtaining distance measurement signals. The signal processing unit is used to receive and process the distance measurement signals of each of the laser ranging sensors to obtain distance data; The controller is configured to execute the following control flow: S1: Receive distance data sent by the signal processing unit; S2: Based on the fixed installation coordinates of each laser ranging sensor in the coordinate system of the end effector body and their corresponding distance data, the current pose parameters of the photovoltaic module surface relative to the end effector body are determined through spatial geometric calculation. The current pose parameters include at least the center position deviation and the tilt angle. S3: Generate a pose adjustment command based on the current pose parameters to drive the end effector body to move, so as to reduce the center position deviation and tilt angle until the center position deviation and tilt angle reach the preset pose tolerance range. S4: After the current pose parameters reach the pose tolerance range, control the end effector body to perform the bonding action, and monitor the displacement of the suction cup in real time through the displacement detection system. When the displacement of all the suction cups reaches the preset bonding displacement threshold, and maintains a preset stable time after reaching the threshold, it is determined that the bonding is complete. S5: After the bonding is determined, the vacuum adsorption system is started to complete the gripping of the photovoltaic module.

[0005] Furthermore, the vacuum adsorption system includes at least two independent adsorption control units; The plurality of suction cups are assigned to each of the adsorption control units, such that each adsorption control unit controls at least one of the suction cups; Each of the aforementioned adsorption control units further includes: at least one vacuum pump and at least one vacuum tank connected to the outlet of the vacuum pump; Each suction cup belonging to the same set of adsorption control unit is connected to the vacuum tank through an independent vacuum branch. Each vacuum branch is provided with a vacuum filter and a solenoid valve in sequence along the vacuum flow direction. The controller is electrically connected to the vacuum pump in each of the adsorption control units and the solenoid valve on each of the vacuum branches, and is used to independently control the start and stop of each of the adsorption control units and to independently control the on and off of each of the vacuum branches.

[0006] Furthermore, the displacement detection system includes a sensor switch and a sensor plate disposed on each of the suction cups; The suction cup is slidably mounted on the end effector body via a buffer rod, and the sensing plate is disposed on the buffer rod; When the suction cup contacts and is pressed against the surface of the photovoltaic module, it moves the buffer rod and the sensing sheet to trigger the inductive switch, so that the inductive switch generates a signal indicating that the suction cup is in place.

[0007] Furthermore, the displacement detection system also includes a micro grating ruler integrated inside each of the buffer rods, the micro grating ruler being used to measure the absolute compressive displacement d of the buffer rod relative to its mounting base; The controller is configured to perform a two-level fusion judgment in step S4, specifically including: When the sensor switch of any suction cup is triggered, it is determined to be a primary fit, and the reading of the laser rangefinder corresponding to that suction cup at the current moment is recorded as the reference distance value L0. After the initial bonding, the controller switches to reading the data of the micro grating ruler and continues to control the end effector body to perform micro-pressure. During the micro-pressure process, the following calculations are performed in real time for each suction cup: Obtain the reading L of the laser rangefinder at the current moment. t; Calculate the normal displacement δ of the photovoltaic module surface derived from laser measurements. laser The calculation formula is as follows: δ laser =L0-L t ; Obtain the compression displacement d of the buffer rod measured by the micro grating ruler at the current moment, and calculate the coupling difference Δ=δ according to the following formula. laser -d; The final fine bonding is considered complete when both of the following conditions are met: The compression displacement d of the buffer rod of all suction cups reaches the preset fine fit displacement threshold. The absolute values ​​of the coupling differences Δ calculated by all suction cups are all less than the preset deformation compatibility threshold ε; the deformation compatibility threshold ε is used to determine whether the photovoltaic module has undergone local elastic deformation beyond the permissible range or has poor contact.

[0008] Furthermore, the end effector body is a rectangular frame structure, which includes at least two parallel longitudinal aluminum profiles and multiple transverse aluminum profiles connected between the two parallel longitudinal aluminum profiles. The aluminum profiles are connected by corner brackets. At least four laser ranging sensors of the laser ranging and positioning system are respectively installed in the four corner areas of the rectangular frame, and are used to measure the vertical distance between the sensor and the four corner areas. The vacuum adsorption system has multiple suction cups mounted on the horizontal aluminum profile via their respective buffer rods. The horizontal aluminum profile is provided with a fixing seat for mounting and adjusting the buffer rods. The signal processing unit filters and denoises the received distance measurement signal, and sends the filtered and denoised distance data to the controller for spatial geometry calculation.

[0009] Furthermore, the controller calculates the current pose parameters in the following manner: Based on the fixed installation coordinates of each laser ranging sensor in the coordinate system of the end effector and the distance data measured by them, the three-dimensional coordinates of at least four measuring points on the surface of the photovoltaic module in the coordinate system of the end effector are calculated. Based on the three-dimensional coordinates of the at least four measurement points, the plane equation of the photovoltaic module surface is fitted. Based on the plane equation, the center position deviation and tilt angle of the photovoltaic module surface are calculated.

[0010] Furthermore, the intelligent end effector also includes: An environmental sensing module, integrated on the end effector body, is used to acquire environmental wind field information and motion state information of the end effector body in real time. A drive actuator, connected to the controller, is used to drive the end effector body to move according to the instructions of the controller; The controller is further configured to synchronously perform feedforward wind load dynamic compensation control when executing step S3 and / or step S4, specifically including: The environmental sensing module acquires in real time the ambient wind speed V acting on the end effector body and the adsorbed photovoltaic module. w Wind direction angle θ w , and the triaxial angular velocity and acceleration of the end effector body; Based on the tilt angle in the current pose parameters and the ambient wind speed V w With wind direction angle θ w And a pre-existing aerodynamic parameter model of the photovoltaic module in the controller, to calculate in real time the predicted normal force F generated by wind load in the normal direction of the photovoltaic module plane. wind and the predicted normal force F wind The predicted moment vector M in the component plane wind The aerodynamic parameter model includes at least the windward area of ​​the photovoltaic module, the location of the wind pressure center, and the aerodynamic coefficient. The controller is based on the predicted torque vector M wind， Based on the dynamic model of the drive actuator, the feedforward compensation algorithm built into the controller is used to calculate the compensation method for M. windRequired compensation torque command ΔM comp ; In step S3 and / or step S4, the final control command received by the drive actuator is: the basic control command and the compensation torque command ΔM. comp The superposition of the basic control commands is generated by the pose adjustment algorithm in step S3 or the control logic that generates the fitting control commands in step S4.

[0011] Furthermore, the feedforward compensation algorithm calculates the compensation torque command ΔM through the following steps. comp : The predicted torque vector M wind The transformed disturbance torque vector M is obtained by transforming the coordinate system based on the photovoltaic module plane to the joint space coordinate system or the drive motor coordinate system of the drive actuator. wind' ; Based on the dynamic model of the drive actuator, the torque vector M to counteract the disturbance is calculated. wind' The required joint drive torque or motor current command is used as an initial compensation amount. The initial compensation amount is filtered to match the frequency band characteristics of the filtered initial compensation amount with the actual response bandwidth of the drive actuator, and then multiplied by an adjustable feedforward gain coefficient K. ff Generate the compensation torque command ΔM comp K ff The value range is (0,1).

[0012] Furthermore, the controller also integrates a distributed fault diagnosis and fault-tolerant control module, which is configured to perform the following steps: For the i-th laser rangefinder, its instantaneous measurement value is defined as d. i (t), within the time window T w Calculate its rate of change The difference Δd between adjacent periods i ; The data validity flag of the laser rangefinder sensor shall be set to invalid if any of the following conditions are met: d i (t) remains equal to zero, has a maximum value, or a specific error code exceeds N. err One sampling period; | |>V max , where V max The limit distance change rate is calculated based on the maximum speed of the end effector body; ∣Δd i |>ΔDmax , where ΔD max This is the maximum reasonable jump threshold set based on the mechanical vibration characteristics; If the valid data flag of a certain laser rangefinder remains invalid for a duration exceeding the first preset time threshold T, fault If the laser rangefinder is determined to be in a hard failure state, it will be removed from the set of currently valid laser rangefinders S. valid Remove from; Let the total number of laser rangefinders be N, and define the effective set of laser rangefinders as S. valid The number of elements is M; If M≥4, then directly use S valid Installation coordinates (x) of all laser rangefinders i ,y i ,z i ) body and its measured value d i By fitting the plane using the least squares method or using the SVD decomposition method, the equation ax+by+cz+d=0 of the photovoltaic module plane in the coordinate system of the end-sensor is solved, and the tilt angle and the deviation of the center position are obtained. If M=3, a spatial plane is determined using data from three effective laser rangefinders. Combined with the encoder feedback value of the drive actuator, the expected pose of the end effector body is obtained through forward kinematics calculation. The expected pose is then fused with the laser plane equation, and the complete current pose parameters are estimated using a Kalman filter. If M < 3, it is determined that the pose cannot be reliably calculated at present, the operation is immediately suspended and a fault alarm message indicating that manual intervention is required is issued.

[0013] Furthermore, during the bonding process in step S4, for the j-th suction cup, the displacement sensor measurement value of that suction cup is defined as δ. j (t); Calculate the average value of all suction cup displacements. and standard deviation δ(t); If |δ j (t)- |>k·δ(t) and continuously exceeds the second preset time threshold T abn If the displacement of suction cup j is abnormal, then k is determined to be a preset statistical outlier coefficient. When suction cup j is determined to be abnormal, the following control parameter reconstruction steps are performed: The target pressure value P of the vacuum generator corresponding to suction cup j. target,j Adjust to P target,j The formula is adjusted as follows, P target,j '=α·P target,jWhere 0 < α < 1 is the pressure attenuation coefficient; Generate the speed command that drives the end effector body to perform the contact action. At that time, a velocity attenuation field is applied to the local region corresponding to suction cup j, so that the normal velocity v of this local region in the body coordinate system of the end effector approaches the velocity v. n,j Satisfy v n,j =β·v n,default Where β is the velocity decay coefficient, 0 < β < 1; The suction cup array is abstracted as an undirected graph G=(V,E), where vertex V represents the suction cup and edge E represents the adjacency relationship; Once suction cup j is determined to be abnormal, iterate through all its adjacent suction cup sets N(j); For each adjacent suction cup p∈N(j), the self-adhesion displacement threshold δ is reached at that adjacent suction cup. th Then, an additional compensation displacement δ is applied. comp,p The calculation formula is as follows: ; Where γ is the global compensation gain. Let (j,p) be the weight of edge (j,p). This is the theoretical displacement value of the suction cup j estimated based on the current pose parameters; In step S4, the controller adjusts the condition for determining the completion of adhesion of the normal suction cup p to: δ p (t)≥δ th +δ comp,p And it maintains a stable duration.

[0014] Unlike existing technologies, the above technical solution provides an intelligent end effector for photovoltaic module installation, including an end effector body, a vacuum adsorption system, a laser ranging and positioning system, a displacement detection system, and a controller. The laser ranging and positioning system includes at least four laser ranging sensors and a signal processing unit for non-contact measurement of the relative pose between the end effector body and the photovoltaic module surface. The controller is configured to execute the following control flow: first, it receives and processes distance data, and determines the current pose parameters of the photovoltaic module surface (including center position deviation and tilt angle) through spatial geometric calculation; then, it generates instructions based on these parameters to drive the end effector body to move, adjusting the pose to an allowable range; next, it controls the end effector to perform a bonding action, and monitors the displacement of all suction cups in real time through the displacement detection system. When the displacement of all suction cups reaches the bonding displacement threshold and stabilizes, bonding is considered complete; finally, the vacuum adsorption system is activated to complete the gripping. This invention achieves automatic high-precision alignment and reliable bonding judgment before photovoltaic module gripping, significantly improving the gripping success rate and operational safety, and is suitable for automated photovoltaic module installation scenarios.

[0015] The above description of the invention is merely an overview of the technical solution of the present invention. In order to enable those skilled in the art to better understand the technical solution of the present invention and to implement it based on the description and drawings, and to make the above-mentioned objectives and other objectives, features and advantages of the present invention easier to understand, the following description is provided in conjunction with the specific embodiments and drawings of the present invention. Attached Figure Description

[0016] The accompanying drawings are only used to illustrate the principles, implementation methods, applications, features, and effects of specific embodiments of the present invention and other related contents, and should not be considered as limitations on the present invention.

[0017] In the accompanying drawings of the instruction manual: Figure 1 This is a schematic diagram of the first structure of the intelligent end effector described in a specific embodiment; Figure 2 This is a schematic diagram of the second structure of the intelligent end effector described in a specific embodiment; Figure 3 This is a schematic diagram of the third structure of the intelligent end effector described in a specific implementation; Figure 4 This is a schematic diagram of the fourth structure of the intelligent end effector described in a specific implementation; Figure 5 This is a schematic diagram of the fifth structure of the intelligent end effector described in a specific implementation; Figure 6 This is a first flowchart illustrating the method steps executed by the controller in a specific implementation embodiment; The reference numerals used in the above figures are explained as follows: 1. End effector body; 11. Horizontal aluminum profile; 12. Vertical aluminum profile; 13. Angle bracket; 14. Suction cup holder; 15. Fixing connection plate; 21. Buffer rod; 22. Vacuum filter; 23. Solenoid valve; 24. Pressure gauge; 25. Vacuum tank; 26. Vacuum pump; 27. Controller; 28. Photovoltaic module; 31. Laser rangefinder sensor; 32. Signal processing unit; 16. Laser mounting bracket. Detailed Implementation

[0018] To illustrate the possible application scenarios, technical principles, implementable specific solutions, and achievable objectives and effects of this invention in detail, the following description, in conjunction with the listed specific embodiments and accompanying drawings, provides a detailed explanation. The embodiments described herein are merely illustrative of the technical solutions of this invention and are therefore intended only as examples, not as limiting the scope of protection of this invention.

