Automatic tightening method for vehicle screws

Abstract

The invention discloses an automatic tightening method for vehicle screws, and relates to the technical field of automatic control, and the method comprises the steps: obtaining the dynamic information of a target vehicle and the initial position information of all target screw holes; based on the dynamic information, the mobile platform is controlled to perform synchronous following motion, so that the mobile platform and the target vehicle keep a preset relative position relationship; based on the initial position information, an execution mechanism installed on the mobile platform is controlled to position a preset target screw hole, the actual position of the preset target screw hole is obtained, and the preset target screw hole is determined from all the target screw holes based on a preset screening rule; determining target positions of all target screw holes based on the actual position and initial position information of the preset target screw holes; and the execution mechanism is controlled to move to target positions corresponding to all the target screw holes, and all the target screws are tightened.

Description

Technical Field

[0001] This application relates to the field of automation control technology, and in particular to an automated method for tightening vehicle screws. Background Technology

[0002] On vehicle assembly lines, screw tightening is typically required on moving vehicle chassis. Currently, common solutions rely on robots at fixed stations or tightening devices moving along pre-set tracks. However, because vehicles are constantly moving on the assembly line, these solutions cannot achieve real-time synchronization with the moving vehicles. This results in relative displacement between the robot's workspace and the vehicle chassis, making it difficult to accurately position and tighten the screws, thus limiting the efficiency and reliability of automated tightening. Therefore, an automated screw tightening method for vehicles is urgently needed to solve the aforementioned technical problems. Summary of the Invention

[0003] The summary section introduces a series of simplified concepts, which will be further explained in detail in the detailed description section. This summary section is not intended to limit the key and essential technical features of the claimed technical solutions, nor is it intended to determine the scope of protection of the claimed technical solutions.

[0004] In a first aspect, this application provides an automated method for tightening vehicle bolts, comprising: Obtain the dynamic information of the target vehicle and the initial position information of all target bolt holes; Based on the dynamic information, the mobile platform is controlled to perform synchronous following motion, so that the mobile platform and the target vehicle maintain a preset relative positional relationship. Based on the initial position information, the actuator installed on the mobile platform is controlled to position the preset target screw hole to obtain the actual position of the preset target screw hole, wherein the preset target screw hole is determined from all target screw holes based on a preset screening rule; Based on the actual position of the preset target screw hole and the initial position information, determine the target position of all target screw holes; The actuator is controlled to move to the target position corresponding to all target screw holes, and all target screws are tightened.

[0005] In some implementations, acquiring the dynamic information of the target vehicle includes: Image sequences of the target vehicle are acquired using a visual capture device; Based on the image sequence, determine the pixel displacement of preset feature points in the target vehicle; The dynamic information of the target vehicle is determined based on the pixel displacement and the acquisition time interval of the image sequence.

[0006] In some implementations, obtaining the initial position information of all target screw holes includes: The target vehicle is identified by its vehicle model to determine its vehicle model identifier. Based on the vehicle model identifier, the corresponding chassis 3D model is determined from the pre-stored vehicle model database, wherein the chassis 3D model contains the initial position coordinates of all target bolt holes; Based on the chassis 3D model, the initial position information of all target screw holes is generated.

[0007] In some embodiments, controlling the actuator mounted on the mobile platform to position a preset target screw hole based on the initial position information, thereby obtaining the actual position of the preset target screw hole, includes: Based on the initial position information and the preset filtering rules, a preset target screw hole and its corresponding initial position coordinates are determined from all target screw holes; Based on the initial position coordinates and the preset observation range threshold, the preset observation area is determined; Move the vision sensor installed at the end of the actuator to the preset observation area; The visual sensor is used to acquire a local image of the preset observation area; The local image is processed based on a preset image recognition algorithm to determine the actual image coordinates of the preset target screw hole in the image coordinate system; Based on the actual image coordinates and the preset calibration parameters of the vision sensor, the actual position of the preset target screw hole in the spatial coordinate system is determined.

[0008] In some embodiments, processing the local image based on a preset image recognition algorithm to determine the actual image coordinates of the preset target screw hole in the image coordinate system includes: The local image is subjected to grayscale conversion and noise filtering to obtain a preprocessed image; Edge detection is performed on the preprocessed image to obtain edge features; Based on the preset screw hole contour features, the edge features are subjected to circular detection to determine the candidate circle center pixel coordinate set of the preset target screw hole; From the set of candidate center pixel coordinates, the candidate center pixel coordinates that are closest to the projected coordinates of the initial position coordinates in the image plane are selected as the actual image coordinates of the preset target screw hole in the image coordinate system.

[0009] In some implementations, determining the target position of all target screw holes based on the actual position of the preset target screw hole and the initial position information includes: Based on the actual position of the preset target screw hole and the initial position coordinates corresponding to the preset target screw hole in the initial position information, the spatial transformation parameters are determined; Based on the spatial transformation parameters and the initial position coordinates of all target screw holes in the initial position information, coordinate correction calculations are performed on the initial position coordinates of all target screw holes to determine the target positions of all target screw holes.

[0010] In some implementations, the step of performing coordinate correction calculations on the initial position coordinates of all target screw holes based on the spatial transformation parameters and the initial position coordinates corresponding to all target screw holes in the initial position information, to determine the target positions of all target screw holes, includes: Based on the spatial transformation parameters, construct the coordinate transformation matrix; Based on the coordinate transformation matrix, a coordinate transformation operation is performed on the initial position coordinates of each target screw hole in the initial position information to determine the corrected position coordinates of each target screw hole; Based on the corrected position coordinates of all target screw holes, the target positions of all target screw holes are determined.

[0011] In some embodiments, the actuator includes a parallel platform and a series robotic arm mounted on the parallel platform, and after determining the target positions of all target screw holes, it further includes: For each target screw hole among all target screw holes, based on the target position of the target screw hole, determine the first pose adjustment parameter of the parallel platform and the second pose adjustment parameter of the serial robotic arm; Based on the first pose adjustment parameters, the parallel platform is controlled to perform a first degree of freedom motion; Based on the second pose adjustment parameters, the serial robotic arm is controlled to perform a second degree of freedom movement so that the tightening shaft installed at the end of the serial robotic arm is aligned with the target screw hole.

[0012] In some embodiments, controlling the actuator to move to the target position corresponding to all target screw holes and tightening all target screws includes: For each of the target screw holes, the actuator is controlled to move to the target position of that target screw hole; When the actuator moves to the target position of the target screw hole, the screw feeding mechanism is controlled to deliver the target screw corresponding to the target screw hole to the preset pickup position; When the target screw reaches the preset pickup position, the tightening shaft at the end of the actuator is controlled to pick up the target screw and screw the target screw into the target screw hole; During the screwing process of the target screw, the output torque and rotation angle of the tightening shaft are controlled by a dual closed-loop servo system based on a preset tightening strategy, so that the target screw reaches the target preload.

[0013] In some embodiments, after the tightening operation of all target screws, the method further includes: Control the mobile platform to move to a preset charging station and charge the mobile platform.

[0014] Secondly, this application proposes an automated tightening device for vehicle screws, comprising: The screw hole data acquisition unit is used to acquire the dynamic information of the target vehicle and the initial position information of all target screw holes; The platform position determination unit is used to control the mobile platform to perform synchronous following motion based on the dynamic information, so that the mobile platform and the target vehicle maintain a preset relative position relationship; A screw hole position correction unit is used to control an actuator installed on the mobile platform to position a preset target screw hole based on the initial position information, thereby obtaining the actual position of the preset target screw hole, wherein the preset target screw hole is determined from all target screw holes based on a preset screening rule; The screw hole position determination unit is used to determine the target position of all target screw holes based on the actual position of the preset target screw hole and the initial position information; The screw tightening operation unit is used to control the actuator to move to the target position corresponding to all target screw holes and tighten all target screws.

[0015] In summary, the automated screw tightening method for vehicles provided in this application offers the ability to perceive the state of the moving carrier and the spatial distribution of the work targets by acquiring the dynamic information of the target vehicle and the initial position information of all target screw holes. Based on the dynamic information, the mobile platform is controlled to synchronously follow the target vehicle's movement, enabling the mobile platform to match the target vehicle's motion state in real time and maintain a stable preset relative positional relationship with the target vehicle. This operation overcomes the problem of work space misalignment caused by continuous movement on the assembly line, providing the actuator with a stationary working reference relative to the vehicle body, ensuring the possibility of assembly in dynamic environments. Based on the initial position information, the actuator is controlled to position the preset target screw holes selected according to preset screening rules. By acquiring the actual position of this single-point screw hole, the abstract coordinates in the theoretical model are transformed into a specific spatial point that can be perceived and measured on the current vehicle entity. This process completes the mapping from digital model to physical entity with minimal perception cost, reducing the computational and time overhead required for full positioning. Based on the actual and initial positions of the preset target screw holes, the target positions of all target screw holes are determined by solving the mapping relationship between the two. This generalizes the single-point measurement error compensation capability to all screw holes in the vehicle chassis, enabling each target screw hole to obtain working coordinates that accurately match the current vehicle's actual pose. This method completes the systematic correction of assembly deviations across the entire vehicle with only one visual measurement, improving the positioning efficiency and accuracy in multi-hole collaborative operation scenarios. The control actuator moves to the target positions corresponding to all target screw holes and performs tightening operations on the target screws. The spatial coordinate commands established in the preceding steps are transformed into actual physical motion and mechanical operations, allowing the screws to be inserted into the error-compensated screw holes and tightened. This effectively improves the efficiency of vehicle chassis screw assembly operations while reducing reliance on manual operation experience and physical strength. Attached Figure Description

[0016] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit this specification. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 This is a schematic flowchart of an automated method for tightening vehicle screws, provided in an embodiment of this application. Figure 2 This is a schematic diagram of an automated screw tightening device for vehicles, provided as an embodiment of this application. Detailed Implementation

[0017] The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus. The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them.

