Single-dish vacuum transfer system and its cycle operation control method

Abstract

This invention discloses a single-disc vacuum transfer system and its cyclic operation control method, belonging to the field of vacuum transfer control technology. The method acquires an initial vacuum degree sequence and an initial steel strip tension value sequence to generate a motion parameter constraint matrix; it extracts motion trajectory reference points through equidistant sampling and fits a wafer transfer reference path; it collects the height and horizontal deflection angle values ​​of the wafer support stage for end-effector attitude compensation to obtain the target placement approach path; it collects the angular displacement sequence and angular velocity sequence of the steel strip active wheel to calculate the placement deviation vector; and it adjusts the constraint boundary values ​​for the next cycle based on the change in the deviation vector between the current and historical placement deviation vectors. This scheme reduces the impact of vacuum fluctuations and steel strip tension changes on the transfer motion, improves the end-effector attitude adaptability, and achieves cyclic adaptive correction of the constraint boundaries.

Description

Technical Field

[0001] This invention relates to the field of vacuum transmission control technology, specifically to a single-disc vacuum transmission system and its cyclic operation control method. Background Technology

[0002] In semiconductor wafer manufacturing, single-disc vacuum transfer systems transfer wafers from the pick-up position to the placement position within a vacuum chamber. The height and horizontal angle of the wafer carrier at the placement position are affected by factors such as installation status and chamber temperature variations. Existing cyclic operation control methods typically employ offline teaching or preset trajectories, controlling the robot's movement under fixed speed and acceleration limits. Vacuum fluctuations and steel belt tension changes within the vacuum chamber are ignored or treated merely as alarm parameters. Vacuum variations alter the damping of the kinematic pairs and the friction characteristics of the seals, while steel belt tension changes affect the transmission of driving force from the drive wheel to the end effector. Fixed motion parameter constraints make it difficult to ensure that the instantaneous speed and acceleration at each trajectory point remain within the allowable range of the current cycle; local over-limits can cause servo saturation, steel belt slippage, or end effector jitter. During placement, the pre-planned path end-effector posture cannot be corrected based on the actual height and horizontal angle of the wafer carrier, and placement position and posture deviations are easily introduced into the wafer-to-carrier contact process.

[0003] Existing methods for handling placement deviations primarily focus on the current detection or alarm process. Some solutions only compensate for the positional error in the current cycle, failing to feed back the variation in deviation between adjacent cycles to the motion constraint boundaries of the next cycle. This leads to the recurrence of similar deviations in continuous cyclic operations. To address these issues, it is necessary to solve the problem of generating motion parameter constraints that match the current cycle under dynamic changes in vacuum level and steel strip tension, and to fit a wafer transport reference path accordingly. Furthermore, it is necessary to solve the problem of adaptively adjusting the motion constraint boundaries using the variation in placement deviations between adjacent cycles. Summary of the Invention

[0004] This invention provides a single-disc vacuum transfer system, which includes a robotic arm and a sealed robotic arm chamber. An automatically opening and closing door is located above the robotic arm chamber. A valve controlling the opening and closing of the chamber is located on one side of the robotic arm chamber. The front end of the robotic arm can pass laterally through the valve. The rear end of the robotic arm is fixed to a steel belt. The two ends of the steel belt are respectively connected to the outer edges of a driving pulley and a driven pulley. The steel belt is fixedly connected to the output belt of a synchronous belt drive mechanism via a first connecting block. The input end of the synchronous belt drive mechanism is connected to a drive mechanism, which drives the belt to rotate.

[0005] A single-disc vacuum transfer system and its cyclic operation control method are applied to the cyclic operation process of the robot in the aforementioned single-disc vacuum transfer system. The purpose is to generate a motion parameter constraint matrix based on the initial vacuum degree sequence and the initial steel strip tension value sequence of the current cycle, fit the wafer transfer reference path under dynamic constraint conditions, and reduce the adverse effects of vacuum environment fluctuations and steel strip tension changes on path execution. The boundary values ​​of the motion parameter constraint matrix for the next cycle are adjusted by the deviation change between the placement deviation vector of the current cycle and the previous cycle, so that the motion constraints in continuous cyclic operation are adaptively corrected according to the actual placement deviation, thereby improving placement consistency and cyclic operation stability.