[0019] In this document, the term "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The term "embodiment" appearing in various places throughout the specification does not necessarily refer to the same embodiment, nor does it specifically limit its independence or connection with other embodiments. In principle, in this invention, as long as there are no technical contradictions or conflicts, the technical features mentioned in each embodiment can be combined in any way to form corresponding implementable technical solutions.

[0020] Unless otherwise defined, the technical terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains; the use of related terms herein is merely for the purpose of describing particular embodiments and is not intended to limit the invention.

[0021] In the first aspect, such as Figure 1-5 As shown, the present invention provides an intelligent end-feed device for photovoltaic module installation, comprising: End effector body 1; A vacuum adsorption system includes multiple suction cups, which are mounted on the end effector body and are used to adsorb photovoltaic modules. A laser ranging and positioning system is installed on the end effector body and is used for non-contact measurement of the relative pose between the end effector body and the surface of the photovoltaic module. A displacement detection system, installed on the end effector body, is used to detect the displacement of the suction cup relative to the end effector body in order to determine the adhesion status between the suction cup and the photovoltaic module. The controller is electrically connected to the vacuum adsorption system, the laser ranging and positioning system, and the displacement detection system, respectively. The laser ranging and positioning system includes: At least four laser ranging sensors are deployed on the end effector body to measure the distance from the laser ranging sensors to different measurement points on the surface of the photovoltaic module, thereby obtaining distance measurement signals. Signal processing unit 32 is used to receive and process the distance measurement signals of each of the laser ranging sensors to obtain distance data; like Figure 6 As shown, the controller is configured to execute the following control flow: S1: Receive distance data sent by the signal processing unit; S2: Based on the fixed installation coordinates of each laser ranging sensor in the coordinate system of the end effector body and their corresponding distance data, the current pose parameters of the photovoltaic module 28 surface relative to the end effector body are determined through spatial geometric calculation. The current pose parameters include at least the center position deviation and the tilt angle. S3: Generate a pose adjustment command based on the current pose parameters to drive the end effector body to move, so as to reduce the center position deviation and tilt angle until the center position deviation and tilt angle reach the preset pose tolerance range. S4: After the current pose parameters reach the pose tolerance range, control the end effector body to perform the bonding action, and monitor the displacement of the suction cup in real time through the displacement detection system. When the displacement of all the suction cups reaches the preset bonding displacement threshold, and maintains a preset stable time after reaching the threshold, it is determined that the bonding is complete. S5: After the bonding is determined, the vacuum adsorption system is started to complete the gripping of the photovoltaic module.

[0022] In this embodiment, the end effector body 1 is the basic support of the intelligent end effector. It is constructed with aluminum profiles to form a rectangular frame structure, which serves as the installation foundation for the vacuum adsorption system, laser ranging and positioning system, displacement detection system, and controller. Its structural strength and installation accuracy provide physical protection for the coordinated operation of each system.

[0023] The vacuum adsorption system is the core execution system for grasping photovoltaic modules. It uses multiple suction cups to generate negative pressure to adsorb and release photovoltaic modules, providing force for the handling of photovoltaic modules.

[0024] The laser ranging and positioning system is a non-contact pose detection system based on the principle of laser ranging. It acquires distance data through a laser ranging sensor, and after processing, it realizes accurate measurement of the relative pose between the end effector body and the surface of the photovoltaic module. The core is non-contact measurement, which avoids premature contact with the photovoltaic module and damage.

[0025] The displacement detection system is a sensing system used to detect the displacement of the suction cup relative to the pickup body. The displacement data is used to determine the degree of adhesion between the suction cup and the photovoltaic module, which is a preliminary detection step to achieve precise adsorption.

[0026] The controller is the core control unit of the intelligent end effector. It can be constructed by combining a PLC with the controller 27. It is electrically connected to the vacuum adsorption system, the laser ranging and positioning system, and the displacement detection system, respectively. It receives the sensing data of each system and outputs control commands to realize the coordinated linkage and automated control of each system.

[0027] Pose parameters refer to parameters used to characterize the spatial position and attitude of the photovoltaic module surface relative to the end effector body. They include at least the center position deviation (the spatial offset between the central axis of the end effector body and the central axis of the photovoltaic module) and the tilt angle (the spatial angle between the suction cup plane of the end effector body and the surface of the photovoltaic module, including the horizontal and vertical directions).

[0028] The permissible position range refers to the preset threshold range of center position deviation and tilt angle to ensure effective adhesion between the suction cup and the photovoltaic module. When the parameters are within this range, the relative position of the end effector body and the photovoltaic module meets the requirements of the adsorption operation.

[0029] The adhesion displacement threshold refers to the minimum displacement of the suction cup relative to the end-feeder body when the suction cup is fully adhered to the photovoltaic module. Reaching this threshold indicates that the suction cup has achieved physical contact with the surface of the photovoltaic module and is properly adhered.

[0030] The stabilization time refers to the time required after the suction cup displacement reaches the bonding displacement threshold. It is used to avoid misjudgment of bonding caused by factors such as instantaneous contact and mechanical vibration, and to ensure the stability of the bonding state.

[0031] The core grasping control process of the intelligent end effector in this embodiment is uniformly scheduled and executed by the controller. Each step is interconnected, realizing automated closed-loop control from pose detection and adjustment to adhesion and adsorption. The detailed working principle is as follows: In step S1, at least four laser ranging sensors of the laser ranging and positioning system continuously collect the distances from themselves to different measurement points on the surface of the photovoltaic module. The collected distance measurement signals are transmitted to the signal processing unit. After the signal processing unit completes the preliminary processing of the signals, it sends the effective distance data to the controller for subsequent pose calculation.

[0032] In step S2, the controller pre-stores the fixed installation coordinates of all laser ranging sensors in the coordinate system of the end effector body. Combined with the distance data corresponding to each sensor received, the distance data is converted with the installation coordinates through spatial geometric calculation methods to obtain the three-dimensional coordinates of each measurement point on the surface of the photovoltaic module in the coordinate system of the end effector body. Then, based on the three-dimensional coordinates, the actual plane parameters of the photovoltaic module surface are fitted, and finally the current pose parameters of the photovoltaic module surface relative to the end effector body are calculated, including the center position deviation and tilt angle, so as to achieve accurate determination of the relative pose of the two.

[0033] In step S3, the controller compares the calculated current pose parameters with the preset pose tolerance range, generates a corresponding pose adjustment command based on the magnitude and direction of the deviation, and sends the command to the drive actuator of the end effector. The drive actuator drives the end effector body to perform translation, rotation, tilt and other movements according to the command, gradually reducing the center position deviation and tilt angle until both parameters fall within the preset pose tolerance range. At this time, the suction cup plane of the end effector body is basically parallel to the surface of the photovoltaic module and the central axis is basically coincident, which meets the pose requirements of the subsequent bonding operation.

[0034] In step S4, after the pose parameters reach the allowable pose range, the controller sends a bonding action command to the drive actuator, controlling the end effector body to press down towards the photovoltaic module to perform the bonding action; at the same time, the displacement detection system monitors the displacement of each suction cup relative to the end effector body in real time, and the controller performs real-time judgment on the displacement data of all suction cups. When the displacement of all suction cups reaches the preset bonding displacement threshold, and this state is maintained for a preset stable time, it is determined that the suction cup and the photovoltaic module have completed bonding, avoiding invalid bonding situations such as single suction cup bonding and instantaneous bonding, and ensuring the comprehensiveness and stability of bonding.

[0035] In step S5, after determining that the bonding is complete, the controller immediately sends a start command to the vacuum adsorption system. The vacuum adsorption system starts working, creating a negative pressure between the suction cup and the photovoltaic module, thus completing the stable gripping of the photovoltaic module and laying the foundation for subsequent photovoltaic module handling and installation operations.

[0036] The above solution uses a laser ranging and positioning system to measure the relative pose of the end effector and the photovoltaic module. It does not require prior contact with the photovoltaic module, effectively avoiding scratches and collision damage to the photovoltaic module that may be caused by contact detection. At the same time, the laser measurement method has a fast response speed and high accuracy, which is suitable for the automated operation requirements of photovoltaic module installation.

[0037] By generating spatial geometry calculations and pose adjustment commands from the controller, the pose adjustment of the end effector body can be completed without manual intervention, which greatly reduces the intensity of manual operation and solves the problems of low efficiency and poor accuracy of manual positioning of traditional end effectors, thus significantly improving the efficiency of photovoltaic module handling.

[0038] The displacement detection system monitors the displacement of all suction cups in real time and sets dual judgment conditions of adhesion displacement threshold and stabilization time to ensure that all suction cups are properly adhered to the surface of the photovoltaic module. This avoids negative pressure leakage and uneven force on the photovoltaic module during subsequent adsorption due to insufficient adhesion, thus providing a prerequisite guarantee for the stability of vacuum adsorption.

[0039] Achieving automated closed-loop control: From pose detection and adjustment to fit determination and adsorption start-up, the entire process is uniformly controlled by the controller. All systems work together to form an automated closed loop, eliminating the need for manual intervention in core operations, improving operational convenience, and effectively avoiding subjective errors from manual operation, thus ensuring the consistency and stability of the grasping operation.

[0040] The combination of non-contact laser measurement and automated control enables the end effector to complete the grasping operation without direct human observation of the photovoltaic module. Even in complex outdoor environments of photovoltaic power stations, such as strong light, backlight, and dust, it can still stably complete the posture detection and grasping, solving the problem of poor environmental adaptability of traditional end effectors.

[0041] In some embodiments, the vacuum adsorption system includes at least two independent adsorption control units; The plurality of suction cups are assigned to each of the adsorption control units, such that each adsorption control unit controls at least one of the suction cups; Each of the adsorption control units further includes: at least one vacuum pump 26 and at least one vacuum tank 25 connected to the outlet of the vacuum pump; Each suction cup belonging to the same set of adsorption control unit is connected to the vacuum tank through an independent vacuum branch. Each vacuum branch is provided with a vacuum filter 22 and a solenoid valve 23 in sequence along the vacuum flow direction. The controller is electrically connected to the vacuum pump in each of the adsorption control units and the solenoid valve on each of the vacuum branches, and is used to independently control the start and stop of each of the adsorption control units and to independently control the on and off of each of the vacuum branches.

[0042] In some embodiments, the adsorption control unit refers to an independent control subunit of the vacuum adsorption system. Each unit contains complete vacuum generation, transmission, and control components, and can independently complete the negative pressure adsorption and release of the suction cup. Each unit is independent of the others, and there is no cross-interference between vacuum paths and control signals.

[0043] A vacuum tank is a pressure vessel used to store vacuum. It is connected to the outlet of a vacuum pump and can stabilize the negative pressure generated by the vacuum pump, providing a continuous and stable vacuum source for the suction cup and avoiding negative pressure fluctuations caused by frequent start-stop of the vacuum pump.

[0044] A vacuum branch refers to an independent vacuum transmission line that connects the vacuum tank to a single suction cup. Each branch provides vacuum for only one suction cup, enabling independent control of the suction cup vacuum supply.

[0045] A vacuum filter is a filter component installed on a vacuum branch line, arranged along the direction of vacuum flow. It is used to filter dust, impurities, etc. in the vacuum pipeline, preventing impurities from entering the suction cup or vacuum pump, and avoiding suction cup seal failure and vacuum pump damage.

[0046] A solenoid valve is an on / off control component installed on a vacuum branch. It is arranged behind the vacuum filter along the vacuum flow direction and is independently controlled by a controller to realize the vacuum supply and cut-off of the corresponding vacuum branch, thereby controlling the adsorption and release of a single suction cup.

[0047] In this embodiment, the vacuum adsorption system adopts a design with multiple independent adsorption control units, and multiple suction cups are reasonably distributed to each unit. Each unit controls at least one suction cup, forming a redundant vacuum adsorption structure. At the same time, each vacuum branch and each adsorption control unit are independently controlled by a controller. The detailed working principle is as follows: The vacuum adsorption system consists of at least two adsorption control units. Each unit is equipped with at least one vacuum pump and one vacuum tank. The vacuum pump provides negative pressure to the vacuum tank, and the vacuum tank stores and stabilizes the negative pressure. Multiple suction cups are distributed to each adsorption control unit, so that each unit controls at least one suction cup. Each suction cup is connected to the vacuum tank of its respective adsorption control unit through an independent vacuum branch. A vacuum filter and a solenoid valve are installed sequentially along the vacuum flow direction on each vacuum branch to achieve filtration and on / off control of vacuum transmission.

[0048] The controller is electrically connected to the vacuum pump in each adsorption control unit and can independently send start and stop commands to the vacuum pump of each unit to achieve independent control of vacuum generation in each adsorption control unit. At the same time, the controller is electrically connected to the solenoid valve on each vacuum branch and can independently send on and off commands to each solenoid valve to achieve independent on and off of each vacuum branch, thereby achieving precise and independent control of adsorption and release of each suction cup.