[0018] The automated screw tightening method for vehicles provided in this application is mainly applied to the final assembly lines of automobiles and commercial vehicles, especially in critical workstations such as chassis assembly, powertrain installation, and battery pack installation, where multiple screws need to be tightened on the bottom of a moving vehicle on the assembly line. In this application scenario, the vehicle is carried by a conveyor line and moves continuously at a constant speed. Traditional fixed automated equipment or moving devices relying on preset tracks cannot be used directly due to the relative displacement between the workspace and the vehicle. Manual operation faces problems such as working with one's head tilted up and high labor intensity. This method aims to achieve fully automated screw insertion and high-precision tightening of all screw holes in the chassis while the vehicle is in continuous motion by combining a strategy of synchronous following of the moving platform, single-hole visual positioning, and full-vehicle coordinate correction. This meets the requirements of modern automobile manufacturing for assembly automation, flexibility, and high reliability.

[0019] To facilitate understanding of this solution, the following explanations are provided for certain specific terms used in this application.

[0020] The dynamic information in this application refers to data reflecting the real-time motion status of the target vehicle on the assembly line, including but not limited to the vehicle's speed, direction of movement, and pose changes in two-dimensional or three-dimensional space. This information is obtained by continuously acquiring images and calculating pixel displacements of preset feature points on the vehicle chassis using a high-frequency vision capture device, and serves as the input basis for the mobile platform to achieve synchronous following motion.

[0021] The initial position information in this application refers to the spatial coordinate data of all target bolt holes under ideal design conditions, which are pre-extracted and stored based on the three-dimensional digital model of the chassis of the target vehicle model. This information originates from the geometric model during the product design phase, characterizes the theoretical position of the bolt holes in the vehicle coordinate system, and does not include actual offsets introduced by manufacturing tolerances, jig runout, or assembly line positioning deviations, providing a global spatial index and expected observation benchmark for positioning operations.

[0022] The mobile platform described in this application refers to an automated mobile base with autonomous navigation and motion control capabilities, whose bottom integrates sensors such as a drive unit, lidar, and odometer. The core function of this platform is to receive dynamic information from the target vehicle and calculate its own motion commands in real time. By dynamically adjusting its speed and direction, it maintains a constant preset relative position with the continuously moving target vehicle on the assembly line, providing a reference coordinate system that is stationary relative to the vehicle body for the work unit mounted on top.

[0023] The synchronous following motion described in this application refers to the process by which a mobile platform autonomously adjusts its own speed and trajectory based on real-time acquired dynamic information of the target vehicle, maintaining a stable spatial relative pose with the moving target vehicle in a global coordinate system. The direct result of this motion is that the workspace of the actuator and the vehicle chassis reach a relatively static state, transforming the dynamic assembly problem into a quasi-static operation problem.

[0024] The actuator in this application refers to an electromechanical integrated device integrated and installed on a mobile platform for performing screw hole alignment and screw tightening operations. The mechanism includes at least a multi-degree-of-freedom robotic arm, an end effector tightening shaft, a vision sensor, and an auxiliary control unit, and its functions cover spatial point movement, visual observation and recognition, screw picking, screw hole alignment, and tightening action execution.

[0025] The preset target screw hole in this application refers to a representative screw hole selected from all target screw holes involved in the current task according to preset screening rules, used for performing the initial visual positioning and error calculation. This screw hole is typically selected in areas with high chassis structural rigidity, strong feature recognizability, and global representativeness in spatial distribution, such as screw holes near powertrain mounting points or the center of symmetry of the frame longitudinal beams. Its function is to establish a spatial mapping relationship between the theoretical model and the physical vehicle with minimal perceptual cost.

[0026] The target screw holes in this application refer to the threaded mounting holes on the vehicle chassis where screws need to be installed, and are the direct targets of the tightening operation. All target screw holes constitute the complete set of workstations for the current task, and their spatial distribution is determined by the vehicle model and is pre-marked in coordinate form in the chassis 3D model.

[0027] The preset screening rules in this application refer to pre-defined logical criteria or algorithmic strategies used to uniquely determine a preset target screw hole from all target screw holes. These rules can be formulated based on a combination of factors, including the screw hole's geometric position in a 3D model, process priority, ease of visual recognition, or historical operational data.

[0028] The actual position in this application refers to the three-dimensional spatial coordinates of the screw hole in the current vehicle spatial coordinate system, calculated after actual observation and image recognition of the preset target screw hole by the vision sensor at the end of the actuator. This position reflects the actual offset between the theoretical model and the physical entity caused by factors such as vehicle assembly errors, assembly line vibrations, or tooling deviations.

[0029] The target position in this application refers to the working coordinates of each target screw hole, determined by coordinate correction calculation after obtaining the actual position of the preset target screw hole and solving the spatial transformation parameters, and matching the screw hole in the current actual vehicle posture. This position is the basis for the center point of the actuator moving above the screw hole and aligning with it.

[0030] The spatial transformation parameters in this application refer to the mathematical description set obtained by solving the geometric mapping relationship between the actual position of the preset target screw hole and its initial position coordinates. This parameter set describes the rigid body transformation relationship from the theoretical design coordinate system to the current vehicle actual pose coordinate system, and its specific form includes, but is not limited to, rotation matrices and translation vectors. This parameter set is used to correct the initial position coordinates of all remaining target screw holes.

[0031] The tightening shaft in this application refers to a specialized work tool installed at the end of an actuator for directly performing screw picking, screwing, and tightening actions. This shaft is driven by a servo motor and integrates a torque sensor, angle encoder, and electromagnetic adsorption sleeve, providing high-precision output and real-time feedback capabilities.

[0032] The screw feeding mechanism of this application refers to an automatic screw supply device with a clip-type hopper design, in which target screws are arranged and stored sequentially. The mechanism uses a pneumatic pusher to deliver the screws one by one to a preset pickup position, which works in conjunction with the electromagnetic adsorption sleeve of the tightening shaft to complete the automated gripping of the screws, eliminating the need for manual loading of each screw individually.

[0033] The dual closed-loop servo control described in this application refers to a high-precision real-time feedback control strategy employed during the screw insertion and tightening stage. This strategy simultaneously establishes and operates two independent closed-loop control loops: a torque loop uses a torque sensor mounted on the tightening shaft as feedback to adjust and limit the output torque in real time; an angle loop uses an encoder built into the servo motor as feedback to measure and control the rotation angle at the endpoint. The two loops work together to ensure that the screw continues to rotate to a preset angle after reaching the engagement state, thereby obtaining the target preload by controlling the bolt tension.

[0034] Please see Figure 1 This is a schematic flowchart of an automated vehicle screw tightening method provided in an embodiment of this application, including: S110. Obtain the dynamic information of the target vehicle and the initial position information of all target bolt holes; For example, when the target vehicle enters the preset work station, the visual capture device continuously acquires images of preset feature points on the target vehicle chassis to obtain an image sequence containing multiple frames. Based on the pixel displacement of the preset feature points between adjacent frames and the acquisition time interval of the image sequence, dynamic information reflecting the real-time speed and direction parameters of the target vehicle is calculated. At the same time, the visual recognition device identifies the vehicle model to determine the vehicle model identifier. Based on the vehicle model identifier, the corresponding chassis 3D model is retrieved from the pre-stored vehicle model database. The initial position coordinates of all target screw holes are extracted from the chassis 3D model to generate the initial position information of all target screw holes.

[0035] S120. Based on dynamic information, control the mobile platform to perform synchronous following motion, so that the mobile platform and the target vehicle maintain a preset relative positional relationship. For example, the mobile platform receives dynamic information reflecting the real-time pose changes of the target vehicle and calculates and generates its own motion control commands based on this information. According to these commands, the mobile platform drives its walking mechanism to adjust its speed and direction, matching its motion state in the global coordinate system with that of the target vehicle, maintaining a preset and stable spatial relative position with the target vehicle throughout the operation. This synchronized following motion allows the actuators mounted on the mobile platform to obtain a nearly stationary working reference frame relative to the moving vehicle chassis, providing a relatively static working environment for positioning and tightening operations.

[0036] S130. Based on the initial position information, control the actuator installed on the mobile platform to position the preset target screw hole to obtain the actual position of the preset target screw hole, wherein the preset target screw hole is determined from all target screw holes based on preset screening rules; For example, after the mobile platform completes synchronized following motion with the target vehicle and establishes a preset relative position relationship, the actuator, based on the spatial distribution data of all target screw holes in the initial position information, determines a preset target screw hole and its corresponding initial position coordinates from all target screw holes according to a preset filtering rule. Using the initial position coordinates of this preset target screw hole as a target guide, the actuator moves its end-mounted vision sensor to a spatial region near these initial position coordinates. The vision sensor acquires local images of this region, processes the acquired images using a preset image recognition algorithm, identifies the imaging features of the preset target screw hole in the image, calculates its coordinates in the image coordinate system, and, combined with the preset calibration parameters of the vision sensor, transforms the image coordinates to a spatial coordinate system to obtain the spatial coordinates of the preset target screw hole in the current actual pose of the vehicle. These spatial coordinates are then determined as the actual position of the preset target screw hole.