[0006] To achieve the above objectives, the present invention provides the following technical solution: The present invention provides a cyclic operation control method for a single-disc vacuum transfer system, comprising: responding to an initialization completion signal, acquiring an initial vacuum degree sequence and an initial steel strip tension value sequence for the current cycle; generating a motion parameter constraint matrix for the current cycle based on the initial vacuum degree sequence and the initial steel strip tension value sequence; extracting multiple motion trajectory reference points in each dimension of the motion parameter constraint matrix using an equidistant sampling method; fitting a wafer transfer reference path for the current cycle based on the multiple motion trajectory reference points; and acquiring the height value of the wafer support stage and the horizontal deflection angle value at the placement position within the current cycle. The method involves performing end-point attitude compensation on the wafer transport reference path based on the height and horizontal deflection of the wafer stage to obtain the target placement approach path. The robot arm is then controlled to execute the placement action for the current cycle based on this target placement approach path. The angular displacement and angular velocity sequences of the steel belt drive pulley are collected during the placement action, and the placement deviation vector for the current cycle is calculated based on these sequences. The historical placement deviation vector from the previous cycle is obtained, and the constraint boundary values ​​of the motion parameter constraint matrix for the next cycle are adjusted based on the deviation change between the current cycle's placement deviation vector and the historical placement deviation vector. This method incorporates vacuum level and steel belt tension into the motion parameter constraints and performs cross-cycle iterative correction based on actual placement errors, ensuring consistency in the wafer placement process under vacuum and steel belt drive conditions.

[0007] As a technical solution of the present invention, the process of generating the motion parameter constraint matrix for the current cycle includes: after the robot arm completes the return-to-origin operation and all axes are ready, receiving the initialization completion signal, triggering the vacuum sensor group to continuously collect multiple instantaneous vacuum values ​​in the robot arm cavity at a preset sampling frequency to form the initial vacuum degree sequence; simultaneously triggering the tension sensor group to continuously collect multiple instantaneous tension values ​​on the driving and driven sides of the steel belt to form the initial steel belt tension value sequence; extracting the maximum and minimum vacuum values ​​from the initial vacuum degree sequence, and extracting the maximum and minimum tension values ​​from the initial steel belt tension value sequence to form a motion parameter boundary array; and constructing a motion parameter constraint matrix based on the motion parameter boundary array, with the upper limit of the steel belt linear velocity and the upper limit of acceleration as variables, wherein the motion parameter constraint matrix includes the acceleration constraint range and velocity constraint range for each motion axis. Thus, the motion parameter constraint matrix can reflect the current vacuum environment and steel belt tension state, ensuring that the path planning parameters are within the acceptable motion range of the equipment.

[0008] Preferably, the row vector dimension of the motion parameter constraint matrix corresponds to the lifting axis, rotation axis and telescopic axis in the joint space of the robot hand, and the column vector dimension corresponds to the velocity constraint range of each axis in the acceleration and deceleration phases. This can directly provide constraints for different joint motion phases and reduce the risk of parameter out-of-bounds errors during the path search process.

[0009] As a technical solution of the present invention, the process of fitting and obtaining the wafer transport reference path includes: dividing the motion space of the wafer from the pick-up position to the placement position into multiple interpolation segments; extracting multiple path points at equal time intervals within the start and end times of each interpolation segment to obtain multiple motion trajectory reference points for each interpolation segment; reading the velocity and acceleration constraint intervals corresponding to the current interpolation segment in the motion parameter constraint matrix; determining whether the instantaneous velocity and instantaneous acceleration of each motion trajectory reference point fall within the velocity and acceleration constraint intervals; if so, retaining the motion trajectory reference point; otherwise, discarding the motion trajectory reference point; and fitting all retained motion trajectory reference points using a cubic spline interpolation method to obtain a continuous and smooth wafer transport reference path, which is represented in a time-parameterized form. This path reduces motion abrupt changes while satisfying vacuum and steel strip tension constraints, which helps to reduce the risk of vibration and wafer displacement during transport.

[0010] Preferably, during the start and end times of each interpolation segment, the equal time interval is adaptively determined based on the response bandwidth of the steel belt drive wheel, and the number of motion trajectory reference points extracted in each interpolation segment is proportional to the total number of interpolation segments in the current cycle.