[0049] During the photovoltaic module grabbing operation, the controller simultaneously activates all adsorption control units, all vacuum pumps operate to establish negative pressure in their corresponding vacuum tanks, all solenoid valves open, and each suction cup obtains negative pressure through an independent vacuum branch to jointly complete the adsorption of the photovoltaic module. When one or more adsorption control units malfunction (such as vacuum pump shutdown or vacuum tank leakage), or a vacuum branch malfunctions (such as solenoid valve damage or pipeline leakage), because each adsorption control unit and each vacuum branch is independent of each other, the fault will not spread to other units or branches. The remaining normal adsorption control units and vacuum branches can still generate negative pressure, achieving stable adsorption of the photovoltaic module through the corresponding suction cup, thus avoiding the photovoltaic module from falling off due to a single system failure.

[0050] After the photovoltaic modules are installed, the controller can stop the vacuum pumps of each adsorption control unit or directly open each solenoid valve according to the operation requirements, so that the suction cups are connected to the atmosphere, the negative pressure disappears, and the photovoltaic modules are released. If it is necessary to adjust the position of the photovoltaic modules locally, the controller can also independently control the release and adsorption of some suction cups to achieve fine adjustment of the photovoltaic modules and improve the installation accuracy.

[0051] The above solution forms a redundant design through multiple independent adsorption control units. When one or more units experience vacuum leakage or malfunction, the remaining normal units can still maintain effective negative pressure adsorption. Structurally, this eliminates the possibility of photovoltaic modules falling due to single-system failure. It solves the problems of traditional vacuum adsorption systems lacking redundancy and having low safety, and is especially suitable for handling large-size and heavy photovoltaic modules.

[0052] Each suction cup corresponds to an independent vacuum branch, and the branch is equipped with an independent solenoid valve, which is independently controlled by the controller to achieve precise adsorption and release of single or partial suction cups. It can not only complete the overall gripping and release, but also make local fine adjustments to the photovoltaic modules according to the installation requirements, thereby improving the accuracy of photovoltaic module installation.

[0053] Each adsorption control unit is equipped with a vacuum tank, which can stabilize the negative pressure generated by the vacuum pump, avoid negative pressure fluctuations caused by frequent start-stop of the vacuum pump, provide a continuous and stable vacuum source for the suction cup, ensure that the negative pressure value remains within the effective range during adsorption, and improve the stability of adsorption.

[0054] The vacuum filter on the vacuum branch can effectively filter dust and impurities, preventing impurities from entering the suction cup and affecting the sealing effect, or entering the vacuum pump and causing damage to the pump body, thus extending the service life of core components such as the suction cup and vacuum pump. At the same time, each adsorption control unit is independent of each other. When a certain unit fails, the unit can be repaired or replaced individually without stopping the whole system for maintenance, reducing equipment maintenance costs and downtime losses.

[0055] In practical applications, the number of adsorption control units and the distribution of suction cups can be flexibly adjusted according to the size and weight of the photovoltaic modules. The negative pressure supply of each suction cup can also be adjusted independently through the controller to adapt to the gripping needs of photovoltaic modules of different specifications, thereby improving the versatility of the vacuum adsorption system.

[0056] In some embodiments, the displacement detection system includes a sensor switch and a sensor plate disposed on each of the suction cups; The suction cup is slidably mounted on the end effector body via a buffer rod 21, and the sensing plate is disposed on the buffer rod; When the suction cup contacts and is pressed against the surface of the photovoltaic module, it moves the buffer rod and the sensing sheet to trigger the inductive switch, so that the inductive switch generates a signal indicating that the suction cup is in place.

[0057] In this embodiment, the buffer rod refers to a sliding rod that connects the suction cup and the end effector body. It has an elastic buffering function. The suction cup is installed at the lower end of the buffer rod, and the upper end of the buffer rod can slide relative to the end effector body. When the suction cup is in contact with the photovoltaic module and is subjected to pressure, the buffer rod can slide axially to achieve a buffering effect and avoid damage to the photovoltaic module or suction cup caused by hard contact.

[0058] An inductive switch is a sensing and triggering component of a displacement detection system. It is installed on the end effector and used in conjunction with the sensing element. When the sensing element moves to a preset position, the inductive switch is triggered and generates an electrical signal, indicating that the suction cup and the photovoltaic module are properly attached.

[0059] The sensor plate is the triggering component of the displacement detection system. It is installed on the buffer rod and moves synchronously with the sliding of the buffer rod. Its movement trajectory matches the trigger position of the induction switch. It is the core component for realizing the displacement detection of the suction cup.

[0060] In this embodiment, the displacement detection system adopts a combination of an inductive switch and an inductive sheet, along with a sliding buffer rod, to achieve accurate detection of the adhesion status between the suction cup and the photovoltaic module. This structure corresponds one-to-one with the suction cups, ensuring that the adhesion status of each suction cup can be detected independently. The detailed working principle is as follows: The buffer rod is slidably mounted on the suction cup mounting base 14 of the end effector body. The suction cup is fixedly mounted on the lower end of the buffer rod, and the sensing plate is fixedly mounted on the rod body of the buffer rod, forming an integrated structure with the buffer rod. The induction switch is independently mounted on the suction cup mounting base of the end effector body for each suction cup, and the installation position of the induction switch matches the sliding trajectory of the induction plate, ensuring that the induction plate can accurately trigger the induction switch when the buffer rod slides to the preset position.

[0061] When the controller controls the end effector to perform the bonding action, the end effector presses down towards the photovoltaic module. The suction cup first contacts the surface of the photovoltaic module. As the end effector continues to press down, the suction cup is subjected to the reaction force of the photovoltaic module. This reaction force is transmitted to the buffer rod, which pushes the buffer rod to slide upward relative to the suction cup fixing seat of the end effector, and at the same time drives the sensing plate installed on the buffer rod to slide upward synchronously.

[0062] When the buffer rod slides upward to the preset distance, that is, when the suction cup is completely attached to the surface of the photovoltaic module, the sensing plate moves to the trigger position of the induction switch and triggers the induction switch. After the induction switch is triggered, it immediately generates an electrical signal indicating that the suction cup is in place and transmits the signal to the controller in real time to complete the detection of the attachment status of a single suction cup.

[0063] The controller receives signals from the induction switches corresponding to all suction cups. When all induction switches are triggered, that is, when all suction cups are properly attached, the controller combines the stabilization time to determine the overall attachment status, providing a signal basis for the subsequent start-up of the vacuum adsorption system. If the induction switch of a certain suction cup is not triggered, it indicates that the suction cup is not properly attached, and the controller will pause the adsorption start-up until all induction switches are triggered.

[0064] The combination of the inductive switch and the sensing element results in a fast response speed, enabling real-time detection of the suction cup's displacement. A signal is generated immediately upon the suction cup's proper contact, with no significant delay, ensuring the controller can promptly acquire contact status information and improving the efficiency of the gripping operation. The suction cup is mounted to the end effector body via a buffer rod. The sliding design of the buffer rod provides elastic cushioning during the contact process, preventing hard contact when the end effector body presses down, which could cause scratches or cracks on the photovoltaic module surface or damage to the suction cup, thus reducing equipment and product wear and tear during operation.

[0065] The combination of the inductive switch and the inductive element is a mechanical contact detection structure. This design is simple, low-cost, and less susceptible to interference from complex environmental factors in photovoltaic power plants, such as strong sunlight, dust, and temperature variations. It offers high detection stability and reliability, making it suitable for long-term outdoor operation. The inductive switch, inductive element, and suction cup are installed one-to-one, allowing independent detection of the adhesion status of each suction cup. The controller can accurately determine the position of suction cups that are not properly adhered, facilitating subsequent troubleshooting and fine-tuning. This ensures that all suction cups are in contact with the surface of the photovoltaic module, guaranteeing the effectiveness of vacuum adsorption.

[0066] The signal indicating proper fit is automatically generated by the sensor switch and transmitted to the controller, eliminating the need for manual observation or judgment. This not only reduces the intensity of manual operation but also avoids subjective errors caused by human judgment, improving the convenience and accuracy of operation and making it suitable for automated and semi-automated paving operations.

[0067] In some embodiments, the displacement detection system further includes a micro-grating ruler integrated inside each of the buffer bars, the micro-grating ruler being used to measure the absolute compressive displacement d of the buffer bar relative to its mounting base; The controller is configured to perform a two-level fusion judgment in step S4, specifically including: When the sensor switch of any suction cup is triggered, it is determined to be a primary fit, and the reading of the laser rangefinder corresponding to that suction cup at the current moment is recorded as the reference distance value L0. After the initial bonding, the controller switches to reading the data of the micro grating ruler and continues to control the end effector body to perform micro-pressure. During the micro-pressure process, the following calculations are performed in real time for each suction cup: Obtain the reading L of the laser rangefinder at the current moment. t; Calculate the normal displacement δ of the photovoltaic module surface derived from laser measurements. laser The calculation formula is as follows: δ laser =L0-L t ; Obtain the compression displacement d of the buffer rod measured by the micro grating ruler at the current moment, and calculate the coupling difference Δ=δ according to the following formula.laser -d; The final fine bonding is considered complete when both of the following conditions are met: The compression displacement d of the buffer rod of all suction cups reaches the preset fine fit displacement threshold. The absolute values ​​of the coupling differences Δ calculated by all suction cups are all less than the preset deformation compatibility threshold ε; the deformation compatibility threshold ε is used to determine whether the photovoltaic module has undergone local elastic deformation beyond the permissible range or has poor contact.

[0068] In this embodiment, the miniature grating ruler refers to a high-precision displacement measuring component integrated inside the buffer rod. Based on the grating measurement principle, it can measure the absolute compression displacement of the buffer rod relative to its mounting base (suction cup fixing base) in real time and accurately. The measurement accuracy is much higher than that of conventional contact sensing components, providing high-precision data for accurate fit determination.

[0069] Primary bonding refers to the first judgment state of displacement detection, which is the bonding state when the sensor switch of any suction cup is triggered. It indicates that the suction cup and the photovoltaic module have achieved initial physical contact, which is the prerequisite state for subsequent fine bonding judgment.

[0070] The reference distance value L0 refers to the distance measurement value collected by the laser rangefinder of the suction cup at the initial bonding moment, which serves as the reference value for calculating the normal displacement of the photovoltaic module surface from the subsequent laser rangefinder value.

[0071] Micro-pressing refers to the small-distance, slow-speed pressing action performed by the controller-controlled pickup body after the initial bonding, so that the suction cup and the surface of the photovoltaic module can be more tightly bonded, while providing operating space for high-precision measurement and fine bonding judgment of the micro grating ruler.

[0072] δ normal displacement of photovoltaic module surface laser It refers to the displacement of the photovoltaic module surface in its normal direction, which is calculated from the change in the measurement value of the laser rangefinder, reflecting the change in the relative distance between the end effector body and the photovoltaic module surface.

[0073] The coupling difference Δ refers to the difference between the normal displacement calculated from the laser measurement value and the compression displacement of the buffer rod measured by the micro grating ruler. This difference can characterize the degree of local elastic deformation of the photovoltaic module and the degree of contact and adhesion of the suction cup.

[0074] The precision bonding displacement threshold refers to the minimum absolute compression displacement of the buffer rod required to achieve a tight fit between the suction cup and the photovoltaic module. Reaching this threshold indicates that the compression of the buffer rod meets the requirements for a tight fit.

[0075] The deformation compatibility threshold ε refers to the maximum threshold of the absolute value of the preset coupling difference. It is used to determine whether the photovoltaic module has undergone local elastic deformation beyond the permissible range, or whether there is poor contact or loose adhesion between the suction cup and the photovoltaic module. Its value is determined according to the material characteristics and structural strength of the photovoltaic module.

[0076] Final fine bonding refers to the second judgment state of displacement detection. It is a high-precision bonding state achieved by the fusion detection of micro grating ruler and laser range sensor on the basis of primary bonding. It indicates that the suction cup and the surface of photovoltaic module are tightly and effectively bonded, without local elastic deformation exceeding the standard or poor contact.

[0077] The displacement detection system in this embodiment integrates a miniature grating ruler on the basis of an inductive switch and an inductive plate, realizing a two-level fusion of laser ranging and grating ruler measurement for adhesion determination. The controller achieves high-precision detection of the suction cup adhesion state through this fusion determination. The detailed working principle is as follows: During the bonding action performed by the pickup body at the controller control end, the induction switch and the induction plate of the displacement detection system work together in real time. When the induction switch of any suction cup is triggered by the induction plate driven by the suction cup, the controller immediately determines that the suction cup has reached the initial bonding state, indicating that the suction cup has achieved initial physical contact with the photovoltaic module. At the same time, the controller captures and records the distance measurement value of the laser rangefinder corresponding to the suction cup at the moment of initial bonding in real time, and sets it as the reference distance value L0, providing a reference for subsequent normal displacement calculation.

[0078] Once a suction cup reaches the initial adhesion state, the controller immediately switches the suction cup's displacement detection mode from "inductive switch detection" to "micro grating ruler + laser rangefinder sensor fusion detection" mode. At the same time, the controller stops the large-distance downward pressure and switches to micro-pressure mode, continuing to press down towards the photovoltaic module at a small distance and slow speed. This ensures that the suction cup can achieve a tighter adhesion to the surface of the photovoltaic module and avoids excessive local elastic deformation of the photovoltaic module due to excessive pressure speed.