[0037] S140. Based on the actual and initial position information of the preset target screw holes, determine the target position of all target screw holes; For example, based on the actual position of the preset target screw hole and the initial position coordinates corresponding to the preset target screw hole in the initial position information, a spatial transformation parameter describing the deviation between the theoretical design coordinate system and the current vehicle actual pose coordinate system is calculated; the spatial transformation parameter is applied to the initial position coordinates corresponding to all target screw holes in the initial position information, and coordinate correction calculation is performed on the initial position coordinates of each target screw hole to match the theoretical design coordinates with the current actual pose of the vehicle, thereby obtaining the target positions of all target screw holes that are adapted to the current vehicle state.

[0038] S150: Control the actuator to move to the target position corresponding to all target screw holes and tighten all target screws.

[0039] For example, after calculating the target positions of all target screw holes, the actuator moves sequentially to the target position corresponding to each target screw hole according to the planned path. When the actuator reaches the target position of the current target screw hole, the screw feeding mechanism delivers the target screw to the preset pickup position. The tightening shaft at the end of the actuator picks up the target screw and screws it into the current target screw hole. During the screwing process, the tightening shaft performs dual closed-loop servo control on the output torque and rotation angle based on a preset tightening strategy, so that the target screw reaches the target preload, thereby completing the tightening operation of the target screw hole. The actuator repeats the above process until all target screw holes have been tightened.

[0040] In summary, this embodiment of the application obtains the dynamic information of the target vehicle and the initial position information of all target screw holes, and simultaneously establishes the perception of the real-time status of the moving carrier and the distribution of the work targets at the start of the operation. Based on the dynamic information, the mobile platform is controlled to follow the target vehicle synchronously, so that the speed and direction of the mobile platform are matched with the running trajectory of the target vehicle in real time, thereby maintaining a stable preset relative position relationship with the target vehicle. This step overcomes the relative displacement of the work area caused by the continuous transmission of the assembly line, and transforms the dynamic assembly environment into a quasi-static work scenario in which the actuator is stationary relative to the vehicle body, providing a stable physical platform for screw hole positioning and screw insertion.

[0041] Based on the initial position information, the control actuator locates a single preset target screw hole selected according to the preset screening rules. The precise spatial coordinates of the screw hole in the current actual pose of the vehicle are obtained through a vision sensor. This step completes the mapping from the digital theoretical model to the physical state of the physical vehicle at the cost of only performing a local observation once. This avoids the huge time overhead and computing power consumption caused by identifying all screw holes one by one in the traditional solution, and improves the efficiency of the initial perception stage.

[0042] Based on the actual and initial position information of the preset target screw holes, by solving the spatial transformation parameters and performing coordinate correction calculations on the initial position coordinates of all target screw holes, the error compensation capability obtained from single-point measurement is generalized to all target screw holes of the entire chassis. This step corrects the overall vehicle assembly deviation caused by factors such as vehicle manufacturing tolerances, assembly line jig sway, and tooling positioning deviations, under the premise of performing visual positioning only once. This enables each target screw hole to obtain working coordinates that match its current actual pose, improving the positioning accuracy and consistency in multi-screw hole collaborative operation scenarios.

[0043] The control actuator moves to the target position corresponding to all target screw holes, and at each target position, the target screw is automatically fed by the screw feeding mechanism, picked up by the tightening shaft, screwed in, and the output torque and rotation angle are controlled by dual closed-loop servo control during the screwing process. This step converts all the spatial commands established by the aforementioned perception, positioning and correction calculation into physical actions, so that the target screw can be reliably guided into the error-compensated screw hole and achieve the target preload. This improves the efficiency, accuracy and process stability of vehicle chassis assembly operations, while greatly reducing the dependence on manual operation experience and high-intensity overhead work.

[0044] The automated vehicle screw tightening system proposed in this application is physically embodied as an integrated mobile composite robot device. This device uses a mobile platform with autonomous navigation and motion control capabilities as its main structure. The platform chassis integrates environmental perception sensors such as LiDAR and odometers, as well as high-power drive wheels, enabling it to autonomously move and precisely position itself on the assembly workshop floor. The control function of the mobile platform is to receive and calculate the dynamic information of the target vehicle, and by adjusting its own speed and trajectory in real time, achieve synchronous following with continuously moving vehicles on the assembly line, establishing a static and stable spatial operating reference relative to the vehicle chassis for all working units mounted on the platform.

[0045] The mobile platform houses the actuator for tightening operations, specifically a multi-degree-of-freedom robotic arm system. To effectively address positioning deviations and low-frequency vibrations during vehicle assembly, the actuator base utilizes a six-degree-of-freedom Stewart parallel platform. This platform boasts high rigidity, high load-bearing capacity, and rapid dynamic response, enabling significant positional adjustments to compensate for vehicle assembly tolerances and assembly line vibrations. A three-degree-of-freedom precision serial robotic arm is mounted atop the parallel platform, used for millimeter-level or even sub-millimeter-level fine-tuning of the end-effector's position and attitude based on macroscopic positioning. Together, these two components form an actuator system with nine-degree-of-freedom dynamic error compensation capabilities. An integrated tool is mounted on the end flange of the serial robotic arm; this tool is a high-precision servo tightening shaft with a built-in torque sensor, angle encoder, and electromagnetic adsorption sleeve.

[0046] The perception module consists of multiple sets of vision devices working together. A fixed industrial 3D scanning camera is installed at the entrance station of the production line to acquire images of the vehicle chassis entering the work area. Automatic vehicle model recognition is achieved through a vehicle model identification model, and the corresponding 3D digital model of the chassis is retrieved from the database. This model has the theoretical spatial coordinates of all the screw holes to be tightened pre-set. The mobile platform or robotic arm is equipped with an ultra-high frequency industrial camera with a frame rate of no less than 1000Hz, continuously acquiring images of preset feature points on the vehicle chassis. By analyzing the pixel displacement of feature points in consecutive frames, the vehicle's speed and direction are calculated in real time, providing high-frequency visual feedback for the synchronous following control of the mobile platform. A high-resolution vision sensor is also configured at the end of the serial robotic arm for local close-range imaging of the target screw holes, achieving sub-millimeter level precision positioning.

[0047] The screw feeding mechanism employs a spring-loaded hopper design, pre-loaded with target screws in sequence. When the tightening shaft moves to the preset pickup position, a pneumatic pusher pushes a single screw to the outlet. The sleeve at the end of the tightening shaft uses an internal electromagnet to generate magnetic force to attract and grip the screw. During the tightening operation, the servo drive system implements a dual closed-loop control strategy for torque and rotation angle on the tightening shaft. First, a constant torque drives the screw to screw in until the screw seat is in contact with the connected part. Then, the tightening shaft is controlled to rotate at a preset angle, and the target preload is accurately achieved by controlling the elastic elongation of the bolt.

[0048] To ensure continuous operation of the system, an autonomous charging module is configured. A pre-installed charging station within the workshop integrates a high-power wireless charging transmitter. After completing all tightening tasks, the mobile platform autonomously plans its path back to the charging station. Non-contact power replenishment is achieved through electromagnetic induction between the chassis receiving module and the transmitter, realizing fully automated energy management of the equipment. This minimizes manual intervention and downtime, supporting long-term, uninterrupted operation.

[0049] In some instances, dynamic information about the target vehicle is obtained, including: Acquire image sequences of the target vehicle using visual capture equipment; Based on the image sequence, determine the pixel displacement of preset feature points in the target vehicle; The dynamic information of the target vehicle is determined based on pixel displacement and the acquisition time interval of the image sequence.

[0050] For example, a vision capture device deployed above the work area continuously captures images of predetermined feature points on the chassis of a moving target vehicle at a preset acquisition frequency, obtaining an image sequence reflecting the continuous change of the vehicle's position over time. The vision capture device typically employs a high-frame-rate industrial camera, whose optical axis is pre-aligned with selected physical feature points on the vehicle chassis that possess high contrast and unique texture, such as specific assembly marks, hole edges, or weld points, ensuring that these feature points remain within range and are captured throughout the vehicle's movement. Each frame in the image sequence contains timestamp information, providing a time reference for analyzing the vehicle's motion state.

[0051] The acquired image sequence is processed to determine the pixel displacement of preset feature points. These preset feature points are typically selected as markers with high contrast or unique patterns, whose positions shift across different frames of the image sequence. Computer vision algorithms, such as feature point tracking or optical flow, are used to identify and match the same preset feature point across two or more consecutive frames. By calculating the change in pixel coordinates of this feature point within the image plane of the preceding and following frames, its displacement vector in the two-dimensional pixel coordinate system is obtained, i.e., the pixel displacement. This pixel displacement reflects the amount of projected motion of the vehicle relative to the visual capture device on the two-dimensional imaging plane within the camera sampling time interval. Analyzing the pixel displacements of multiple feature points improves the accuracy of motion state calculation.