[0011] As a technical solution of the present invention, the process of obtaining the target placement approach path includes: when the end effector of the robot reaches a preset deceleration zone near the placement position, triggering a laser displacement sensor mounted on the substrate of the wafer support stage to measure the height difference between the upper surface of the wafer support stage and the reference plane of the robot end effector, as the height value of the wafer support stage; simultaneously triggering an angle sensor mounted on the side of the wafer support stage to measure the angle between the central axis of the wafer support stage and the feed direction of the robot end effector, as the horizontal deflection angle value; calculating the height compensation amount of the end effector in the vertical direction based on the height value of the wafer support stage, and calculating the rotation compensation amount of the end effector in the horizontal plane based on the horizontal deflection angle value; superimposing the height compensation amount and the rotation compensation amount onto the end pose parameters of the wafer transfer reference path to generate the target placement approach path. Through end effector attitude compensation, the influence of the wafer support stage height and horizontal angle deviation on the placement position can be corrected, improving the placement position accuracy.

[0012] Preferably, the measurement reference for the height of the plate bearing stage is the lower surface plane of the robot end effector when it is in a horizontal posture, and the measurement reference for the horizontal deflection angle is the theoretical feed direction vector of the end effector in the horizontal plane.

[0013] As a technical solution of the present invention, the process of calculating the placement deviation vector includes: according to the time parameter sequence of the target placement path, outputting position and speed commands for each interpolation point to the servo driver sequentially, causing the steel belt drive wheel to rotate according to the commands, driving the end effector to move along the target placement path; during the placement action, continuously collecting the number of rotary encoder pulses on the steel belt drive wheel at a fixed sampling period, converting the pulse count at each sampling moment into instantaneous angular displacement values ​​to form the angular displacement sequence, and simultaneously collecting the instantaneous rotational speed of the steel belt drive wheel to form the angular velocity sequence; obtaining the placement position deviation based on the difference between the angular displacement value at the end of the angular displacement sequence and the expected angular displacement value corresponding to the theoretical placement position; obtaining the speed tracking deviation based on the difference between the angular velocity value at each sampling moment and the commanded speed value at the corresponding moment in the angular velocity sequence, and combining the placement position deviation and the speed tracking deviation into the placement deviation vector. This quantifies the placement execution error of the current cycle, providing a basis for adjusting the constraint boundary in the next cycle.

[0014] As a technical solution of the present invention, the process of adjusting the constraint boundary value of the next cycle includes: reading the historical placement deviation vector recorded at the end of the previous cycle from the system storage area; calculating the difference vector between the placement deviation vector of the current cycle and the historical placement deviation vector, as the deviation change; extracting the position deviation change component and the velocity deviation change component from the deviation change; if the position deviation change component is greater than zero, subtracting a first adjustment step size from the velocity constraint upper limit value of the current motion parameter constraint matrix; if the position deviation change component is less than zero, increasing the velocity constraint upper limit value of the current motion parameter constraint matrix; if the velocity deviation change component is greater than zero, subtracting a second adjustment step size from the acceleration constraint upper limit value of the current motion parameter constraint matrix; if the velocity deviation change component is less than zero, increasing the acceleration constraint upper limit value of the current motion parameter constraint matrix, thereby obtaining the constraint boundary value of the motion parameter constraint matrix for the next cycle. This adjustment process dynamically changes the constraint boundary according to the deviation change trend between adjacent cycles, making the motion parameters of the next cycle closer to the current equipment state, thereby gradually reducing the placement deviation.

[0015] The technical effects and advantages provided by the present invention in the above technical solution are as follows: Based on the maximum and minimum vacuum values ​​in the initial vacuum degree sequence and the maximum and minimum tension values ​​in the initial steel strip tension value sequence, a boundary array of motion parameters is constructed. A motion parameter constraint matrix is ​​generated, with the upper limits of the steel strip linear velocity and acceleration as variables. The row vectors of the constraint matrix correspond to the acceleration constraint intervals of each motion axis, and the column vectors correspond to the velocity constraint intervals of each motion axis. Multiple motion trajectory reference points are extracted in each dimension of the motion parameter constraint matrix using equidistant sampling. The instantaneous velocity and instantaneous acceleration of each reference point are compared with the corresponding velocity and acceleration constraint intervals. Reference points exceeding the constraint intervals are discarded, and the remaining reference points are fitted into a time-parameterized wafer transport reference path using cubic spline interpolation. Unlike conventional path planning methods that use fixed speed and acceleration limits, the actual measured values ​​of vacuum level and steel belt tension within the current cycle are introduced into the constraint boundary generation process. This causes the motion parameter constraints to change with the vacuum environment and tension state, preventing some trajectory points from exceeding speed or acceleration limits due to the current cycle boundary offset. This reduces servo saturation, steel belt slippage, and end-effector jitter caused by these issues. The reference points after constraint screening are then interpolated using cubic splines, resulting in a smooth and continuous path that is more suitable for the cyclic transmission requirements of single-disc vacuum transmission systems under varying operating conditions.