[0079] During the micro-pressing process, the controller simultaneously collects two types of data in real time for each suction cup that has achieved initial adhesion: one is the current distance measurement value L of the laser rangefinder corresponding to that suction cup. t Secondly, the absolute compressive displacement d of the buffer rod relative to its mounting base is measured by a micro-grating ruler integrated inside the buffer rod; and two calculations are performed in real time based on the collected data: (1) Based on the reference distance value L0 and the current laser measurement value L t Calculate the normal displacement δ of the photovoltaic module surface. laser The calculation formula is δ laser =L0-L tThis value reflects the change in the relative distance between the end effector body and the surface of the photovoltaic module during the micro-pinch pressing process; (2) Based on the calculated normal displacement δ laser The coupling difference Δ is calculated from the compressive displacement d measured by the micro grating ruler using the formula Δ=δ. laser -d, this value is the difference between the relative displacement measured by the laser and the absolute compressive displacement measured by the grating ruler. Its magnitude directly reflects the degree of local elastic deformation of the photovoltaic module and the degree of contact and adhesion of the suction cup.

[0080] During the micro-pressure process, the controller makes real-time judgments on the fusion detection data and calculation results of all suction cups. Only when both conditions are met simultaneously will it be determined that all suction cups have reached the final fine bonding state: (1) The absolute compression displacement d of the buffer rod of all suction cups reaches the preset fine bonding displacement threshold, indicating that all buffer rods are compressed to the preset degree and the suction cups and the surface of the photovoltaic module achieve close physical contact; (2) The absolute value of the coupling difference Δ calculated by all suction cups is less than the preset deformation compatibility threshold ε, indicating that the photovoltaic module has not undergone local elastic deformation beyond the permissible range during the bonding process, and all suction cups have achieved effective contact with the surface of the photovoltaic module without false bonding, poor contact, etc.

[0081] Once the controller determines that the final fine bonding state has been reached, it combines the determination of the stable duration of maintaining this state, and finally outputs a bonding completion signal, and starts the subsequent vacuum adsorption system to complete the gripping of the photovoltaic module.

[0082] The above solution adopts a two-stage fusion judgment mode of "primary bonding + final fine bonding". Combining the rapid triggering of the inductive switch with the high-precision measurement of the micro grating ruler and laser range sensor, it not only realizes the rapid initial judgment of the bonding state, but also completes the high-precision fine judgment, ensuring that the suction cup and the surface of the photovoltaic module achieve a tight and effective bonding, fundamentally avoiding situations such as false bonding and poor contact, and providing a higher precision guarantee for the stability of subsequent vacuum adsorption.

[0083] By calculating the coupling difference Δ and setting the deformation compatibility threshold ε, the local elastic deformation of the photovoltaic module during the bonding process can be monitored in real time. When the deformation exceeds the permissible range, the controller can stop pressing in time to avoid permanent deformation or breakage of the photovoltaic module due to excessive pressure, thus effectively protecting the structural integrity of the photovoltaic module.

[0084] The miniature grating ruler is integrated inside the buffer rod, which can accurately measure the absolute compression displacement of the buffer rod, making up for the shortcomings of the inductive switch, which can only achieve qualitative triggering and cannot quantitatively measure displacement. At the same time, combined with the non-contact measurement of the laser range sensor, it realizes the fusion of "contact absolute measurement + non-contact relative measurement", improving the accuracy and comprehensiveness of displacement detection.

[0085] The combination of micro-press mode and fusion detection enables the controller to perform fine control of the bonding process at small distances and slow speeds, avoiding problems such as hard contact and excessive pressure that may be caused by traditional pressing methods. At the same time, real-time data analysis and judgment can provide timely feedback on the bonding status, making it easier for the controller to adjust the pressing speed and distance according to the actual situation, thus improving the level of intelligence of bonding control.

[0086] In some embodiments, the end effector body is a rectangular frame structure, which includes at least two parallel longitudinal aluminum profiles and multiple transverse aluminum profiles connected between the two parallel longitudinal aluminum profiles, and the aluminum profiles are connected by corner brackets. At least four laser ranging sensors of the laser ranging and positioning system are respectively installed in the four corner areas of the rectangular frame, and are used to measure the vertical distance between the sensor and the four corner areas. The vacuum adsorption system has multiple suction cups mounted on the horizontal aluminum profile via their respective buffer rods. The horizontal aluminum profile is provided with a fixing seat for mounting and adjusting the buffer rods. The signal processing unit filters and denoises the received distance measurement signal, and sends the filtered and denoised distance data to the controller for spatial geometry calculation.

[0087] In this embodiment, the longitudinal aluminum profile 12 refers to the longitudinal load-bearing rod of the rectangular frame of the end effector body. At least two rods are provided and are parallel to each other. They provide the mounting base for components such as the laser rangefinder sensor and the fixed connection plate, and are the longitudinal skeleton of the rectangular frame.

[0088] The horizontal aluminum profile 11 refers to the horizontal rods connecting two vertical aluminum profiles. There are multiple horizontal rods, which together with the vertical aluminum profiles form a rectangular frame structure, providing an installation base for components such as buffer rods and suction cup fixing seats. The number and installation spacing can be flexibly adjusted according to the size of the photovoltaic modules.

[0089] Corner bracket 13 refers to the connector used to connect the longitudinal aluminum profiles and the transverse aluminum profiles. The corner bracket can achieve a firm and precise connection between the aluminum profiles, ensuring the stability and installation accuracy of the rectangular frame structure, while also facilitating the disassembly, assembly and adjustment of the frame.

[0090] The four corner areas refer to the four corners of the rectangular frame of the end effector. The laser rangefinder is installed in this area to measure the distance to the corresponding measurement points at the four corners of the photovoltaic module surface. This ensures that the distribution of measurement points can cover the main area of ​​the photovoltaic module surface, providing comprehensive and effective data for pose parameter calculation.

[0091] A suction cup mounting bracket is a suction cup mounting component installed on a horizontal aluminum profile. It is used to fix the buffer rod and the suction cup, and has a position adjustment function. The installation position can be adjusted according to the size and specifications of the photovoltaic module, so as to achieve flexible adjustment of the suction cup layout.

[0092] Filtering refers to the basic processing method of the distance measurement signal by the signal processing unit. It is used to eliminate random noise and interference signals in the signal, such as the fluctuations in the measurement signal caused by strong sunlight and dust in the outdoor photovoltaic power station, and to improve the stability of the distance measurement signal.

[0093] Noise reduction processing refers to the in-depth processing method of distance measurement signals by the signal processing unit. It is used to remove outliers and jump values ​​in the signal, such as erroneous measurement values ​​caused by momentary occlusion of laser rangefinders, to ensure the validity and accuracy of distance data.

[0094] This embodiment clearly defines the specific structure of the end effector body, the installation position of the laser rangefinder sensor, the installation method of the suction cup, and the processing method of the signal processing unit, making the structural layout of each system more reasonable and the data processing more standardized. The detailed working principle is as follows: The end effector body adopts a rectangular frame structure, which is constructed from at least two parallel longitudinal aluminum profiles and multiple transverse aluminum profiles. The connection points of the longitudinal and transverse aluminum profiles are firmly connected by corner brackets. The corner brackets ensure the connection accuracy and structural stability between the aluminum profiles, giving the rectangular frame sufficient structural strength to bear the weight of each system and external forces during operation. At the same time, the aluminum profiles are lightweight and corrosion-resistant, making them suitable for the outdoor working environment of photovoltaic power stations, and the frame structure is easy to disassemble, assemble, and adjust.

[0095] The laser ranging and positioning system has at least four laser ranging sensors, which are respectively installed at the four corners of the rectangular frame of the end effector body. This installation layout makes the four sensors symmetrically distributed in a rectangle. During operation, they can simultaneously illuminate the four corner measurement points on the surface of the photovoltaic module, and the distribution of measurement points covers the main area of ​​the photovoltaic module surface, ensuring that the collected distance data can fully reflect the planar features of the photovoltaic module surface. Each laser ranging sensor independently measures its own vertical distance to the corresponding measurement point, providing a comprehensive and effective distance measurement signal for subsequent pose parameter calculation.

[0096] Multiple suction cups of the vacuum adsorption system are mounted on a horizontal aluminum profile via their respective buffer rods. Suction cup mounting seats are fixedly installed on the horizontal aluminum profile, and the buffer rods are slidably mounted on the suction cup mounting seats. The suction cup mounting seats have a position adjustment function, which can flexibly adjust the installation position along the length of the horizontal aluminum profile. At the same time, the number and installation spacing of the horizontal aluminum profiles can be adjusted according to the size and specifications of the photovoltaic modules, thereby realizing flexible adjustment of the number and layout of suction cups. This allows the layout of the suction cups to match the surface of photovoltaic modules of different sizes and specifications, ensuring that the photovoltaic modules are subjected to uniform force during adsorption.

[0097] After the laser rangefinder sensor transmits the acquired distance measurement signal to the signal processing unit, the signal processing unit first filters the signal, using filtering algorithms to eliminate random noise and environmental interference signals, such as fluctuations in the measurement signal caused by strong light or dust, making the distance measurement signal more stable. Then, the filtered signal is denoised to remove outliers and jump values, such as erroneous measurements caused by momentary sensor occlusion, ensuring the validity of the distance data. Finally, the signal processing unit sends the filtered and denoised effective distance data to the controller, providing an accurate and reliable data foundation for the controller's spatial geometry calculation.

[0098] The rectangular frame of the end effector provides a precise and stable installation foundation for each system. The corner layout of the laser rangefinder ensures comprehensive pose detection. The adjustable installation of the suction cup ensures uniform adsorption. The filtering and noise reduction of the signal processing unit ensures data accuracy. The structure and processing methods of each part work together to enable the various systems of the intelligent end effector to work together more efficiently and stably. The accuracy and stability of the grasping operation are guaranteed from both structural layout and data processing aspects.

[0099] The above solution uses aluminum profiles to build a rectangular frame, which is connected by corner brackets. The structure has high strength and good stability and can withstand various external forces during operation. At the same time, aluminum profiles are lightweight and corrosion-resistant, making them suitable for the complex outdoor environment of photovoltaic power stations. The frame structure is easy to disassemble, assemble and adjust, and the number and installation spacing of aluminum profiles can be flexibly adjusted according to the size of the photovoltaic modules, which improves the adaptability of the end effector.

[0100] The laser rangefinder is installed at the four corners of the rectangular frame in a symmetrical rectangular distribution. The measurement points cover the main area of ​​the photovoltaic module surface. The collected distance data can comprehensively reflect the planar features of the photovoltaic module surface, avoiding the deviation in pose parameter calculation caused by uneven distribution of measurement points, and improving the comprehensiveness and accuracy of pose detection.

[0101] The suction cups are mounted on the suction cup mounting base of the horizontal aluminum profile via a buffer rod. The mounting base can be flexibly adjusted to change the installation position, and the number and spacing of the horizontal aluminum profiles can also be adjusted as needed. The number and layout of the suction cups can be adjusted according to photovoltaic modules of different sizes and specifications, so that the adsorption points of the suction cups can be evenly distributed on the surface of the photovoltaic module, ensuring that the photovoltaic module is subjected to uniform force during adsorption, and avoiding deformation and breakage of the photovoltaic module due to uneven force.

[0102] The signal processing unit performs dual filtering and noise reduction on the distance measurement signal, effectively eliminating signal noise and outliers caused by environmental interference and sensor failures, ensuring the accuracy and reliability of the distance data, providing a precise data foundation for the controller's spatial geometry calculation, and thus improving the accuracy of the pose parameter calculation.

[0103] The installation positions and methods of the end effector body, laser rangefinder, and suction cup are clearly defined, making the layout of each system more reasonable and compact, avoiding mutual interference between components, and improving space utilization. At the same time, the reasonable structural layout and standardized data processing methods enable the systems to work together more efficiently and stably, further improving the overall operating efficiency and stability of the intelligent end effector.

[0104] In some embodiments, the controller calculates the current pose parameters in the following manner: Based on the fixed installation coordinates of each laser ranging sensor in the coordinate system of the end effector and the distance data measured by them, the three-dimensional coordinates of at least four measuring points on the surface of the photovoltaic module in the coordinate system of the end effector are calculated. Based on the three-dimensional coordinates of the at least four measurement points, the plane equation of the photovoltaic module surface is fitted. Based on the plane equation, the center position deviation and tilt angle of the photovoltaic module surface are calculated.

[0105] In this embodiment, the end-effector body coordinate system refers to a three-dimensional rectangular coordinate system established with the end-effector body as the reference. It is the reference coordinate system that defines the installation coordinates of the laser rangefinder sensor and the measurement point coordinates of the photovoltaic module. Its origin and coordinate axis direction are fixed to the end-effector body to ensure that the measurement and calculation of all spatial position parameters have a unified reference.

[0106] The three-dimensional coordinates of the measurement point refer to the x, y, and z axis coordinates of the measurement point on the surface of the photovoltaic module in the coordinate system of the end-effector. They are calculated from the fixed installation coordinates of the laser rangefinder sensor and the measurement distance data, and are the core parameters characterizing the spatial position of the measurement point.

[0107] The plane equation is a mathematical equation used to characterize the spatial plane of the photovoltaic module surface. It is generally in the form of ax + by + cz + d = 0, where a, b, c, and d are plane parameters, which are obtained by mathematical fitting of the three-dimensional coordinates of at least four measurement points. It is the core basis for solving the pose parameters of the photovoltaic module.