[0052] Based on the determined pixel displacements and the known acquisition time intervals between image frames, the dynamic information of the target vehicle is determined. Specifically, the pixel displacements are converted into displacements in actual three-dimensional space according to the intrinsic and extrinsic parameters of the vision capture device, and then divided by the corresponding acquisition time interval to calculate the instantaneous velocity and direction of the target vehicle in three-dimensional space. This dynamic information is a vector containing the vehicle's linear velocity and direction angle at the current moment. Its update frequency is consistent with the frame rate of the vision capture device, enabling it to reflect the changes in the vehicle's motion state on the assembly line in real time and providing accurate input for the synchronous following control of the mobile platform.

[0053] In summary, this embodiment of the application achieves the perception of the motion state of a moving vehicle by monitoring preset feature points and acquiring image sequences through a visual capture device, providing a high-frequency feedback data source. Pixel displacement is calculated based on the image sequence, directly measuring the vehicle's microscopic motion with the precision of visual perception, avoiding the cumulative errors that may be introduced by relying on indirect calculations. Using a known acquisition time interval, the pixel displacement is converted into motion speed and direction; this dynamic information characterizes the vehicle's real-time pose change in the global coordinate system. This series of processes ensures the accuracy of the acquired dynamic information, providing a decision-making basis for the mobile platform to achieve precise synchronous following motion, quickly responding to changes in vehicle motion, and maintaining a stable relative working relationship.

[0054] In some instances, the initial position information of all target screw holes is obtained, including: Perform vehicle model identification on the target vehicle to determine its vehicle model identifier; Based on the vehicle model identifier, the corresponding chassis 3D model is determined from the pre-stored vehicle model database. The chassis 3D model contains the initial position coordinates of all target bolt holes. Based on the 3D model of the chassis, the initial position information of all target screw holes is generated.

[0055] For example, when a target vehicle enters a pre-defined work area, a vehicle model recognition process is initiated. This process uses high-resolution vision sensors deployed at the assembly line entrance or work start point to acquire images of the vehicle's overall chassis outline, specific structural features, or pre-defined markings. The acquired vehicle feature images are fed into a pre-trained vehicle model recognition model for analysis and matching. This model's database stores typical feature data for all possible vehicle models. By comparing and calculating the acquired features with those in the database, a unique vehicle model identifier is output, thereby accurately determining the specific model of the current target vehicle.

[0056] It should be noted that the pre-trained vehicle recognition model described above is a computer vision model built upon a deep convolutional neural network. In the offline phase, this model is trained using supervised learning on a large dataset of labeled vehicle chassis images covering all possible vehicle models. During training, the model learns to extract multi-level feature representations from the input images. These features effectively distinguish differences between different vehicle models in terms of overall chassis outline, key component layout, structural shape, or specific identification textures. In the online recognition phase, when the acquired vehicle feature images are input into the model, the forward propagation computation outputs a feature vector. This vector is then compared for similarity or classified against typical feature data of various vehicle models stored in the model's memory, mapping and outputting the corresponding vehicle identification code. The model's role is to transform raw pixel data into high-level semantic information, namely, the vehicle model category.

[0057] Based on the identified vehicle model identifier, the system accesses a pre-stored vehicle model database. This database contains high-precision 3D digital chassis models pre-built and stored for each supported vehicle model. These models were created during the vehicle design phase or through offline precision measurements. The models not only contain the structural geometry of the chassis but also annotate the theoretical center positions of all bolt holes to be tightened in 3D coordinates. Through a query operation, the system retrieves the 3D chassis model that perfectly corresponds to the current target vehicle based on the vehicle model identifier.

[0058] After successfully establishing the chassis 3D model, the 3D coordinate data of all the bolt holes to be tightened, marked in the model, are read. This coordinate data is typically based on a coordinate system associated with the chassis design datum. This coordinate data is then formatted and organized, for example, sorted and indexed according to the assembly sequence or region of the bolt holes, generating a structured dataset containing the theoretical initial positions of all target bolt holes—that is, the initial position information of all target bolt holes—providing a spatial reference for subsequent positioning and coordinate correction calculations.

[0059] In summary, this application employs a high-resolution vision sensor for non-contact image acquisition, providing raw data for vehicle model recognition. A pre-trained deep convolutional neural network model processes the images, leveraging its powerful feature learning capabilities to transform chassis visual information into high-level semantic features suitable for classification. This allows the vehicle model recognition process to adapt to different lighting conditions and viewing angles, improving robustness and accuracy. Based on the vehicle identification code output by the model, a pre-stored database is accessed to retrieve the corresponding 3D chassis model, ensuring that the initial bolt hole position reference data relied upon by subsequent operations originates from the design origin, and its accuracy and reliability are not limited by on-site measurement conditions. The initial bolt hole position information is extracted and structured from the 3D model, transforming the design data into a spatial index that can directly drive error compensation, providing a unified theoretical benchmark for calculating the target positions of all bolt holes on the vehicle.

[0060] It should be noted that in the vehicle assembly lines to which this method is applicable, especially in workstations requiring work on the vehicle's underside, such as chassis assembly and powertrain installation, the target vehicle is typically in the middle of the assembly process. Its exterior is often partially obscured by assembly fixtures, moving brackets, or safety barriers, making it difficult to obtain a complete image of the vehicle's exterior for model identification from conventional side or top views. However, the structural features of the vehicle chassis, such as the layout of longitudinal and transverse beams, the location of key positioning holes, and the shape of the battery pack or fuel tank mounting surfaces, are determined during the vehicle design phase and are already present in the early stages of the final assembly process. Therefore, in practical implementation, this method preferably achieves model identification by analyzing the structural feature information of the vehicle's underside area, thereby overcoming the identification obstacles in the aforementioned application scenarios.

[0061] In some instances, based on dynamic information, the mobile platform is controlled to perform synchronous following motion, maintaining a preset relative positional relationship between the mobile platform and the target vehicle, including: For example, the mobile platform receives acquired dynamic information reflecting the motion state and pose changes of the target vehicle. This dynamic information includes the vehicle's instantaneous velocity and direction of motion in the global coordinate system. Based on this dynamic information, the control system within the mobile platform calculates and generates motion control commands that match its own drive mechanism. The goal of these commands is to drive the mobile platform to adjust its own velocity and direction of motion so that its composite motion vector in the global coordinate system matches the motion vector of the target vehicle.

[0062] Specifically, the control process compares the deviation between the current pose of the mobile platform and the pose predicted based on dynamic information to maintain the preset relative positional relationship, and continuously adjusts the output of the drive motor using a closed-loop control algorithm (such as PID control) to correct the motion trajectory of the mobile platform. The preset relative positional relationship refers to the three-dimensional spatial offset and attitude angle that the work origin on the mobile platform (e.g., the center of the actuator base) should maintain relative to a fixed reference point on the target vehicle (e.g., the geometric center of the chassis), as defined in the work planning stage. Through the above adjustments, the mobile platform can achieve synchronous following with the target vehicle moving at a constant or variable speed on the assembly line, creating a work space reference that is approximately stationary relative to the moving vehicle chassis for the actuators fixedly mounted on the mobile platform.

[0063] In some instances, based on initial position information, the actuator mounted on the mobile platform is controlled to position a preset target screw hole, thereby obtaining the actual position of the preset target screw hole, including: Based on the initial position information and preset filtering rules, the preset target screw holes and their corresponding initial position coordinates are determined from all target screw holes; Based on the initial location coordinates and the preset observation range threshold, the preset observation area is determined; Move the vision sensor installed at the end of the actuator to the preset observation area; A local image of a preset observation area is acquired using a visual sensor; The local image is processed based on a preset image recognition algorithm to determine the actual image coordinates of the preset target screw hole in the image coordinate system; Based on the actual image coordinates and the preset calibration parameters of the vision sensor, the actual position of the preset target screw hole in the spatial coordinate system is determined.

[0064] For example, based on the initial position information of all target screw holes obtained from the chassis 3D model and a preset screening rule, one screw hole is selected from all screw holes to be tightened as a preset target screw hole, and the initial position coordinates of the screw hole in the chassis 3D model coordinate system are obtained. The preset screening rule is a pre-set logical criterion, which is determined based on factors such as the screw hole's process priority, the representativeness of its spatial location (e.g., located in the geometric center area of ​​the chassis or a mounting point with high structural rigidity), or the degree of correlation with vehicle positioning feature points. Its purpose is to select a key screw hole that can reflect the overall vehicle posture as a reference point for global positioning.

[0065] After determining the preset target screw hole and its initial position coordinates, a preset observation area is determined based on these initial position coordinates and a preset observation range threshold. This preset observation range threshold is a pre-defined three-dimensional spatial range value, defining the allowable search boundaries along the X, Y, and Z axes of the spatial coordinate system, centered on the initial position coordinates. The determination of the preset observation area aims to provide a limited search space for precise visual positioning, avoiding blind scanning in unreferenced chassis areas, thereby reducing the scope of image processing and improving positioning efficiency.

[0066] The control mechanism moves the vision sensor mounted at the end of the actuator to a calculated preset observation area. Based on the spatial coordinates of the preset observation area, the actuator drives its joint axes to move in a coordinated manner, ultimately accurately positioning the optical center of the vision sensor at the optimal observation point within that area. This point ensures that the theoretical position of the preset target screw hole is within the depth of field and clear imaging field of view of the vision sensor.