[0016] During the placement action, the angular displacement and angular velocity sequences of the steel belt drive pulley are collected. The placement position deviation is obtained from the difference between the angular displacement at the termination moment and the expected angular displacement corresponding to the theoretical placement position. The speed tracking deviation is obtained from the difference between the angular velocity at each sampling moment and the commanded speed, forming the placement deviation vector for the current cycle. The historical placement deviation vector recorded in the previous cycle is read, and the deviation change between the current deviation vector and the historical deviation vector is calculated. The position deviation change component and the speed deviation change component are extracted from the deviation change. When the position deviation change component is greater than zero, the first adjustment step is subtracted from the upper limit of the speed constraint in the current motion parameter constraint matrix to tighten the speed constraint boundary for the next cycle; when the position deviation change component is less than zero, the first adjustment step is added to the upper limit of the speed constraint to moderately loosen the speed constraint boundary for the next cycle. The speed deviation change component acts on the upper limit of the acceleration constraint in the same way. Unlike conventional methods that only detect or compensate for the deviation in the current placement, the deviation change between adjacent cycles is used as feedback to adjust the motion constraint boundary of the next cycle. The constraint boundary is adjusted by adding or subtracting according to the actual placement deviation change direction, which can suppress the accumulation of repetitive placement deviation in continuous cycle operation, reduce the probability of the same position deviation and speed tracking deviation in subsequent cycles, and improve the consistency of placement in multi-cycle operation. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0018] Figure 1 This is a flowchart of the cyclic operation control method for a single-disc vacuum transfer system; Figure 2 This is a flowchart of the wafer transfer reference path generation method; Figure 3 This is a flowchart for obtaining the placement deviation vector; Figure 4 This is a schematic diagram of the reference points and speed constraints for the telescopic shaft motion trajectory of a single-disc vacuum transmission system; Figure 5 The curves show the root mean square values ​​of the placement deviation and speed tracking deviation as a function of the cycle period. Figure 6 This is a three-dimensional structural diagram of a single-disc vacuum transfer system; Figure 7 This is a partial cross-sectional view of a single-disc vacuum transfer system; Among them, 101, robotic arm; 1, wafer support; 2, robotic arm chamber; 3, door; 4, valve; 5, steel belt; 6, steel belt drive pulley; 7, steel belt driven pulley; 8, drive mechanism; 9, drive pulley; 10, first driven pulley; 11, second driven pulley; 12, first synchronous belt; 13, second synchronous belt; 14, wafer. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] See Figure 6 and Figure 7 This invention provides a single-disc vacuum transfer system, comprising a robotic arm 101 arranged laterally, with a wafer support 1 at its front end for placing a wafer 14. The single-disc vacuum transfer system also includes a sealed robotic arm chamber 2, with an automatically opening and closing door 3 above the chamber. A valve 4 controlling the opening and closing of the chamber is located on one side of the robotic arm chamber 2, allowing the front end of the robotic arm to pass laterally through the valve 4. The robotic arm chamber 2 and a wafer processing chamber are located on either side of the valve 4. When the valve 4 is open, the robotic arm chamber 2 is connected to the processing chamber; when the valve 4 is closed, the robotic arm chamber 2 is sealed and separated from the processing chamber.