[0108] This embodiment achieves accurate and standardized pose calculation through standardized spatial coordinate transformation, plane equation fitting, and pose parameter derivation. The detailed working principle is as follows: The controller pre-stores the fixed installation coordinates (x, y) of all laser rangefinder sensors in the end-effector body coordinate system. i ,y i ,z i ) body Furthermore, the laser ranging sensors all measure the distance to the photovoltaic module surface along the normal direction of their location; the controller receives the distance measurement values ​​d from each sensor transmitted by the signal processing unit. i Then, based on the coordinate conversion rules of the spatial coordinate system, the sensor installation coordinates are combined with the measurement distance to calculate the three-dimensional coordinates (x, y, y) of the measurement point on the photovoltaic module surface corresponding to each sensor in the end-effector's body coordinate system. i ,y i ,z i ) body This completes the conversion from distance data to spatial location data, providing basic data for plane equation fitting.

[0109] The controller collects three-dimensional coordinate data from at least four measurement points. Due to minor errors during the measurement process, the measurement points are not entirely on the same ideal plane. Therefore, the controller uses the least squares method to fit the three-dimensional coordinates of all measurement points to a plane. By minimizing the sum of the squares of the distances from each measurement point to the fitted plane, the plane equation that best fits the actual surface of the photovoltaic module, ax + by + cz + d = 0, is obtained. This fitting method can effectively eliminate the influence of measurement errors from a single laser ranging sensor, environmental interference, and other factors on the plane determination, ensuring that the fitted plane equation is highly matched with the actual surface of the photovoltaic module.

[0110] Based on the fitted plane equation and the preset size parameters of the photovoltaic module, the controller calculates the actual coordinates of the geometric center of the photovoltaic module surface in the coordinate system of the end-effector. At the same time, the controller pre-stores the reference coordinates of the geometric center of the end-effector in its own coordinate system. By comparing the two coordinates, the controller calculates the offset of the geometric center of the photovoltaic module from the geometric center of the end-effector in the x, y, and z axes. This offset is the center position deviation, which comprehensively characterizes the degree of offset between the two in spatial position.

[0111] The normal vector of the plane equation ax+by+cz+d=0 is (a,b,c). The controller pre-stores the reference value of the normal vector of the end-effector's suction cup plane in its own coordinate system. By calculating the spatial angle between the normal vector of the fitted plane on the photovoltaic module surface and the reference normal vector of the end-effector's suction cup plane, the tilt angles of the photovoltaic module surface relative to the end-effector body in the horizontal and vertical directions are calculated respectively, which fully characterizes the deviation of the two in spatial attitude. If the normal vector of the fitted plane is in the same direction as the reference normal vector, the tilt angle is 0, indicating that the two planes are completely parallel.

[0112] The controller integrates the calculated center position deviation (x, y, z axis offset) with the tilt angle (horizontal and vertical angle) to form complete current pose parameters. It then compares these parameters with the preset pose tolerance range to provide a precise and quantitative basis for generating subsequent pose adjustment commands.

[0113] The above scheme establishes a unified coordinate system for the end effector body, converts distance measurement data into three-dimensional spatial coordinates, and then obtains quantified center position deviation and tilt angle through plane fitting and geometric calculation, replacing the traditional qualitative judgment method. This makes the calculation of pose parameters more accurate and quantifiable, providing clear numerical basis for pose adjustment.

[0114] Using the least squares method to fit the coordinates of multiple measurement points to a plane can effectively eliminate local data deviations caused by factors such as measurement errors of a single laser rangefinder and outdoor environmental interference, ensuring that the fitted plane equation is highly matched with the actual surface of the photovoltaic module, thereby improving the accuracy and reliability of pose parameter calculation.

[0115] The calculated center position deviation includes the offset in the x, y, and z axes, and the tilt angle includes the angle between the horizontal and vertical directions. It can characterize the relative pose deviation between the photovoltaic module and the end effector body from all angles and perspectives, ensuring that subsequent pose adjustment can achieve accurate spatial positioning without adjustment blind spots.

[0116] Both plane equation fitting and pose parameter calculation are based on the three-dimensional coordinates of the measurement point and the preset size parameters of the photovoltaic module. The calculation basis can be flexibly adjusted according to the size parameters of photovoltaic modules of different specifications without changing the core calculation logic, thus improving the universality and adaptability of the pose calculation method.

[0117] In some embodiments, the intelligent end effector further includes: An environmental sensing module, integrated on the end effector body, is used to acquire environmental wind field information and motion state information of the end effector body in real time. A drive actuator, connected to the controller, is used to drive the end effector body to move according to the instructions of the controller; The controller is further configured to synchronously perform feedforward wind load dynamic compensation control when executing step S3 and / or step S4, specifically including: The environmental sensing module acquires in real time the ambient wind speed V acting on the end effector body and the adsorbed photovoltaic module. w Wind direction angle θ w , and the triaxial angular velocity and acceleration of the end effector body; Based on the tilt angle in the current pose parameters and the ambient wind speed V w With wind direction angle θ w And a pre-existing aerodynamic parameter model of the photovoltaic module in the controller, to calculate in real time the predicted normal force F generated by wind load in the normal direction of the photovoltaic module plane. wind and the predicted normal force F wind The predicted moment vector M in the component plane wind The aerodynamic parameter model includes at least the windward area of ​​the photovoltaic module, the location of the wind pressure center, and the aerodynamic coefficient. The controller is based on the predicted torque vector M wind， Based on the dynamic model of the drive actuator, the feedforward compensation algorithm built into the controller is used to calculate the compensation method for M. wind Required compensation torque command ΔM comp ; In step S3 and / or step S4, the final control command received by the drive actuator is: the basic control command and the compensation torque command ΔM. comp The superposition of the basic control commands is generated by the pose adjustment algorithm in step S3 or the control logic that generates the fitting control commands in step S4.

[0118] In this embodiment, the environmental perception module refers to the multi-sensor data acquisition module integrated into the end effector body. It can collect environmental wind field information and end effector body motion status information in real time, providing basic data for dynamic compensation of wind load. It generally includes wind speed and direction sensors, three-axis gyroscopes, three-axis accelerometers, etc.

[0119] The drive actuator refers to the power execution component connected to the controller. It is the power source for the movement of the end effector body and can realize the translation, rotation, tilting and other movements of the end effector body according to the control command of the controller. It generally includes servo motors, reducers, robotic arm joints, etc.

[0120] Feedforward wind load dynamic compensation control is a control method that anticipates disturbances and performs active compensation in advance. By calculating the disturbance force and disturbance torque generated by wind load in real time, compensation commands are generated in advance and superimposed on the basic control commands to offset the influence of wind load on the end effector movement and improve control accuracy.

[0121] Wind field information refers to parameters characterizing the outdoor wind environment of a photovoltaic power station, including at least the ambient wind speed V. w (Airflow speed) and wind angle θ w The angle between the wind direction and a certain coordinate axis of the end-feeder's coordinate system is the core environmental parameter for calculating wind loads.

[0122] The aerodynamic parameter model refers to the mathematical model pre-stored in the controller for calculating the wind load on the photovoltaic module. It includes core parameters such as the windward area of ​​the photovoltaic module, the position of the wind pressure center, and the aerodynamic coefficient. These parameters are pre-calibrated and stored according to the size, structure, and material characteristics of the photovoltaic module.

[0123] Predicting normal force F wind This refers to the force exerted by wind load in the normal direction of the photovoltaic module plane, which causes the end-effector body to shift in the normal direction, affecting the stability of bonding and adsorption.

[0124] Predicted torque vector M wind It refers to the torque generated by wind load in the plane of photovoltaic module. It is a three-dimensional vector that causes the end effector to rotate and tilt, thus disrupting its relative posture with the photovoltaic module.

[0125] Basic control commands refer to the original control commands generated by the pose adjustment algorithm or the bonding control logic, which are used to drive the end effector to complete the pose adjustment or bonding operation, without considering external interference factors such as wind load.

[0126] This embodiment addresses the interference problem caused by outdoor wind farms in photovoltaic power plants by optimizing the motion control of the end effector to resist interference. The detailed working principle is as follows: First, environmental and motion status information is collected. Specifically, the environmental sensing module integrated into the end-sensor continuously works to collect two types of core data in real time: one is the outdoor environmental wind field information of the photovoltaic power station, including the environmental wind speed V acting on the end-sensor and the photovoltaic module. w Wind direction angle θ w Secondly, the motion state information of the end effector body, including the three-axis angular velocity and three-axis acceleration of the end effector body; all collected data are transmitted to the controller in real time to provide dynamic basic data for wind load calculation.

[0127] Then, wind load-related parameters are retrieved. Specifically, after receiving environmental and motion status information, the controller retrieves the corresponding aerodynamic parameter model from its internal storage based on the specifications of the photovoltaic module currently in operation. This aerodynamic parameter model includes pre-calibrated parameters such as the windward area, wind pressure center position, and aerodynamic coefficient of the photovoltaic module of that specification. At the same time, the controller retrieves the tilt angle from the calculated current pose parameters as the attitude basis for wind load calculation.

[0128] Then, wind load prediction calculations are performed, specifically including: the controller based on the tilt angle of the photovoltaic modules and the real-time collected ambient wind speed V. w With wind direction angle θ w By combining the retrieved aerodynamic parameter model and using relevant calculation formulas from fluid mechanics and rigid body mechanics, the two core effects of wind load on photovoltaic modules are calculated in real time: one is the predicted normal force F generated in the direction normal to the photovoltaic module plane. wind Secondly, the predicted moment vector M generated in the plane of the photovoltaic module. wind The calculation process is performed in real time and is dynamically updated as the wind field information and end-feeder pose change, ensuring accurate prediction of wind load.

[0129] Then, the compensation torque command is calculated, specifically including: the controller pre-stores the dynamic model of the drive actuator, which reflects the correspondence between the input command and output torque and motion state of the drive actuator; the controller calculates the predicted torque vector M. wind Substituting the dynamic model of the drive actuator, the built-in feedforward compensation algorithm is used to calculate the compensation torque command ΔM required to counteract the wind load torque interference. comp This instruction serves as the basis for torque compensation in driving the actuator.

[0130] Then, the composite control command is generated and executed, specifically including: when the controller executes step S3 (pose adjustment) and / or step S4 (fitting action), the calculated compensation torque command ΔM is generated. comp The control commands are superimposed on the basic control commands to generate composite control commands. The drive actuator receives and executes these composite control commands, simultaneously offsetting the wind load torque vector M while completing the pose adjustment or alignment operation. wind To prevent interference, ensure that the movement of the end effector body is not affected by the outdoor wind field and maintain a stable relative position with the photovoltaic module.

[0131] Then, real-time updates of dynamic compensation are performed, specifically including: the environmental perception module continuously collects wind field and motion state information, and the controller dynamically updates the wind load calculation results and compensation torque commands based on real-time data, so that the feedforward compensation control can be adjusted in real time with changes in the wind field, realize dynamic compensation of wind load, and ensure that the high precision of end effector control can still be maintained under wind field fluctuations.

[0132] The above scheme adopts feedforward wind load dynamic compensation control, which predicts the torque interference generated by wind load in advance and generates compensation commands to offset the influence of wind field on end-effector movement from the control source. This solves the problems of end-effector position offset and unstable bonding caused by outdoor wind field fluctuations in photovoltaic power plants, and greatly improves the control accuracy of position adjustment and bonding operations.

[0133] Unlike traditional feedback compensation, feedforward compensation does not require waiting for the disturbance to occur before making adjustments. Instead, it calculates the disturbance in advance and compensates for it by collecting wind field information in real time, which has the characteristics of no lag and fast response. At the same time, the compensation command is dynamically updated with the wind field information, realizing real-time tracking and compensation of wind load, and adapting to the dynamic changes of outdoor wind field.

[0134] By counteracting the normal force and torque vector generated by wind load, the displacement and tilt of the end effector body caused by wind field are effectively avoided, ensuring that the suction cup can accurately adhere to the photovoltaic module during the bonding operation, maintaining the stability of negative pressure during the adsorption operation, and reducing the risk of bonding failure and module detachment caused by wind field interference.

[0135] The addition of dynamic wind load compensation control enables the intelligent end effector to work stably in complex outdoor environments with wind. Even under fluctuating wind speeds, it can still maintain high-precision motion control, solving the problem of poor wind resistance of traditional end effectors and their ability to operate only in windless / light wind environments, thus greatly expanding the range of environmental adaptability.

[0136] The compensation torque command achieves the compensation function by superimposing it on the basic control command. It does not change the original position adjustment and fit control logic of the end effector. It can be seamlessly integrated with the original control algorithm. At the same time, the compensation intensity can be flexibly adjusted according to the actual wind field conditions, which has strong compatibility and flexibility.

[0137] In some embodiments, the feedforward compensation algorithm calculates the compensation torque command ΔM through the following steps. comp : The predicted torque vector M wind The transformed disturbance torque vector M is obtained by transforming the coordinate system based on the photovoltaic module plane to the joint space coordinate system or the drive motor coordinate system of the drive actuator. wind' ; Based on the dynamic model of the drive actuator, the torque vector M to counteract the disturbance is calculated. wind' The required joint drive torque or motor current command is used as an initial compensation amount. The initial compensation amount is filtered to match the frequency band characteristics of the filtered initial compensation amount with the actual response bandwidth of the drive actuator, and then multiplied by an adjustable feedforward gain coefficient K. ff Generate the compensation torque command ΔM comp K ff The value range is (0,1).

[0138] In this embodiment, the photovoltaic module planar coordinate system refers to a local coordinate system established with the surface of the photovoltaic module as a reference, which defines the predicted torque vector M. wind The original coordinate system, whose origin is generally the geometric center of the photovoltaic module.

[0139] The joint space coordinate system / drive motor coordinate system refers to the coordinate system established with the joints or drive motors of the drive actuator as the reference. It is the reference coordinate system for the drive actuator to receive control commands, ensuring that the compensation torque command can be directly recognized and executed by the drive actuator.