[0067] After the vision sensor is in place, image acquisition is performed to obtain a local image of the preset observation area. This local image is presented in the form of a digital pixel matrix, containing the texture features of the chassis surface within the observation area, the edge contours of the screw holes, and background interference information such as oil stains and reflections. During the acquisition process, the vision sensor adaptively adjusts the exposure parameters and focal length according to the ambient light to ensure that the image has sufficient contrast and sharpness.

[0068] After acquiring the local image, a preset image recognition algorithm is invoked to process the image data. This algorithm, once trained, is capable of recognizing specific visual features of the screw hole in the image. The algorithm performs digital image processing and feature analysis on the input local image, outputting the position of the preset target screw hole in the image plane coordinate system (i.e., pixel coordinate system). This position is expressed in the form of image coordinates, such as the row and column number of the pixel where the center point of the screw hole is located, which is the actual image coordinate of the preset target screw hole in the image coordinate system.

[0069] Based on the determined actual image coordinates and the preset calibration parameters obtained by the vision sensor through pre-calibration, the actual position of the preset target screw hole in the spatial coordinate system is determined. The preset calibration parameters of the vision sensor specifically include intrinsic parameters and extrinsic parameters: intrinsic parameters describe the inherent geometric characteristics of sensor imaging, such as focal length, principal point coordinates, and lens distortion coefficients. They establish a mathematical mapping relationship between the coordinates in pixels on the image plane and the normalized coordinates in physical dimensions in the sensor coordinate system; extrinsic parameters define the three-dimensional spatial position and rotational attitude of the optical center of the vision sensor in a certain reference coordinate system (usually the coordinate system of the end-effector of the actuator), i.e., the rotation matrix and translation vector.

[0070] By combining data representing the precise pose of the vision sensor in the mobile platform's base coordinate system, provided by the actuator's kinematic model, a complete coordinate transformation chain is constructed. This chain starts from the 2D image pixel coordinate system, first using intrinsic parameters to eliminate distortion and transform to the sensor coordinate system, then using extrinsic parameters to transform to the end-effector coordinate system, and finally using forward kinematics calculations based on the actuator's joint angle data to transform back to the mobile platform's base coordinate system or the global coordinate system. Through continuous and reversible matrix transformation operations, the three-dimensional coordinates of the preset target screw hole in the spatial coordinate system are ultimately calculated, i.e., its actual position, realizing the mapping from 2D image information to 3D spatial position.

[0071] In summary, this application's embodiments optimize the allocation of computing resources by selecting key screw holes as positioning references through preset screening rules; improve the efficiency of visual search by defining the search area through preset observation range thresholds; control the visual sensor to move to the area to acquire local images, providing a high-quality data source for feature recognition; extract target coordinates from the image using a preset image recognition algorithm, achieving non-contact precise positioning; and convert image coordinates into spatial coordinates using preset calibration parameters, ensuring the accuracy and usability of the positioning results in the global coordinate system.

[0072] In some instances, a local image is processed based on a preset image recognition algorithm to determine the actual image coordinates of the preset target screw hole in the image coordinate system, including: A preprocessed image is obtained by performing grayscale conversion and noise filtering on a local image. Edge detection is performed on the preprocessed image to obtain edge features; Based on the preset screw hole contour features, circle detection is performed on the edge features to determine the set of candidate circle center pixel coordinates for the preset target screw hole; From the set of candidate center pixel coordinates, select the candidate center pixel coordinates that are closest to the projected coordinates of the initial position coordinates in the image plane, and use them as the actual image coordinates of the preset target screw hole in the image coordinate system.

[0073] For example, a preprocessed image is obtained by performing grayscale conversion and noise filtering on a local image. Grayscale conversion maps a color or three-channel image to a single-channel grayscale image using algorithms such as weighted averaging or channel separation. Its purpose is to remove interference from color information on screw hole edge recognition and reduce the data dimensionality of subsequent processing, thereby improving computational efficiency. Noise filtering employs median filtering or Gaussian filtering algorithms, using a sliding window to statistically or weightedly average the neighborhood of image pixels, suppressing isolated noise points and smoothing fine texture interference, while preventing excessive blurring of the step edges of the screw hole contour. The preprocessed image output after grayscale conversion and noise filtering exhibits a significantly improved signal-to-noise ratio, and the grayscale contrast between the screw hole area and the background metal surface is clearer.

[0074] Edge detection is performed on the preprocessed image to obtain edge features. Edge detection employs gradient-based operators, with the Canny edge detection algorithm being preferred. This algorithm first performs Gaussian smoothing on the preprocessed image to further suppress residual noise. Then, it uses the Sobel operator to calculate the gradient magnitude and direction in the horizontal and vertical directions, generating a gradient magnitude image. Subsequently, non-maximum suppression is performed, preserving local maximum pixels along the gradient direction and suppressing non-maximum points, thinning the edge ridges to a single pixel width. Finally, a dual-threshold hysteresis processing is applied: a high threshold is used to identify strong edge pixels and initiate edge connections, while a low threshold is used to track and connect weak edge pixels along the edge direction. The final output is a binary image composed of continuous edge points, i.e., the edge features. These edge features record the locations of all drastic gray-level gradient changes in the local image in pixel coordinates, presenting various geometric contours of the chassis surface, including screw hole edges, weld seams, scratches, and oil stain boundaries.

[0075] Based on preset screw hole contour features, circular detection is performed on edge features to determine the candidate set of center pixel coordinates for the target screw hole. The preset screw hole contour features refer to a pre-defined set of parameters describing the projected geometry of the target screw hole in the image, mainly including the empirical range of the screw hole imaging radius, the circularity threshold, and edge closure requirements. Circular detection employs the Hough circle detection algorithm, which treats each edge pixel in the edge features as a point on the candidate circumference and accumulates the center coordinates and radius in the three-dimensional parameter space based on its gradient direction. The algorithm, constrained by the radius range in the preset screw hole contour features, only votes on radius parameters within this range, effectively reducing the parameter search space and suppressing interference from irrelevant contours. After traversing all edge pixels, the parameter combinations corresponding to local peak points in the parameter space are identified as candidate circles, and their center coordinates and radii are recorded and output, forming the candidate set of center pixel coordinates. This set includes the center of the true target screw hole and may also include false centers formed by other approximately circular structures on the chassis or edge noise.

[0076] From the set of candidate center pixel coordinates, the candidate center pixel coordinates that are closest to the projected coordinates of the initial position coordinates in the image plane are selected as the actual image coordinates of the preset target screw hole in the image coordinate system. First, using the pre-calibrated intrinsic parameter matrix and distortion coefficients of the vision sensor, the initial position coordinates of the preset target screw hole in the spatial coordinate system are mapped to the image plane through perspective projection transformation, obtaining the projected coordinates of the theoretical position of the screw hole in the pixel coordinate system. These projected coordinates represent the position where the preset target screw hole should appear in the image under ideal conditions without any assembly deviations. Then, the Euclidean distance between each candidate center pixel coordinate in the candidate center pixel coordinate set and this projected coordinate is calculated one by one, and the candidate center coordinate with the smallest distance value is selected as the final matching result. This selection strategy is based on the engineering prior that the actual screw hole position will not deviate significantly from the theoretical position within the normal production tolerance range of the vehicle, effectively eliminating false detection results far from the target area and ensuring the uniqueness and accuracy of the positioning output. The output actual image coordinates are the precise center position of the preset target screw hole in the current local image.

[0077] In summary, this embodiment transforms the original high-dimensional, noisy image data into a low-dimensional, high signal-to-noise ratio preprocessed image by performing grayscale conversion and noise filtering on local images, reducing interference and computational overhead in subsequent processing. Edge detection is performed on the preprocessed image, converting pixel-level grayscale information into geometrically meaningful binary edge features, completing the abstraction from the original image to a structured contour, allowing the screw hole edge and background interference to be expressed in the same feature space. Based on preset screw hole contour features, circular detection is performed on the edge features. The parameter search space is limited to a reasonable range using radius range and circularity constraints, and a voting mechanism is used to determine a finite number of candidate circle centers from a massive number of discrete edges. Finally, Euclidean distance is used as a benchmark to filter the candidate results, and geometric verification is performed to uniquely determine the center position of the actual screw hole in the image.

[0078] In some embodiments, the target positions of all target screw holes are determined based on the actual and initial position information of the preset target screw holes, including: Based on the actual position and initial position information of the preset target screw hole, determine the spatial transformation parameters; Based on the initial position coordinates of all target screw holes in the spatial transformation parameters and initial position information, coordinate correction calculations are performed on the initial position coordinates of all target screw holes to determine the target positions of all target screw holes, including: Construct a coordinate transformation matrix based on spatial transformation parameters; Based on the coordinate transformation matrix, the initial position coordinates of each target screw hole in the initial position information are transformed to determine the corrected position coordinates of each target screw hole. Based on the corrected position coordinates of all target screw holes, determine the target positions of all target screw holes.