[0021] The rear end of the robotic arm is fixed to the steel belt 5 via a second connecting block. Both ends of the steel belt 5 are connected to the outer edges of the driving pulley 6 and the driven pulley 7, respectively. The steel belt 5 is fixedly connected to the output belt of the synchronous belt drive mechanism via a first connecting block. The input end of the synchronous belt drive mechanism is connected to the drive mechanism 8, and is driven to rotate by the drive mechanism 8. Specifically, the synchronous belt drive mechanism includes a driving pulley 9, a first driven pulley 10, and a second driven pulley 11. The driving pulley 9 is connected to the first driven pulley 10 via a first synchronous belt 12, and the first driven pulley 10 is connected to the second driven pulley 11 via a second synchronous belt 13. The driving pulley 9 is fixedly sleeved on the output shaft of the drive mechanism 8. The second synchronous belt 13 is fixedly connected to the steel belt 5 via the first connecting block.

[0022] The drive mechanism 8 drives the active pulley 9 to rotate, which in turn drives the second synchronous belt 13 to move via the first synchronous belt 12 and the first driven pulley 10. During the movement of the second synchronous belt 13, the steel belt 5 moves via the first connecting block, which in turn moves the robotic arm 101, enabling the robotic arm 101 to reciprocate and transport the wafer 14. The drive mechanism 8 is motor-driven and is used to drive the active pulley 9 to rotate. The specific configuration and settings can be adapted by those skilled in the art according to the specific application requirements; this part is a conventional technique in the field and will not be elaborated here.

[0023] To prevent the wafer transport process from affecting the vacuum level of the process chamber, after the robotic arm 101 retracts into the robotic arm chamber 2, valve 4 is closed, and then door 3 is opened. At this time, the robotic arm chamber 2 is no longer in a vacuum state as it is connected to the outside. The wafer support 1 is located directly below door 3, and the wafer 14 falls from the opened door 3 onto the wafer support 1. Then, door 3 is closed, and the automatic vacuum mode is activated. The vacuum pumping device connected to the bottom surface of the robotic arm chamber 2 performs vacuuming treatment inside the robotic arm chamber 2. The system first detects the presence of the wafer using a laser sensor located below the wafer support 1. Next, the vacuum levels of the robotic arm chamber 2 and the process chamber are detected and compared using vacuum sensors to confirm whether the vacuum conditions set for robotic arm transfer are met. Once the transfer conditions are met, the control valve 4 opens, and the drive mechanism 8 drives the active pulley 9 to rotate, causing the robotic arm 101 to extend from the robotic arm chamber 2 and reach the designated position in the process chamber. After the wafer 14 is supported by the wafer support stage in the process chamber, the robotic arm 101 detaches from the bottom surface of the wafer 14, and the robotic arm 101 retracts back into the robotic arm chamber 2 for the next wafer retrieval, thus achieving cyclical operation of the robotic arm 101. In this embodiment, the final distance traveled by the robotic arm 101 is twice the distance traveled by the second synchronous belt 13.

[0024] In this embodiment, in order to ensure the horizontal movement of the robot, support and guide mechanisms are also provided on both sides of the second synchronous belt 13. The support and guide mechanisms include guide rails and sliders. The guide rails are installed on the outer shell of the second synchronous belt 13, and the sliders are fixed to the first connecting block as one piece through the connecting plate. The guide rails and sliders cooperate to ensure that the first connecting block can only slide smoothly along the direction of the guide rails.

[0025] The aforementioned steel belt structure enables the "linear motion" of the end effector, rather than the circular motion common in traditional manipulators. Unlike atmospheric manipulators that rely on simple transmission methods such as belts and lead screws, the steel belt structure used in a vacuum utilizes a high-strength, corrosion-resistant stainless steel precision belt for transmission via "fixed traction." This transmission method is fundamentally different from common belt or chain drives. The two ends of the steel belt are not connected to form a loop, but are directly connected to the outer edges of the driving pulley 6 and the driven pulley 7. Because there is no relative sliding between the steel belt and the pulleys, this structure provides extremely high transmission accuracy, ensuring precise manipulator movement. The core component responsible for power transmission, the steel belt 5, is made of stainless steel, which is corrosion-resistant, has low gas release, and is dust-free. Due to its metallic properties, it generates almost no dust during vacuum friction.