[0140] Disturbance torque vector M wind′ This refers to the predicted torque vector M wind After coordinate system transformation, the torque vector in the joint space coordinate system / drive motor coordinate system of the drive actuator is the interference torque that the drive actuator actually needs to counteract.

[0141] Joint drive torque / motor current command refers to the core control input for driving the actuator. Joint drive torque is used to control the joints of the robotic arm, and motor current command is used to control the servo motor. It is the direct basis for realizing torque output.

[0142] The initial compensation amount refers to the amount used to offset the disturbance torque vector M. wind′ The original joint driving torque or motor current command calculated from the dynamic model of the drive actuator is not filtered or adjusted for gain.

[0143] Frequency band characteristics refer to the frequency distribution characteristics of a signal. The actual response bandwidth of the actuator is the range of signal frequencies that it can effectively respond to. Signals outside this range cannot be recognized by the actuator and will cause vibration of the actuator.

[0144] Feedforward gain coefficient K ff It refers to the coefficient used to adjust the intensity of the compensation torque command. The value range is (0,1]. It can be flexibly adjusted according to the actual working environment and control accuracy requirements to avoid over-compensation or under-compensation.

[0145] This embodiment achieves precision and practicality of the compensation torque command through coordinate system transformation, preliminary compensation calculation, filtering, and gain adjustment, ensuring that the drive actuator can effectively perform the compensation action. The detailed working principle is as follows: Predicted torque vector M wind The control commands for the drive actuators are defined in the planar coordinate system of the photovoltaic module, while they need to be input in the joint space coordinate system / drive motor coordinate system. Therefore, the controller first uses rotation and translation transformation algorithms of the space coordinate system to transform the predicted torque vector M. wind The transformed disturbance torque vector M is obtained by transforming the photovoltaic module's planar coordinate system to the joint space coordinate system / drive motor coordinate system of the drive actuator. wind′ This transformation process ensures that the reference coordinate system of the torque vector is consistent with the control coordinate system of the driving actuator, so that the compensation command can be accurately identified by the actuator.

[0146] The initial compensation calculation refers to the controller retrieving the pre-stored dynamic model of the drive actuator. This model includes the core characteristics of the drive actuator, such as inertia parameters, damping parameters, and transmission ratio, reflecting the quantitative relationship between the input joint drive torque / motor current and the output torque. The controller then transforms the disturbance torque vector M... wind′ Substituting into the dynamic model, the joint driving torque or motor current command required to counteract the disturbance torque is obtained through inverse dynamics calculation, and used as the initial compensation amount; this calculation process realizes the quantitative conversion from disturbance torque to actuator control input.

[0147] The filtering of the preliminary compensation amount refers to the fact that the drive actuator has an actual response bandwidth. High-frequency signals exceeding this bandwidth cannot be effectively responded to by the actuator and will also cause high-frequency vibrations in the mechanism, affecting motion stability. At the same time, the preliminary compensation amount may contain high-frequency noise caused by wind field fluctuations and measurement errors. Therefore, the controller adopts a low-pass filtering algorithm that matches the response bandwidth of the actuator to filter the preliminary compensation amount, remove the high-frequency components, and make the frequency band characteristics of the filtered preliminary compensation amount completely match the actual response bandwidth of the drive actuator, ensuring that the compensation amount can be stably and effectively responded to by the actuator.

[0148] Adjusting the feedforward gain coefficient refers to setting the feedforward gain coefficient K for the initial compensation amount after filtering, in order to avoid over-compensation or under-compensation caused by errors in aerodynamic parameter modeling or wind field measurement. ff The value of this coefficient ranges from (0,1] and can be flexibly adjusted according to the actual wind farm conditions and operational control accuracy requirements of the photovoltaic power station; the controller multiplies the filtered preliminary compensation amount by this feedforward gain coefficient K. ff The final compensation torque command ΔM is obtained. comp When the wind field fluctuates significantly, K can be appropriately increased. ff To increase the compensation intensity; when the wind field is relatively stable, K can be appropriately reduced. ff To avoid overcompensation.

[0149] Output and superposition of compensation torque command: The controller will calculate the final compensation torque command ΔM. comp The data is output in real time and superimposed with the basic control commands in steps S3 and S4 to generate a composite control command, which is then sent to the drive actuator to achieve accurate and stable compensation for wind load torque interference.

[0150] The above scheme uses coordinate system transformation to make the reference of the torque vector consistent with the control reference of the drive actuator, thus avoiding the failure of compensation commands caused by coordinate system inconsistency. At the same time, through filtering, the frequency band characteristics of the compensation amount are matched with the response bandwidth of the actuator, ensuring that the compensation command can be stably and effectively responded to by the actuator without problems such as high-frequency vibration.

[0151] By setting the feedforward gain coefficient K ff It can flexibly adjust the intensity of the compensation command, effectively avoiding over-compensation or under-compensation caused by aerodynamic parameter model calibration errors and wind field information measurement errors, thus improving the robustness of the feedforward compensation algorithm. Even with certain errors, it can still achieve good compensation results.

[0152] The filtering process removes high-frequency noise and signals exceeding the response bandwidth of the actuator from the initial compensation amount, avoiding vibration and impact of the actuator caused by high-frequency signals, ensuring the motion stability of the driven actuator when performing compensation actions, and extending the service life of the actuator.

[0153] The core steps of the feedforward compensation algorithm can flexibly adjust the dynamic model and control command form according to the type of driving actuator (such as servo motor drive, hydraulic drive). By simply changing the coordinate system transformation parameters and the dynamic model, it can be adapted to different types of actuators, thus improving the versatility of the algorithm.

[0154] In some embodiments, the controller further integrates a distributed fault diagnosis and fault tolerance control module, which is configured to perform the following steps: For the i-th laser rangefinder, its instantaneous measurement value is defined as d. i (t), within the time window T w Calculate its rate of change The difference Δd between adjacent periods i ; The data validity flag of the laser rangefinder sensor shall be set to invalid if any of the following conditions are met: d i (t) remains equal to zero, has a maximum value, or a specific error code exceeds N. err One sampling period; | |>V max , where V max The limit distance change rate is calculated based on the maximum speed of the end effector body; ∣Δd i |>ΔD max , where ΔD max This is the maximum reasonable jump threshold set based on the mechanical vibration characteristics; If the valid data flag of a certain laser rangefinder remains invalid for a duration exceeding the first preset time threshold T, fault If the laser rangefinder is determined to be in a hard failure state, it will be removed from the set of currently valid laser rangefinders S. valid Remove from; Let the total number of laser rangefinders be N, and define the effective set of laser rangefinders as S. valid The number of elements is M; If M≥4, then directly use S valid Installation coordinates (x) of all laser rangefinders i ,y i ,z i ) body and its measured value d i By fitting the plane using the least squares method or using the SVD decomposition method, the equation ax+by+cz+d=0 of the photovoltaic module plane in the coordinate system of the end-sensor is solved, and the tilt angle and the deviation of the center position are obtained. If M=3, a spatial plane is determined using data from three effective laser rangefinders. Combined with the encoder feedback value of the drive actuator, the expected pose of the end effector body is obtained through forward kinematics calculation. The expected pose is then fused with the laser plane equation, and the complete current pose parameters are estimated using a Kalman filter. If M < 3, it is determined that the pose cannot be reliably calculated at present, the operation is immediately suspended and a fault alarm message indicating that manual intervention is required is issued.

[0155] In this embodiment, the distributed fault diagnosis and fault-tolerant control module refers to the intelligent diagnosis and fault-tolerant module integrated into the controller (which can be implemented by a computer program). It uses distributed logic to perform independent fault diagnosis on each laser ranging sensor and adopts different pose calculation strategies according to the number of effective sensors to achieve fault-tolerant control under sensor failure and ensure the continuity of operation.

[0156] Instantaneous measurement value d i (t) refers to the real-time distance measurement value of the i-th laser rangefinder at time t.

[0157] Time window T w This refers to the preset statistical time range for fault diagnosis data, used to calculate the rate of change of sensor measurements and the difference between adjacent periods, ensuring the timeliness and accuracy of the diagnostic results.

[0158] rate of change This refers to the measurement value of the i-th laser rangefinder sensor within the time window T. w The rate of change within reflects the dynamic trend of the measured value.

[0159] The difference between adjacent periods Δd i It refers to the difference in measurement values ​​between two adjacent sampling periods of the i-th laser rangefinder sensor, reflecting the degree of jump in the measurement value.

[0160] The data validity flag is a logic bit used to identify the validity of the measurement data from the laser rangefinder. It has two states: "valid" and "invalid". The controller updates this flag in real time based on the fault diagnosis results.

[0161] A hard failure state refers to a permanent fault state of a laser rangefinder sensor, where the valid data flag bit of the sensor remains invalid for an extended period of time, indicating that the laser rangefinder sensor can no longer function properly and needs to be eliminated or replaced.

[0162] Effective laser rangefinder sensor set S valid It refers to the set of all laser rangefinders whose data validity flag is "valid", and the number of its elements M is the number of currently valid sensors.

[0163] Singular Value Decomposition (SVD) is a mathematical method for matrix decomposition. It can be used to fit a plane equation based on the coordinates of a small number of measurement points, and can still ensure the accuracy of plane fitting when the number of effective sensors is small.

[0164] The Kalman filter is an optimal estimation algorithm that achieves accurate estimation of the system state by fusing measurement data from multiple sensors with system state prediction. When the number of effective sensors is 3, it fuses the laser plane equation with the encoder data of the drive actuator to estimate the complete pose parameters of the photovoltaic module.

[0165] This embodiment addresses potential malfunctions in laser rangefinder sensors by implementing real-time fault diagnosis and fault-tolerant control for pose calculation. This ensures that even with partial sensor failure, the intelligent end effector can maintain operation as much as possible. The detailed working principle is as follows: First, the distributed fault diagnosis and fault-tolerant control module independently monitors each laser ranging sensor in a distributed manner, and collects the instantaneous measurement value d of the i-th sensor in real time. i (t); simultaneously within the preset time window T w Internally, it calculates two core characteristic parameters of each sensor in real time: one is the rate of change of the measured value. Secondly, the difference Δd between adjacent sampling periods. i .

[0166] Then, the distributed fault diagnosis and fault-tolerant control module sets a data valid flag bit for each sensor and presets three invalid judgment conditions. If a sensor meets any of the conditions, its data valid flag bit is immediately set to invalid: (1) Measured value d i (t) remains equal to zero, the device maximum value, or a specific error code, and exceeds N. err One sampling period indicates that the sensor has a hardware fault or signal transmission interruption; (2) rate of change | |>V max Vmax The rate of change of the limit distance calculated based on the maximum speed of the end effector body indicates that the rate of change of the measured value exceeds the physical limit and is abnormal data; (3) the difference between adjacent cycles |Δd i |>ΔD max , where ΔD max The maximum reasonable jump threshold set according to the mechanical vibration characteristics indicates that the measured value has an irregular jump, which is abnormal data.

[0167] Then, the distributed fault diagnosis and fault-tolerant control module continuously monitors the sensors whose data validity flag is "invalid". If the data validity flag of a certain laser ranging sensor remains invalid for a duration exceeding the first preset time threshold T, the system will take action. fault If the sensor fails to perform properly, it is determined to be in a hard failure state, indicating that the sensor can no longer function normally; the module immediately removes the laser rangefinder from the set of valid laser rangefinders S. valid The sensor data will be removed from the calculation and will no longer be used for subsequent pose calculations.

[0168] Then, fault-tolerant pose calculation based on the number of effective sensors is performed, specifically by real-time statistical analysis of the set S of effective laser ranging sensors. valid The number of elements M is determined, and different pose calculation strategies are adopted according to different values ​​of M to achieve fault tolerance control, as follows: When M≥4, directly use S valid Installation coordinates (x) of all valid sensors i ,y i ,z i ) body and its measured value d i The plane equation ax+by+cz+d=0 of the photovoltaic module surface is fitted by the least squares method or SVD decomposition method. Then, the tilt angle and center position deviation are calculated according to the plane equation. The calculation logic is consistent with the normal state, ensuring the accuracy of pose calculation. When M=3, a spatial plane is first determined using the measurement data from three effective sensors to obtain the preliminary laser plane equation; at the same time, the encoder feedback value of the drive actuator is retrieved, and the expected pose of the end effector body is obtained through forward kinematics calculation; then the laser plane equation and the expected pose data of the end effector are input into the Kalman filter, and the optimal estimation of the complete current pose parameters of the photovoltaic module is achieved through data fusion, making up for the deficiency of insufficient number of effective sensors; When M < 3, since three points are needed to determine a spatial plane, the plane equation cannot be reliably fitted using laser data. The module immediately determines that the pose cannot be reliably calculated and triggers the fault-tolerant protection mechanism: immediately suspends all operations of the end effector and sends a fault alarm message to the operator requiring manual intervention, reminding the operator to check for sensor faults.

[0169] The above scheme adopts distributed independent monitoring and quantitative feature judgment, which can diagnose the faults of each laser ranging sensor in real time. Through the dual constraints of multiple judgment conditions and time thresholds, it effectively avoids misdiagnosis caused by environmental interference and instantaneous signal fluctuations, and ensures the accuracy of fault diagnosis results.

[0170] Different pose calculation strategies are adopted based on the number of effective sensors. Pose calculation can still be achieved when M≥3, avoiding the downtime of the entire machine due to the failure of a single or a few sensors. This greatly improves the fault tolerance of the intelligent end effector, ensures the continuity of photovoltaic module installation, and reduces the loss of work interruption caused by equipment failure.