[0079] For example, after obtaining the actual position of the preset target screw hole, spatial transformation parameters are calculated. The actual position of the preset target screw hole is obtained by the vision sensor locating the screw hole on the current vehicle chassis in a synchronous following state on the moving platform, resulting in three-dimensional spatial coordinates. These coordinates directly reflect the true position of the screw hole in the current vehicle's actual pose. Simultaneously, the preset target screw hole corresponds to a pre-stored initial position coordinate in the initial position information. This coordinate originates from the chassis 3D model of the corresponding vehicle model and represents the theoretical position of the screw hole under ideal design conditions. These two coordinates describe the same physical point expressed in two different coordinate systems: the initial position coordinate is located in the theoretical model coordinate system, while the actual position is located in the current vehicle's actual pose coordinate system. Due to factors such as manufacturing tolerances, jig sway, or tooling positioning deviations on the assembly line, the vehicle's actual pose will experience rigid body displacement between the theoretical model and the actual pose. Therefore, there is a rotational and translational transformation relationship between these two coordinate systems. This method solves for the rigid body transformation, treating the actual position and initial position coordinates as a pair of corresponding points. Using singular value decomposition or least squares optimization, it calculates the rotation matrix and translation vector from the theoretical model coordinate system to the current vehicle's actual pose coordinate system. These two parameters are then used to determine the spatial transformation parameters. These parameters describe the overall assembly deviation and spatial pose offset of the vehicle at the current moment.

[0080] After obtaining the spatial transformation parameters containing the rotation matrix and translation vector, a homogeneous coordinate transformation matrix is ​​constructed based on these parameters. The coordinate transformation matrix adopts a 4×4 homogeneous coordinate form, encapsulating the rotation matrix and translation vector within the same mathematical operator. Its upper left 3×3 sub-block is the rotation matrix obtained earlier, which describes the rotation relationship between the theoretical model coordinate system and the current vehicle's actual pose coordinate system in three-dimensional space, specifically expressed as continuous rotation angles around the X, Y, and Z axes. The first three rows of the fourth column correspond to the three components of the translation vector, representing the translation offset along the X, Y, and Z axes, respectively. The fourth row is fixed at [0 0 0 1]. This homogeneous transformation matrix expresses the rigid body transformation mapping relationship from the theoretical model coordinate system to the current vehicle's actual pose coordinate system, enabling any spatial point defined in the theoretical model coordinate system to be transformed to its corresponding point in the current vehicle's actual pose coordinate system through a single matrix multiplication operation. The construction of this matrix integrates discrete rotation and translation parameters into a single mathematical expression, providing a mathematical model for coordinate correction of the initial position coordinates of all target screw holes.

[0081] Based on the constructed coordinate transformation matrix, this method performs coordinate correction calculations on the initial position coordinates of all target screw holes. The initial position information includes the initial position coordinates of all target screw holes involved in the current task in the theoretical model coordinate system, with each coordinate stored as a three-dimensional vector. For each target screw hole, the method first expands the initial position coordinates of each target screw hole into a homogeneous coordinate vector, that is, adds the value 1 to the end of its three-dimensional coordinates to form a four-dimensional column vector; multiplies this four-dimensional vector on the left by the constructed homogeneous coordinate transformation matrix, performing matrix multiplication to obtain the transformed four-dimensional homogeneous coordinate vector; extracts the first three components of this vector, which are the corrected position coordinates of the screw hole in the current vehicle's actual pose coordinate system after rigid body transformation. This coordinate transformation operation essentially applies the same rotation and translation to each screw hole point in the theoretical model coordinate system as the preset target screw hole, thereby synchronously mapping the theoretical positions of all screw holes to spatial positions that match the current vehicle's actual pose. This operation fully preserves the relative geometric topological relationships between all target bolt holes, ensuring that the corrected coordinate point set still follows the design structure of the vehicle chassis and will not undergo non-rigid deformations such as distortion, scaling, or shearing due to coordinate transformation. This process achieves systematic compensation from the measured deviation of a single reference point to the positional deviation of bolt holes throughout the entire vehicle.

[0082] After transforming the initial position coordinates of all target screw holes, the corrected position coordinates for each screw hole are organized to form a target position data set that is structurally consistent with the initial position information but numerically aligned with the actual vehicle pose. The target position of each target screw hole in this set is the spatial coordinate point to which the actuator needs to move and perform screw insertion and tightening operations. This target position information fully describes the position of all screw holes to be tightened on the chassis under the current dynamic following state of the vehicle, serving as the spatial command reference for actuator movement, screw hole alignment, pose fine-tuning, and screw tightening operations.

[0083] In summary, this embodiment of the application quantifies the positional deviation obtained from single-point visual measurement into a rigid body pose deviation description of the entire vehicle by solving the spatial transformation parameters between the actual position and initial position coordinates of the preset target screw hole. This achieves an upgrade from local perception information to a global error model, ensuring that the result of a single positioning is no longer limited to a single screw hole. A homogeneous coordinate transformation matrix is ​​constructed based on the spatial transformation parameters, unifying rotation and translation into matrix operations. This avoids repetitive calculations of the transformation relationship for each screw hole individually, improving computational efficiency. By performing the same coordinate transformation matrix multiplication on the initial position coordinates of each target screw hole, the pose deviation observed at the preset target screw hole is transferred to all screw holes in the chassis, ensuring the physical consistency of error compensation. That is, all screw holes in the vehicle are treated as a rigid whole, shifting synchronously with the actual vehicle pose, rather than being corrected independently. The final set of target positions for all target screw holes not only matches the actual position of the current vehicle, but also retains the original relative geometric relationship between each screw hole. Thus, when the actuator moves to each target position in sequence, there is no need to perform a global search or secondary positioning for each screw. The precise alignment of the tightening shaft and the screw hole can be achieved solely by relying on the target position, which improves the assembly efficiency and positioning reliability in multi-screw hole collaborative operation scenarios.

[0084] In some instances, the actuator includes a parallel platform and a serial robotic arm mounted on the parallel platform, and after determining the target positions of all target screw holes, it also includes: For each target screw hole among all target screw holes, based on the target position of the target screw hole, determine the first pose adjustment parameter of the parallel platform and the second pose adjustment parameter of the serial robotic arm; Based on the first pose adjustment parameters, control the parallel platform to perform the first degree of freedom motion; Based on the second pose adjustment parameters, the serial robotic arm is controlled to perform a second degree of freedom movement so that the tightening shaft installed at the end of the serial robotic arm is aligned with the target screw hole.

[0085] For example, the actuator consists of a six-DOF parallel platform and a three-DOF serial manipulator mounted on the parallel platform. The parallel platform has a large stroke and high rigidity spatial pose adjustment capability, while the serial manipulator has the characteristics of flexible end effector and fast response. The two are serially arranged in spatial structure, jointly providing motion compensation capability for nine independent degrees of freedom. After determining the target positions of all target screw holes, for each target screw hole, the target position coordinates of the screw hole in the current mobile platform base coordinate system are first obtained. Since this target position is the spatial point that the tightening shaft end effector ultimately needs to reach and align with the central axis of the screw hole, and the tightening shaft is fixed to the end effector of the serial manipulator, the target position needs to be mapped to the multi-DOF motion space of the actuator. According to the kinematic model of the parallel platform and the kinematic model of the serial manipulator, the target position is decomposed into two levels of pose adjustment target: the parallel platform needs to adjust the base of the serial manipulator to a macroscopic reference pose that can cover the workspace near the target position; the serial manipulator, based on this, uses its end effector movement to accurately align the tightening shaft with the target screw hole.

[0086] Based on the target position of the target screw hole, the first pose adjustment parameters of the parallel platform are determined. The upper platform of the parallel platform is fixed to the base of the serial robotic arm, and the lower platform is fixed to the base of the moving platform. Taking the current pose of the upper platform of the parallel platform as the initial state, the length changes of the six legs or the six-dimensional pose offset of the upper platform in the spatial coordinate system required for the upper platform to move from the current pose to a certain target pose are calculated through inverse kinematics. This set of parameters is determined as the first pose adjustment parameters. The principle for selecting this target pose is to ensure that during the subsequent movement of the serial robotic arm, the angles of all its joints are within the workspace, while the end effector can reach the target position with high positioning accuracy and fast response speed.

[0087] Based on the determined first posture adjustment parameters, the parallel platform is subjected to first-degree-of-freedom motion control. The parallel platform adopts a six-degree-of-freedom Stewart parallel mechanism, and its motion control is achieved by adjusting the extension and retraction of the electric cylinders of the six outriggers. The target extension and retraction of the six electric cylinders in the first posture adjustment parameters are converted into speed and angle commands for each servo motor and sent to the servo drivers of each outrigger. The servo drivers drive the electric cylinders to extend and retract precisely according to the commands. By collecting feedback signals from the built-in displacement sensors of each cylinder in real time, a position closed loop is formed to ensure that the six outriggers move synchronously and in a coordinated manner, so that the upper platform of the parallel platform can quickly and smoothly reach the desired macroscopically compensated posture. This step is mainly used to compensate for the spatial position and posture deviations caused by the overall assembly tolerance of the vehicle, the sway of the assembly line lifting fixture, and large vibrations, and to coarsely position the working space of the actuator to the vicinity of the target screw hole, providing a proximal reference for the fine adjustment of the serial robotic arm.