[0026] See Figure 1 This invention also provides a cyclic operation control method for a single-disc vacuum transfer system, mainly applied to the reciprocating motion control process of a robot in the aforementioned single-disc vacuum transfer system. The method includes: responding to an initialization completion signal, acquiring the initial vacuum degree sequence and the initial steel strip tension value sequence for the current cycle; generating a motion parameter constraint matrix for the current cycle based on the initial vacuum degree sequence and the initial steel strip tension value sequence; extracting multiple motion trajectory reference points in each dimension of the motion parameter constraint matrix using equidistant sampling; fitting the multiple motion trajectory reference points to obtain the wafer transfer reference path for the current cycle; acquiring the height and horizontal deflection angle of the wafer support platform at the placement position within the current cycle; performing end-effector attitude compensation on the wafer transfer reference path based on the height and horizontal deflection angle to obtain the target placement approach path; controlling the robot to execute the placement action for the current cycle based on the target placement approach path; acquiring the angular displacement sequence and angular velocity sequence of the steel strip drive wheel during the placement action; and calculating the placement deviation vector for the current cycle based on the angular displacement sequence and angular velocity sequence. Obtain the historical placement deviation vector of the previous cycle, and adjust the constraint boundary values ​​of the motion parameter constraint matrix for the next cycle based on the change in deviation between the current cycle's placement deviation vector and the historical placement deviation vector.

[0027] In practical implementation, the single-disc vacuum transfer system's cyclic operation control method receives an initialization completion signal after the robot arm completes its return-to-origin operation and all axes are ready. This initialization completion signal is issued by the robot arm controller after the lifting, rotating, and telescopic axes have completed their return-to-origin operations; it is a rising edge signal. In response to the initialization completion signal, the vacuum sensor group is triggered to continuously collect multiple instantaneous vacuum levels within the robot arm's chamber at a preset sampling frequency. The vacuum sensor group includes three vacuum gauges, which are respectively installed on the top, bottom, and side walls of the robot arm's chamber. The preset sampling frequency is set to 100Hz, the sampling period is 10ms, and the sampling duration is 2s. Each vacuum gauge collects 200 instantaneous vacuum levels, and the 600 instantaneous vacuum levels collected by the three gauges are arranged in chronological order to form an initial vacuum level sequence.

[0028] Simultaneously, the tension sensor group continuously acquires multiple instantaneous tension values ​​from both the driving and driven sides of the steel belt. The tension sensor group includes a tension sensor on the driving side and a tension sensor on the driven side. The driving side tension sensor is mounted on the driving pulley bearing housing, and the driven side tension sensor is also mounted on the driven pulley bearing housing. The sampling frequency of the tension sensor group is set to 100Hz, the sampling period is 10ms, and the sampling duration is 2s. Each tension sensor acquires 200 instantaneous tension values. These 200 values ​​are then arranged in chronological order to form an initial steel belt tension value sequence.

[0029] In practice, the maximum and minimum vacuum values ​​are recorded by comparing each of the 600 instantaneous vacuum values ​​in the initial vacuum degree sequence. Similarly, the maximum and minimum tension values ​​are recorded by comparing each of the 400 instantaneous tension values ​​in the initial steel strip tension value sequence. The maximum and minimum vacuum values, along with the maximum and minimum tension values, are then arranged in the order of maximum vacuum, minimum vacuum, maximum tension, and minimum tension to form a boundary array of motion parameters.

[0030] In practical implementation, a motion parameter constraint matrix is ​​constructed based on the boundary array of motion parameters. The row vector dimension of the motion parameter constraint matrix corresponds to the lifting, rotation, and extension axes in the joint space of the robot hand, and the row vectors correspond to the acceleration constraint intervals of each motion axis; the column vector dimension of the motion parameter constraint matrix corresponds to the velocity constraint intervals of each axis during the acceleration and deceleration phases. When constructing the motion parameter constraint matrix, the upper limit of the acceleration constraint is calculated according to the following formula: in, Indicates the first Upper limit of acceleration constraint value for each motion axis The values ​​can be set for lifting axis, rotation axis, and telescopic axis. The unit of the upper limit of the acceleration constraint is meters per second squared or radians per second squared, depending on the type of motion axis. This represents the tension boundary weight, with a value ranging from 0 to 1. In this embodiment... The value is set to 0.6, based on the fact that the impact of steel strip tension fluctuation on acceleration constraint is greater than that of vacuum degree fluctuation, and that steel strip tension fluctuation accounts for 60% of the weight allocation between the steel strip preload adjustment range and the chamber vacuum degree control accuracy. This represents the maximum tension value in the initial steel strip tension value sequence, expressed in Newtons. This represents the minimum tension value in the initial steel strip tension value sequence, expressed in Newtons. This represents the maximum vacuum level in the initial vacuum level sequence, expressed in Pascals. This represents the minimum vacuum level in the initial vacuum level sequence, expressed in Pascals. Indicates the first The rated acceleration values ​​for each motion axis are as follows: the rated acceleration value for the lifting axis is 1.2 m / s², the rated acceleration value for the rotating axis is 2.0 rad / s², and the rated acceleration value for the telescopic axis is 1.5 m / s².