[0171] When M≥4, the conventional least squares / SVD decomposition method is used to ensure the accuracy of the solution. When M=3, the laser data and encoder data are fused by Kalman filter to achieve the optimal estimation of the pose parameters, which minimizes the impact of insufficient number of effective sensors on the accuracy of the solution and ensures that the pose accuracy requirements of the operation can still be met under fault-tolerant conditions.

[0172] The laser rangefinder sensor malfunction is handled in a tiered manner, with "preliminary judgment of invalid data" and "final judgment of hard failure," avoiding over-processing of transient malfunctions. At the same time, when M<3, the operation is immediately suspended and a manual intervention alarm is issued, realizing safety protection under fault conditions and avoiding safety accidents such as damage to photovoltaic modules and collisions of end effectors caused by pose calculation failures.

[0173] Distributed fault diagnosis can accurately locate the faulty sensor, and operators can directly check and replace the faulty sensor based on the alarm information without having to test all sensors one by one, which greatly reduces the time and cost of manual maintenance; at the same time, fault-tolerant control allows the equipment to continue to operate for a period of time in a faulty state, buying time for manual maintenance.

[0174] In some embodiments, during the bonding process in step S4, for the j-th suction cup, the displacement sensor measurement value of that suction cup is defined as δ. j (t); Calculate the average value of all suction cup displacements. and standard deviation δ(t); If |δ j (t)- |>k·δ(t) and continuously exceeds the second preset time threshold T abn If the displacement of suction cup j is abnormal, then k is determined to be a preset statistical outlier coefficient. When suction cup j is determined to be abnormal, the following control parameter reconstruction steps are performed: The target pressure value P of the vacuum generator corresponding to suction cup j. target,j Adjust to P target,j The formula is adjusted as follows, P target,j '=α·P target,j Where 0 < α < 1 is the pressure attenuation coefficient; Generate the speed command that drives the end effector body to perform the contact action. At that time, a velocity attenuation field is applied to the local region corresponding to suction cup j, so that the normal velocity v of this local region in the body coordinate system of the end effector approaches the velocity v. n,j Satisfy v n,j =β·v n,default Where β is the velocity decay coefficient, 0 < β < 1; The suction cup array is abstracted as an undirected graph G=(V,E), where vertex V represents the suction cup and edge E represents the adjacency relationship; Once suction cup j is determined to be abnormal, iterate through all its adjacent suction cup sets N(j); For each adjacent suction cup p∈N(j), the self-adhesion displacement threshold δ is reached at that adjacent suction cup. th Then, an additional compensation displacement δ is applied. comp,p The calculation formula is as follows: ; Where γ is the global compensation gain. Let (j,p) be the weight of edge (j,p). This is the theoretical displacement value of the suction cup j estimated based on the current pose parameters; In step S4, the controller adjusts the condition for determining the completion of adhesion of the normal suction cup p to: δ p (t)≥δ th +δ comp,p And it maintains a stable duration.

[0175] In this embodiment, the displacement sensor measures the value δ j (t) refers to the displacement measurement value of the displacement sensor (inductive switch + miniature grating ruler) corresponding to the j-th suction cup at time t, which is the basic data for diagnosing abnormal suction cup displacement.

[0176] average value It refers to the arithmetic mean of the measurements of all suction cup displacement sensors at time t, reflecting the overall average adhesion of the suction cup.

[0177] The standard deviation δ(t) refers to the standard deviation of all suction cup displacement sensor measurements at time t, reflecting the dispersion of each suction cup displacement measurement, i.e. the consistency of the fit.

[0178] The statistical outlier coefficient k is a preset outlier determination coefficient used to determine whether a single suction cup displacement value is an abnormal outlier value. It can generally be set to 2-3.

[0179] Abnormal suction cup displacement refers to a deviation between the displacement measurement value of a single suction cup and the overall average value that exceeds the statistically reasonable range and persists for a certain period of time, indicating that the suction cup has problems such as poor adhesion, jamming, or buffer rod failure.

[0180] Vacuum generator target pressure value P target,j This refers to the preset target negative pressure value of the vacuum generator corresponding to the j-th suction cup, which is the pressure basis for achieving adsorption.

[0181] The pressure attenuation coefficient α is a coefficient used to adjust the target pressure value of the abnormal suction cup. 0 < α < 1, which reduces the adsorption pressure of the abnormal suction cup and avoids negative pressure leakage caused by poor adhesion.

[0182] The velocity attenuation field refers to the velocity attenuation region set for the local area of ​​the abnormal suction cup in the coordinate system of the end effector body, so as to reduce the normal approach velocity of the area and avoid damage caused by hard contact.

[0183] Normal approach velocity v n It refers to the velocity in the normal direction of the end effector body moving towards the photovoltaic module, and is the core motion parameter of the bonding operation.

[0184] The speed attenuation coefficient β is a coefficient used to adjust the normal approach speed of a local area of ​​the abnormal suction cup. 0 < β < 1, which reduces the downward pressure speed in that area and facilitates the adhesion adjustment of the abnormal suction cup.

[0185] An undirected graph G=(V,E) is used to abstractly represent the topology of a suction cup array, where vertex V represents each suction cup and edge E represents the adjacency relationship between two suction cups, with no direction distinction.

[0186] The set of adjacent suction cups N(j) refers to the set of all adjacent suction cups of the j-th abnormal suction cup, reflecting the spatial adjacency relationship between suction cups.

[0187] Compensation displacement δ comp,p This refers to the additional compression displacement set for adjacent suction cups of abnormal suction cups, used to compensate for insufficient adhesion of abnormal suction cups and ensure the uniformity of overall adhesion of photovoltaic modules.

[0188] The global compensation gain γ is a global coefficient used to adjust the magnitude of the compensation displacement, which is set according to the material characteristics and structural strength of the photovoltaic module.

[0189] edge weight w jp It refers to the weight of edge (j,p) in an undirected graph, reflecting the degree of spatial association between the j-th abnormal suction cup and the p-th adjacent suction cup. The closer the distance, the greater the weight.

[0190] Theoretical displacement value It refers to the expected displacement value of the j-th abnormal suction cup estimated based on the current pose parameters, and is the benchmark for calculating the compensation displacement.

[0191] This embodiment ensures effective bonding of photovoltaic modules even when some suction cups malfunction, by adjusting the pressure and speed of the abnormal suction cups and setting compensating displacement for adjacent suction cups. The detailed working principle is as follows: During the bonding process in step S4, the controller collects the displacement sensor measurements of all suction cups in real time, and records the displacement measurement value of the j-th suction cup as δ. j (t); Simultaneously, real-time statistical analysis is performed on the displacement measurements of all suction cups to calculate the average displacement of all suction cups at time t. The mean value reflects the overall adhesion of the suction cups, while the standard deviation reflects the consistency of adhesion of each suction cup.

[0192] The controller presets the statistical outlier coefficient k and the second preset time threshold T. abn For each suction cup, an anomaly determination is made based on its displacement measurement value: if the displacement value of a certain suction cup j satisfies |δ j (t)- |>k·δ(t), meaning the displacement value of the suction cup deviates from the overall average value beyond the statistically reasonable range, and this state continues for more than the second preset time threshold T. abn If the controller detects an abnormal displacement of the suction cup j, it indicates that the suction cup has problems such as poor adhesion, stuck buffer rod, or sensor failure.

[0193] When suction cup j is determined to be in an abnormal displacement state, the controller first reconstructs the basic control parameters of the abnormal suction cup and takes two adjustment measures to avoid the abnormal suction cup from affecting the overall bonding effect: (1) Adsorption pressure adjustment: adjust the target pressure value P of the vacuum generator corresponding to suction cup j. target,j Adjust to P target,j The formula is adjusted as follows, P target,j '=α·P target,j , where 0<α<1 is the pressure attenuation coefficient, reducing the target negative pressure value of the abnormal suction cup, avoiding vacuum leakage caused by poor adhesion of the suction cup, and ensuring the negative pressure stability of other normal suction cups; (2) Adhesion speed adjustment: the speed command v of the adhesion operation of the pickup body at the driving end is generated. cmd At that time, a velocity attenuation field is applied to the local region corresponding to suction cup j, so that the normal velocity v in this region in the end-effector body coordinate system approaches the velocity v. n,jSatisfy v n,j =β·v n,default Where 0 < β < 1 is the speed attenuation coefficient; reducing the downward pressure speed in the abnormal suction cup area avoids damage to the photovoltaic module or suction cup caused by hard contact, and at the same time allows time for the adhesion adjustment of the abnormal suction cup.

[0194] The controller then abstracts the entire suction cup array into an undirected graph G=(V,E), where each vertex V of the undirected graph uniquely corresponds to a suction cup, and an edge E is set between any two adjacent suction cups. The existence of an edge indicates that the two suction cups have a spatial adjacency relationship. When suction cup j is determined to be abnormal, the controller quickly retrieves all its adjacent suction cups based on the undirected graph, forming an adjacent suction cup set N(j), which provides a basis for subsequent compensation displacement settings.

[0195] To compensate for insufficient adhesion of abnormal suction cup j and ensure the uniformity of overall adhesion on the photovoltaic module surface, the controller sets a compensation displacement δ for each adjacent suction cup p∈N(j). comp,p And it is calculated in real time using the following formula: This formula ensures that the compensation displacement of adjacent suction cups is proportional to the fitting deviation of abnormal suction cups and to the degree of spatial correlation between the two, thus achieving precise local compensation.

[0196] The controller refactors the bonding completion judgment conditions for normal suction cups (especially adjacent suction cups of abnormal suction cups), adjusting the bonding completion judgment condition for adjacent suction cup p to: δ p (t)≥δ th +δ comp,p And maintain the preset stable duration; that is, adjacent suction cups need to complete an additional compensation displacement δ on top of reaching the original adhesion displacement threshold. comp,p The compression, through the excessive adhesion of adjacent suction cups, compensates for the insufficient adhesion of abnormal suction cups, ensuring effective adhesion between the entire surface of the photovoltaic module and the end effector suction cup plane, thus providing a guarantee for subsequent vacuum adsorption.

[0197] The above solution uses statistical analysis to determine the displacement outliers of a single suction cup by using the mean and standard deviation. Combined with the constraint of time threshold, it makes the anomaly judgment more objective and accurate, avoiding misjudgments caused by mechanical vibration and instantaneous interference. At the same time, it can effectively identify real anomalies such as poor adhesion and jamming.

[0198] By reducing the target adsorption pressure of the abnormal suction cup, vacuum leakage caused by poor adhesion is avoided, ensuring the negative pressure stability of other normal suction cups; by setting a velocity decay field to reduce the downward pressure speed in the abnormal area, damage to photovoltaic modules or suction cups caused by hard contact is avoided, and the negative impact of the abnormal suction cup is controlled within a local area.

[0199] By setting compensation displacement for adjacent suction cups of abnormal suction cups and reconstructing their bonding judgment conditions, the excessive bonding of adjacent suction cups can compensate for the insufficient bonding of abnormal suction cups. This effectively solves the problem of uneven bonding of photovoltaic modules caused by local suction cup abnormalities, ensuring effective bonding between the entire surface of the photovoltaic module and the end-effector suction cup plane, and providing a guarantee for the stability of subsequent vacuum adsorption.

[0200] To address the issue of abnormal suction cup displacement, control parameters are reconstructed from three dimensions: pressure, speed, and displacement. Simultaneously, local compensation is performed by combining the topology of the suction cup array, making fault-tolerant control more refined and targeted. It can adapt to local suction cup abnormalities without requiring significant adjustments to the overall bonding operation.

[0201] In the event of abnormal displacement of some suction cups, the bonding operation can continue without interruption. The operation can be maintained through parameter reconstruction and local compensation, which greatly improves the robustness of the intelligent end effector in complex working environments, reduces the interruption of operation caused by abnormality of local components, and improves the efficiency of photovoltaic module installation.

[0202] In some embodiments, a fixed connecting plate 15 is installed above the longitudinal aluminum profile of the end effector body. The fixed connecting plate 15 is a connecting component between the end effector body and the photovoltaic module laying robot. It is used to firmly and accurately install the entire intelligent end effector at the end of the laying robot, realize the transmission of power and control signals between the robot and the end effector, and form the connection basis for the collaborative work between the intelligent end effector and the laying robot, thus ensuring the movement stability of the end effector during the handling and laying process.

[0203] Each adsorption control unit is equipped with a pressure gauge 24 in its vacuum pipeline. The pressure gauge 24 is electrically connected to the controller and can monitor the negative pressure value in the vacuum pipeline in real time and transmit the negative pressure data to the controller in real time. The controller can judge the working status of the vacuum adsorption system in real time based on the monitoring data of the pressure gauge. When the negative pressure value is lower than the preset adsorption threshold, it will issue an alarm in time and take measures such as pressure replenishment and shutdown. When the negative pressure value reaches the release threshold, it will control the solenoid valve 23 to open and complete the component release, which further improves the control accuracy and safety of the vacuum adsorption system.

[0204] After the vacuum adsorption system successfully picks up the photovoltaic module, when the end effector moves the module to the preset installation location, the controller controls the solenoid valve to switch on and off, allowing the suction cup to connect with the atmosphere. This eliminates the negative pressure between the suction cup and the photovoltaic module, completing the release of the module. Throughout the adsorption-release process, the controller monitors the pressure changes of the vacuum system in real time using a pressure gauge. Based on these pressure changes, the controller actively controls the adsorption and release of the suction cup, achieving closed-loop pressure control of the vacuum adsorption system and further improving the accuracy of adsorption and release.