[0088] After the parallel platform completes the motion defined by the first pose adjustment parameter, based on the actual pose of the current serial manipulator base and the target position of the target screw hole, the second pose adjustment parameter of the serial manipulator is determined through inverse kinematics solution of the serial manipulator. The serial manipulator consists of multiple rotary joints, and there is a definite geometric mapping relationship between its end-effector pose and joint variables. Using the current joint angles as initial values ​​and the desired end-effector pose corresponding to the target position as input, the combination of joint angles that meets the accuracy requirements is solved using analytical or numerical iterative methods. This set of joint angle values ​​or the motion increment relative to the current angle is determined as the second pose adjustment parameter. The serial manipulator is a three-degree-of-freedom precision serial mechanism, and each joint is driven by a high-resolution servo motor and a harmonic reducer. The target angle values ​​of the three joints in the second pose adjustment parameter are sent to the servo drivers of each joint. The drivers form a position closed loop based on the angle signal fed back by the encoder in real time, driving the joint motors to rotate precisely to the commanded angle. During the movement, the control system acquires the current pose of the tightening shaft at the end of the robotic arm in real time and iteratively compares it with the target position of the target screw hole until the central axis of the tightening shaft coincides with the central axis of the target screw hole in three-dimensional space, with an alignment accuracy reaching the sub-millimeter level. This step, based on the coarse positioning of the parallel platform, compensates for minor positional deviations introduced by factors such as local vehicle deformation, screw hole machining tolerances, and visual positioning residuals, ultimately achieving high-precision alignment between the tightening shaft and the target screw hole.

[0089] It should be noted that the movements of the parallel platform and the serial robotic arm are not completely decoupled in sequence throughout the process. They are coordinated based on real-time feedback. That is, the serial robotic arm can start the pre-positioning movement before the parallel platform has fully converged, so as to achieve composite error compensation and high-speed response of nine degrees of freedom.

[0090] In summary, this embodiment of the application constructs a nine-degree-of-freedom composite mechanism by connecting a six-degree-of-freedom parallel platform and a three-degree-of-freedom serial robotic arm, achieving a hierarchical division of error compensation capabilities. The parallel platform executes a large-stroke, high-rigidity first-degree-of-freedom motion based on the first posture adjustment parameter, offsetting macroscopic posture deviations introduced by overall vehicle assembly tolerances, lifting fixture sway, and assembly line vibration, quickly and coarsely positioning the tightening shaft to the workspace near the target screw hole. The serial robotic arm executes a high-precision, fast-response second-degree-of-freedom motion based on the second posture adjustment parameter, further compensating for sub-millimeter errors such as local deformation, screw hole machining tolerances, and visual positioning residuals on the near-end reference provided by the parallel platform, ensuring that the central axis of the tightening shaft coincides with the central axis of the screw hole. This collaborative control strategy not only expands the actuator's adaptability to changes in vehicle posture but also ensures the accuracy of end-effector positioning, shortens the single-hole alignment time, and provides a stable and reliable spatial reference for subsequent screw insertion and high-preload tightening, improving the efficiency of multi-screw hole collaborative operations while ensuring consistent assembly quality.

[0091] In some instances, the actuator is controlled to move to the target position corresponding to all target screw holes to tighten all target screws, including: For each target screw hole among all target screw holes, control the actuator to move to the target position of that target screw hole; When the actuator moves to the target position of the target screw hole, the screw feeding mechanism will deliver the target screw corresponding to the target screw hole to the preset pickup position; When the target screw reaches the preset pickup position, the tightening shaft at the end of the control actuator picks up the target screw and screws the target screw into the target screw hole. During the screwing process of the target screw, the output torque and rotation angle of the tightening shaft are controlled by a dual closed-loop servo system based on a preset tightening strategy, so that the target screw can reach the target preload.

[0092] For example, after the actuator completes the calculation of the target positions for all target screw holes, it enters the batch tightening execution stage. For each target screw hole, the actuator first drives its multi-degree-of-freedom motion system to move the end-effector tightening shaft to a preset starting point above the target position, based on the target position of that screw hole. This movement process is driven by the actuator's motion controller, which generates continuous trajectory commands based on forward kinematics and path planning algorithms, driving the parallel platform and the serial robotic arm to move in coordination, ensuring that the tightening shaft arrives at the target position along a smooth, collision-free path. Upon arrival, the central axis of the tightening shaft and the central axis of the target screw hole have been spatially aligned through the preceding error compensation and alignment operations, establishing a geometric reference for screw pickup and insertion.

[0093] Once the actuator moves to the target position of the current target screw hole and completes its attitude adjustment, the control system sends a screw supply trigger signal to the screw supply mechanism. The screw supply mechanism employs a cartridge-type hopper structure, pre-loading multiple target screws according to their specifications and tightening sequence. Upon receiving the trigger signal, the screw supply mechanism drives its built-in pneumatic pusher or stepper motor to push the foremost target screw at the hopper outlet to the preset pickup position. The preset pickup position is a calibrated fixed point precisely matched to the spatial coordinates of the sleeve at the end of the tightening shaft in the pickup posture. Simultaneously, the screw supply mechanism monitors the target screw's position in real time using a position sensor or photoelectric sensor. When the target screw is detected to be stably positioned at the preset pickup position, the screw supply mechanism sends a ready signal to the control system, completing the screw supply action.

[0094] Upon receiving the screw-supply ready signal, the control system immediately applies a pickup command to the electromagnetic adsorption sleeve at the end of the tightening shaft. The electromagnet coil inside the sleeve is energized, generating a strong magnetic field. This magnetic field is concentrated in the central area through the magnetically conductive structure on the sleeve's end face, thus firmly adsorbing the target screw head, located at the preset pickup position, onto the sleeve's end face. During pickup, the tightening shaft makes a slight advance along its axis to ensure the screw head is fully embedded in the sleeve and coaxially aligned with the sleeve's central axis. Simultaneously, a positioning detection switch inside the sleeve confirms in real time that the screw is fully inserted and against the inner wall of the sleeve. The actuator then lifts and carries the target screw out of the screw-supply area. After removal, the actuator repositions the tightening shaft above the target screw hole according to its current position and smoothly advances the shaft along the screw hole's axis until the screw tip contacts the beginning of the screw hole's thread. The control system then drives the tightening shaft to rotate forward at a set speed, so that the target screw can be screwed into the screw hole by itself through the helical motion. During the screwing process, the feed speed and rotation speed of the tightening shaft maintain a predetermined kinematic matching relationship to ensure that the threads are properly engaged and that there is no stripping or jamming, thereby completing the screw picking and screwing action.

[0095] Once the screw is screwed in to a certain depth, the control system activates a preset tightening strategy to perform dual closed-loop servo control on the tightening shaft output. This control strategy simultaneously employs two independent feedback loops. The torque control loop uses a torque sensor installed inside the tightening shaft as the feedback element to collect the current output torque value in real time and compare it with the target contact torque threshold specified in the preset tightening strategy. A PID controller adjusts the servo motor's current output based on the deviation, causing the actual torque to approach and stabilize at the target contact torque threshold. The angle control loop uses a photoelectric encoder at the rear of the servo motor as the feedback element to continuously record the absolute angle of rotation of the tightening shaft from the moment of contact. When the actual torque first reaches the target contact torque threshold, the control system immediately switches the control mode or superimposes it into the angle control stage. In this stage, the tightening shaft continues to rotate at a set low speed with a preset angle increment until the preset angle threshold is reached. Through the coordinated adjustment of the torque and angle dual loops, the axial tension of the target screw is controlled within the design range, thereby obtaining the target preload and ultimately completing the tightening operation on the target screw hole. After the tightening of the target screw hole is completed, the actuator moves to the target position of the next target screw hole, and repeats the above process until all target screw holes have been tightened.

[0096] In summary, this embodiment of the application controls the actuator to move to the target position of the target screw hole, transforming the spatial coordinate command established by error compensation calculation into the physical movement of the tightening shaft, ensuring that the starting point of screw insertion is consistent with the actual position of the screw hole. By controlling the screw feeding mechanism to transport the target screw to the preset pickup position, automatic screw supply is achieved, eliminating the manual loading step and shortening the operation time for a single screw. By controlling the tightening shaft to electromagnetically pick up the screw at the preset pickup position, fast and reliable gripping is achieved, avoiding mechanical clamping alignment deviations and improving the pickup success rate and operational continuity. During the screwing process, dual closed-loop servo control of the output torque and rotation angle based on a preset tightening strategy transforms the tightening process from a single torque limitation to preload regulation, effectively compensating for interference factors such as differences in thread friction coefficients, ensuring that each screw receives a stable and consistent target preload, and improving the reliability of assembly quality.

[0097] In some instances, after tightening all target screws, the following steps are also included: Control the mobile platform to move to the preset charging station and charge the mobile platform.

[0098] For example, after the mobile platform completes the screw tightening operation on all target screw holes and the control system confirms that the current task has been completed, the mobile platform receives a return command from the central control system at the end of the task. This command is automatically generated based on the task completion status or a preset work cycle.

[0099] The mobile platform first reads the preset charging station location data pre-stored in the control system. This preset charging station is a dedicated workstation deployed in a designated area of ​​the final assembly workshop and integrated with a high-power wireless charging transmitter. Based on its onboard LiDAR, the mobile platform scans the environmental point cloud in real time and matches it with the pre-stored environmental map using real-time positioning and map building algorithms to determine its current position in the global coordinate system. Then, it calls the path planning algorithm to generate a collision-free optimal navigation path, starting from the current position and targeting the preset charging station location, while avoiding static obstacles and dynamic mobile devices.

[0100] The mobile platform drives its independently suspended drive wheels according to the path, and the left and right wheel speeds are differentially adjusted by a servo controller to achieve trajectory tracking. When the mobile platform enters the parking area of ​​a preset charging station, it switches to a low-speed alignment mode. Using infrared positioning sensors and magnetic navigation sensors installed on the platform chassis, it performs relative pose detection with the corresponding reference marks on the charging station. By adjusting, the wireless charging receiving coil of the mobile platform chassis is aligned with the wireless charging transmitting coil in the charging station in the vertical projection direction and kept within a preset distance threshold.