[0205] The laser rangefinder sensor 31 is fixed to both sides of the longitudinal aluminum profile of the end effector body by the laser mounting bracket 16. The laser mounting bracket has angle and position adjustment functions, which can flexibly adjust the installation angle and position of the laser rangefinder sensor according to the size and installation height of the photovoltaic module, ensuring that the sensor can accurately illuminate the measurement point on the surface of the photovoltaic module, thereby improving the installation flexibility and measurement accuracy of the laser rangefinder sensor.

[0206] The position and angle information of the photovoltaic modules collected by the laser ranging and positioning system, as well as the calculated pose parameters, can all be visualized through the display module of the controller. Operators can use this visualized information to manually verify the pose adjustment and adsorption operation of the end effector without directly observing the surface of the photovoltaic modules. Even in complex outdoor environments where direct observation is not possible, such as strong light, backlight, and dust, the operation verification can still be completed, improving the convenience and safety of operation.

[0207] Finally, it should be noted that although the above embodiments have been described in the description and drawings of this invention, this should not limit the scope of patent protection of this invention. Any technical solutions that are based on the essential concept of this invention, utilize the content described in the description and drawings of this invention to make equivalent structural or procedural substitutions or modifications, as well as the direct or indirect application of the technical solutions of the above embodiments to other related technical fields, are all included within the scope of patent protection of this invention.

Claims

1. A smart end-sensor for photovoltaic module installation, characterized in that, include: End effector body; A vacuum adsorption system includes multiple suction cups, which are mounted on the end effector body and are used to adsorb photovoltaic modules. A laser ranging and positioning system is installed on the end effector body and is used for non-contact measurement of the relative pose between the end effector body and the surface of the photovoltaic module. A displacement detection system, installed on the end effector body, is used to detect the displacement of the suction cup relative to the end effector body in order to determine the adhesion status between the suction cup and the photovoltaic module. The controller is electrically connected to the vacuum adsorption system, the laser ranging and positioning system, and the displacement detection system, respectively. The laser ranging and positioning system includes: At least four laser ranging sensors are deployed on the end effector body to measure the distance from the laser ranging sensors to different measurement points on the surface of the photovoltaic module, thereby obtaining distance measurement signals. The signal processing unit is used to receive and process the distance measurement signals of each of the laser ranging sensors to obtain distance data; The controller is configured to execute the following control flow: S1: Receive distance data sent by the signal processing unit; S2: Based on the fixed installation coordinates of each laser ranging sensor in the coordinate system of the end effector body and their corresponding distance data, the current pose parameters of the photovoltaic module surface relative to the end effector body are determined through spatial geometric calculation. The current pose parameters include at least the center position deviation and the tilt angle. S3: Generate a pose adjustment command based on the current pose parameters to drive the end effector body to move, so as to reduce the center position deviation and tilt angle until the center position deviation and tilt angle reach the preset pose tolerance range. S4: After the current pose parameters reach the pose tolerance range, control the end effector body to perform the bonding action, and monitor the displacement of the suction cup in real time through the displacement detection system. When the displacement of all the suction cups reaches the preset bonding displacement threshold, and maintains a preset stable time after reaching the threshold, it is determined that the bonding is complete. S5: After the bonding is determined, the vacuum adsorption system is started to complete the gripping of the photovoltaic module.

2. The intelligent end-sensor for photovoltaic module installation as described in claim 1, characterized in that, The vacuum adsorption system includes at least two independent adsorption control units; The plurality of suction cups are assigned to each of the adsorption control units, such that each adsorption control unit controls at least one of the suction cups; Each of the aforementioned adsorption control units further includes: at least one vacuum pump and at least one vacuum tank connected to the outlet of the vacuum pump; Each suction cup belonging to the same set of adsorption control unit is connected to the vacuum tank through an independent vacuum branch. Each vacuum branch is provided with a vacuum filter and a solenoid valve in sequence along the vacuum flow direction. The controller is electrically connected to the vacuum pump in each of the adsorption control units and the solenoid valve on each of the vacuum branches, and is used to independently control the start and stop of each of the adsorption control units and to independently control the on and off of each of the vacuum branches.

3. The intelligent end-sensor for photovoltaic module installation as described in claim 1, characterized in that, The displacement detection system includes an induction switch and an induction plate disposed on each of the suction cups; The suction cup is slidably mounted on the end effector body via a buffer rod, and the sensing plate is disposed on the buffer rod; When the suction cup contacts and is pressed against the surface of the photovoltaic module, it moves the buffer rod and the sensing sheet to trigger the inductive switch, so that the inductive switch generates a signal indicating that the suction cup is in place.

4. The intelligent end-sensor for photovoltaic module installation as described in claim 3, characterized in that, The displacement detection system also includes a micro grating ruler integrated inside each of the buffer rods, which is used to measure the absolute compressive displacement d of the buffer rod relative to its mounting base; The controller is configured to perform a two-level fusion judgment in step S4, specifically including: When the sensor switch of any suction cup is triggered, it is determined to be a primary fit, and the reading of the laser rangefinder corresponding to that suction cup at the current moment is recorded as the reference distance value L0. After the initial bonding, the controller switches to reading the data of the micro grating ruler and continues to control the end effector body to perform micro-pressure. During the micro-pressure process, the following calculations are performed in real time for each suction cup: obtaining a reading L of a laser distance sensor at the current time t; The surface normal displacement δ of the photovoltaic module calculated from the laser measurement value laser , the calculation formula is as follows: δ laser = L0- L t ; Obtain the compression displacement d of the buffer rod measured by the micro grating ruler at the current moment, and calculate the coupling difference Δ=δ according to the following formula. laser -d; The final fine bonding is considered complete when both of the following conditions are met: The compression displacement d of the buffer rod of all suction cups reaches the preset fine fit displacement threshold. The absolute values ​​of the coupling differences Δ calculated by all suction cups are all less than the preset deformation compatibility threshold ε; the deformation compatibility threshold ε is used to determine whether the photovoltaic module has undergone local elastic deformation beyond the permissible range or has poor contact.

5. The intelligent end-sensor for photovoltaic module installation as described in claim 1, characterized in that, The end effector body is a rectangular frame structure, which includes at least two parallel longitudinal aluminum profiles and multiple transverse aluminum profiles connected between the two parallel longitudinal aluminum profiles. The aluminum profiles are connected by corner brackets. At least four laser ranging sensors of the laser ranging and positioning system are respectively installed in the four corner areas of the rectangular frame, and are used to measure the vertical distance between the sensor and the four corner areas. The vacuum adsorption system has multiple suction cups mounted on the horizontal aluminum profile via their respective buffer rods. The horizontal aluminum profile is provided with a fixing seat for mounting and adjusting the buffer rods. The signal processing unit filters and denoises the received distance measurement signal, and sends the filtered and denoised distance data to the controller for spatial geometry calculation.

6. The intelligent end-sensor for photovoltaic module installation as described in claim 5, characterized in that, The controller calculates the current pose parameters in the following manner: Based on the fixed installation coordinates of each laser ranging sensor in the coordinate system of the end effector and the distance data measured by them, the three-dimensional coordinates of at least four measuring points on the surface of the photovoltaic module in the coordinate system of the end effector are calculated. Based on the three-dimensional coordinates of the at least four measurement points, the plane equation of the photovoltaic module surface is fitted. Based on the plane equation, the center position deviation and tilt angle of the photovoltaic module surface are calculated.

7. The intelligent end-sensor for photovoltaic module installation as described in claim 1, characterized in that, Also includes: An environmental sensing module, integrated on the end effector body, is used to acquire environmental wind field information and motion state information of the end effector body in real time. A drive actuator, connected to the controller, is used to drive the end effector body to move according to the instructions of the controller; The controller is further configured to synchronously perform feedforward wind load dynamic compensation control when executing step S3 and / or step S4, specifically including: The environmental sensing module acquires in real time the ambient wind speed V acting on the end effector body and the adsorbed photovoltaic module. w Wind direction angle θ w , and the triaxial angular velocity and acceleration of the end effector body; Based on the tilt angle in the current pose parameters and the ambient wind speed V w With wind direction angle θ w And a pre-existing aerodynamic parameter model of the photovoltaic module in the controller, to calculate in real time the predicted normal force F generated by wind load in the normal direction of the photovoltaic module plane. wind and the predicted normal force F wind The predicted moment vector M in the component plane wind The aerodynamic parameter model includes at least the windward area of ​​the photovoltaic module, the location of the wind pressure center, and the aerodynamic coefficient. The controller is based on the predicted torque vector M wind， Based on the dynamic model of the drive actuator, the feedforward compensation algorithm built into the controller is used to calculate the compensation method for M. wind Required compensation torque command ΔM comp ; In step S3 and / or step S4, the final control command received by the drive actuator is: the basic control command and the compensation torque command ΔM. comp The superposition of the basic control commands is generated by the pose adjustment algorithm in step S3 or the control logic that generates the fitting control commands in step S4.

8. The intelligent end-sensor for photovoltaic module installation as described in claim 7, characterized in that, The feedforward compensation algorithm calculates the compensation torque command ΔM through the following steps. comp : The predicted torque vector M wind The transformed disturbance torque vector M is obtained by transforming the coordinate system based on the photovoltaic module plane to the joint space coordinate system or the drive motor coordinate system of the drive actuator. wind' ; Based on the dynamic model of the drive actuator, the torque vector M to counteract the disturbance is calculated. wind' The required joint drive torque or motor current command is used as an initial compensation amount. The initial compensation amount is filtered to match the frequency band characteristics of the filtered initial compensation amount with the actual response bandwidth of the drive actuator, and then multiplied by an adjustable feedforward gain coefficient K. ff Generate the compensation torque command ΔM comp K ff The value range is (0,1).

9. The intelligent end-sensor for photovoltaic module installation as described in claim 1, characterized in that, The controller also integrates a distributed fault diagnosis and fault tolerance control module, which is configured to perform the following steps: For the i-th laser rangefinder, its instantaneous measurement value is defined as d. i (t), within the time window T w Calculate its rate of change The difference Δd between adjacent periods i ; like The data validity flag of the laser rangefinder sensor is set to invalid if any of the following conditions are met: d i (t) remains equal to zero, has a maximum value, or a specific error code exceeds N. err One sampling period; | |>V max , where V max The limit distance change rate is calculated based on the maximum speed of the end effector body. ∣Δd i |>ΔD max , where ΔD max This is the maximum reasonable jump threshold set based on the mechanical vibration characteristics; If the valid data flag of a certain laser rangefinder remains invalid for a duration exceeding the first preset time threshold T, fault If the laser rangefinder is determined to be in a hard failure state, it will be removed from the set of currently valid laser rangefinders S. valid Remove from; Let the total number of laser rangefinders be N, and define the effective set of laser rangefinders as S. valid The number of elements is M; If M≥4, then directly use S valid Installation coordinates (x) of all laser rangefinders i ,y i ,z i ) body and its measured value d i By fitting the plane using the least squares method or using the SVD decomposition method, the equation ax+by+cz+d=0 of the photovoltaic module plane in the coordinate system of the end-sensor is solved, and the tilt angle and the deviation of the center position are obtained. If M=3, a spatial plane is determined using data from three effective laser rangefinders. Combined with the encoder feedback value of the drive actuator, the expected pose of the end effector body is obtained through forward kinematics calculation. The expected pose is then fused with the laser plane equation, and the complete current pose parameters are estimated using a Kalman filter. If M < 3, it is determined that the pose cannot be reliably calculated at present, the operation is immediately suspended and a fault alarm message indicating that manual intervention is required is issued.

10. The intelligent end-sensor for photovoltaic module installation as described in claim 9, characterized in that, During the bonding process in step S4, for the j-th suction cup, the displacement sensor measurement value of that suction cup is defined as δ. j (t); Calculate the average value of all suction cup displacements. and standard deviation δ(t); If |δ j (t)- |>k·δ(t) and continuously exceeds the second preset time threshold T abn If the displacement of suction cup j is abnormal, then k is determined to be a preset statistical outlier coefficient. When suction cup j is determined to be abnormal, the following control parameter reconstruction steps are performed: The target pressure value P of the vacuum generator corresponding to suction cup j. target,j Adjust to P target,j The formula is adjusted as follows, P target,j '=α·P target,j Where 0 < α < 1 is the pressure attenuation coefficient; Generate the speed command that drives the end effector body to perform the contact action. At that time, a velocity attenuation field is applied to the local region corresponding to suction cup j, so that the normal velocity v of this local region in the body coordinate system of the end effector approaches the velocity v. n,j Satisfy v n,j =β·v n,default Where β is the velocity decay coefficient, 0 < β < 1; The suction cup array is abstracted as an undirected graph G=(V,E), where vertex V represents the suction cup and edge E represents the adjacency relationship; Once suction cup j is determined to be abnormal, iterate through all its adjacent suction cup sets N(j); For each adjacent suction cup p∈N(j), the self-adhesion displacement threshold δ is reached at that adjacent suction cup. th Then, an additional compensation displacement δ is applied. comp,p The calculation formula is as follows: ； Where γ is the global compensation gain. Let (j,p) be the weight of edge (j,p). This is the theoretical displacement value of the suction cup j estimated based on the current pose parameters; In step S4, the controller adjusts the condition for determining the completion of adhesion of the normal suction cup p to: δ p (t)≥δ th +δ comp,p And it maintains a stable duration.

Images