[0101] Once the mobile platform is in position, the charging station control system detects the load connection and immediately initiates wireless power transmission. The transmitter inverter rectifies and filters the mains frequency AC power, converting it into high-frequency AC power, which is applied to the transmitting coil. According to Faraday's law of electromagnetic induction, the receiving coil induces an AC electromotive force with the same frequency as the transmitter in the alternating magnetic field. This electromotive force is then converted into DC power by the receiving end's rectifier and filter circuit, and the battery management system provides constant current and constant voltage charging for the power battery pack mounted on the mobile platform.

[0102] Throughout the charging process, the battery management system and the charging station controller exchange data such as battery state of charge, temperature, and health via a wireless communication module, adjust the charging power, and cut off energy transmission once the battery capacity reaches a preset threshold, completing the charging operation. The mobile platform remains in standby mode during charging, waiting for a trigger signal from the next target vehicle entering the work station to achieve uninterrupted cyclical operation.

[0103] This embodiment eliminates the waiting and operation time of manual scheduling or manual plugging and unplugging of charging guns by controlling the mobile platform to move to the preset charging station after all tightening operations are completed. This allows the mobile platform to autonomously complete energy replenishment without relying on shift breaks, improving the time utilization of the equipment. Controlling the mobile platform relies on LiDAR and pre-stored maps to autonomously plan the path and navigate to the preset charging station. This process ensures that the mobile platform can reliably navigate to the designated charging station in complex and dynamic workshop environments without the need for physical guide rails or magnetic strips, enhancing adaptability to different workshop layouts.

[0104] By deploying a high-power wireless charging transmitter module fixed at a charging station and a receiver module integrated into a mobile platform chassis, non-contact power transmission is achieved using the principle of electromagnetic induction. This avoids the physical wear, poor contact, and electric arc safety risks that may be caused by repeated plugging and unplugging of the charging interface, and enhances the electrical safety and ease of operation during the charging process.

[0105] Continuous battery status monitoring and dynamic adjustment of charging parameters during charging ensure the power battery always operates within its optimal charge / discharge range, extending battery cycle life and reducing battery replacement frequency and maintenance costs. Two-way real-time communication between the battery management system and the charging controller during charging enables dynamic adaptation of charging power and full lifecycle battery health management. This effectively slows battery capacity degradation while rapidly replenishing charge, reducing long-term equipment maintenance costs. This autonomous charging mechanism eliminates the need for the mobile platform to return to a fixed standby area or rely on manual battery swapping, allowing for continuous 24 / 7 response to production schedules and providing energy security for automated tightening of vehicle chassis bolts.

[0106] Please see Figure 2 The diagram below illustrates the structure of an automated vehicle screw tightening device according to an embodiment of this application, comprising: The screw hole data acquisition unit 21 is used to acquire the dynamic information of the target vehicle and the initial position information of all target screw holes; The platform position determination unit 22 is used to control the mobile platform to perform synchronous following motion based on dynamic information, so that the mobile platform and the target vehicle maintain a preset relative position relationship. The screw hole position correction unit 23 is used to control the actuator installed on the mobile platform to position the preset target screw hole based on the initial position information, so as to obtain the actual position of the preset target screw hole. The preset target screw hole is determined from all target screw holes based on preset screening rules. The screw hole position determination unit 24 is used to determine the target position of all target screw holes based on the actual position and initial position information of the preset target screw holes; The screw tightening operation unit 25 is used to control the actuator to move to the target position corresponding to all target screw holes and tighten all target screws.

[0107] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

[0108] Although preferred embodiments have been described in this specification, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications that fall outside the scope of this specification.

[0109] Obviously, those skilled in the art can make various modifications to this specification without departing from its spirit and scope. Therefore, this specification is intended to include any modifications that fall within the scope of the claims and their equivalents.

Claims

1. An automated method for tightening vehicle screws, characterized in that, include: Obtain the dynamic information of the target vehicle and the initial position information of all target bolt holes; Based on the dynamic information, the mobile platform is controlled to perform synchronous following motion, so that the mobile platform and the target vehicle maintain a preset relative positional relationship. Based on the initial position information, the actuator installed on the mobile platform is controlled to position the preset target screw hole to obtain the actual position of the preset target screw hole, wherein the preset target screw hole is determined from all target screw holes based on a preset screening rule; Based on the actual position of the preset target screw hole and the initial position information, determine the target position of all target screw holes; The actuator is controlled to move to the target position corresponding to all target screw holes, and all target screws are tightened.

2. The method according to claim 1, characterized in that, The acquisition of the target vehicle's dynamic information includes: Image sequences of the target vehicle are acquired using a visual capture device; Based on the image sequence, determine the pixel displacement of preset feature points in the target vehicle; The dynamic information of the target vehicle is determined based on the pixel displacement and the acquisition time interval of the image sequence.

3. The method according to claim 1, characterized in that, The process of obtaining the initial position information of all target screw holes includes: The target vehicle is identified by its vehicle model to determine its vehicle model identifier. Based on the vehicle model identifier, the corresponding chassis 3D model is determined from the pre-stored vehicle model database, wherein the chassis 3D model contains the initial position coordinates of all target bolt holes; Based on the chassis 3D model, the initial position information of all target screw holes is generated.

4. The method according to claim 1, characterized in that, The step of controlling the actuator mounted on the mobile platform to position the preset target screw hole based on the initial position information, and obtaining the actual position of the preset target screw hole, includes: Based on the initial position information and the preset filtering rules, a preset target screw hole and its corresponding initial position coordinates are determined from all target screw holes; Based on the initial position coordinates and the preset observation range threshold, the preset observation area is determined; Move the vision sensor installed at the end of the actuator to the preset observation area; The visual sensor is used to acquire a local image of the preset observation area; The local image is processed based on a preset image recognition algorithm to determine the actual image coordinates of the preset target screw hole in the image coordinate system; Based on the actual image coordinates and the preset calibration parameters of the vision sensor, the actual position of the preset target screw hole in the spatial coordinate system is determined.

5. The method according to claim 4, characterized in that, The step of processing the local image based on a preset image recognition algorithm to determine the actual image coordinates of the preset target screw hole in the image coordinate system includes: The local image is subjected to grayscale conversion and noise filtering to obtain a preprocessed image; Edge detection is performed on the preprocessed image to obtain edge features; Based on the preset screw hole contour features, the edge features are subjected to circular detection to determine the candidate circle center pixel coordinate set of the preset target screw hole; From the set of candidate center pixel coordinates, the candidate center pixel coordinates that are closest to the projected coordinates of the initial position coordinates in the image plane are selected as the actual image coordinates of the preset target screw hole in the image coordinate system.

6. The method according to claim 1, characterized in that, The step of determining the target position of all target screw holes based on the actual position of the preset target screw hole and the initial position information includes: Based on the actual position of the preset target screw hole and the initial position coordinates corresponding to the preset target screw hole in the initial position information, the spatial transformation parameters are determined; Based on the spatial transformation parameters and the initial position coordinates of all target screw holes in the initial position information, coordinate correction calculations are performed on the initial position coordinates of all target screw holes to determine the target positions of all target screw holes.

7. The method according to claim 6, characterized in that, The step of performing coordinate correction calculations on the initial position coordinates of all target screw holes based on the spatial transformation parameters and the initial position coordinates corresponding to all target screw holes in the initial position information, and determining the target positions of all target screw holes, includes: Based on the spatial transformation parameters, construct the coordinate transformation matrix; Based on the coordinate transformation matrix, a coordinate transformation operation is performed on the initial position coordinates of each target screw hole in the initial position information to determine the corrected position coordinates of each target screw hole; Based on the corrected position coordinates of all target screw holes, the target positions of all target screw holes are determined.

8. The method according to claim 7, characterized in that, The actuator includes a parallel platform and a series robotic arm mounted on the parallel platform, and after determining the target positions of all target screw holes, it further includes: For each target screw hole among all target screw holes, based on the target position of the target screw hole, determine the first pose adjustment parameter of the parallel platform and the second pose adjustment parameter of the serial robotic arm; Based on the first pose adjustment parameters, the parallel platform is controlled to perform a first degree of freedom motion; Based on the second pose adjustment parameters, the serial robotic arm is controlled to perform a second degree of freedom movement so that the tightening shaft installed at the end of the serial robotic arm is aligned with the target screw hole.

9. The method according to claim 1, characterized in that, The step of controlling the actuator to move to the target position corresponding to all target screw holes and tightening all target screws includes: For each of the target screw holes, the actuator is controlled to move to the target position of that target screw hole; When the actuator moves to the target position of the target screw hole, the screw feeding mechanism is controlled to deliver the target screw corresponding to the target screw hole to the preset pickup position; When the target screw reaches the preset pickup position, the tightening shaft at the end of the actuator is controlled to pick up the target screw and screw the target screw into the target screw hole; During the screwing process of the target screw, the output torque and rotation angle of the tightening shaft are controlled by a dual closed-loop servo system based on a preset tightening strategy, so that the target screw reaches the target preload.

10. The method according to claim 1, characterized in that, After tightening all target screws, the process also includes: Control the mobile platform to move to a preset charging station and charge the mobile platform.

Images