[0031] The lower limit of the acceleration constraint is set to 20% of the upper limit of the acceleration constraint. The upper limit of the velocity constraint is obtained by multiplying the upper limit of the allowable linear velocity of the steel strip by the ratio of the minimum tension to the maximum tension. The upper limit of the allowable linear velocity of the steel strip is stored in the system parameters; in this embodiment, the upper limit of the allowable linear velocity of the steel strip is 0.8 m / s. The lower limit of the velocity constraint is set to 10% of the upper limit of the velocity constraint. The motion parameter constraint matrix contains four velocity constraint columns: the first velocity constraint column corresponds to the lower limit of the velocity constraint during the acceleration phase, the second velocity constraint column corresponds to the upper limit of the velocity constraint during the acceleration phase, the third velocity constraint column corresponds to the lower limit of the velocity constraint during the deceleration phase, and the fourth velocity constraint column corresponds to the upper limit of the velocity constraint during the deceleration phase. The motion parameter constraint matrix contains three motion axis rows: the first motion axis row corresponds to the lifting axis, the second motion axis row corresponds to the rotation axis, and the third motion axis row corresponds to the telescopic axis. In each motion axis row, the first, second, third, and fourth speed constraint columns are filled with the corresponding acceleration phase speed constraint lower limit, acceleration phase speed constraint upper limit, deceleration phase speed constraint lower limit, and deceleration phase speed constraint upper limit values, respectively. Each motion axis row also includes additional storage locations for acceleration constraint lower limit and acceleration constraint upper limit values, which are filled with the corresponding motion axis's acceleration constraint lower limit and acceleration constraint upper limit values.

[0032] In specific implementation, please refer to Figure 2The movement space of the wafer from the pick-up position to the placement position is divided into multiple interpolation segments. The starting point of the movement space is the pick-up position, and the ending point is the placement position. The total travel distance of the movement space is determined by the effective travel distance of the robot's telescopic axis. In this embodiment, the total travel distance of the movement space is the spatial distance that the robot's telescopic axis moves from the coordinate value of the pick-up position to the coordinate value of the placement position. The total travel distance of the movement space is divided into 20 interpolation segments, and the total number of interpolation segments is set to 20. The duration of each of the 20 interpolation segments is equal, and the duration of each interpolation segment is obtained by dividing the total duration of the movement time axis by the total number of interpolation segments. The total duration of the movement time axis is determined based on the total travel distance of the movement space and the upper limit of the telescopic axis movement speed constraint. Specifically, the calculation method is to divide the total travel distance of the movement space by the upper limit of the telescopic axis movement speed constraint. The start and end times of each interpolation segment are determined according to the interpolation segment number. The start time of the first interpolation segment is the start time of the motion, and the end time of the first interpolation segment is the start time of the motion plus the duration of one interpolation segment. The start time of the second interpolation segment is the end time of the first interpolation segment, and the end time of the second interpolation segment is the start time of the second interpolation segment plus the duration of one interpolation segment. The start and end times of subsequent interpolation segments are determined according to the same time recursion method. The end time of the 20th interpolation segment is the end time of the motion.

[0033] In practical implementation, multiple path points are extracted at equal time intervals within the start and end times of each interpolation segment to obtain multiple motion trajectory reference points for each interpolation segment. The equal time intervals are adaptively determined based on the response bandwidth of the steel belt drive pulley. The response bandwidth of the steel belt drive pulley is stored in the system parameters; in this embodiment, the response bandwidth of the steel belt drive pulley is 20Hz. The equidistant sampling time interval is calculated according to the following formula: in, Indicates the time interval for equidistant sampling, in seconds; This represents the sampling density coefficient, which is dimensionless and ranges from 2 to 5. In this embodiment... The value is set to 4, based on the Nyquist sampling theorem that the sampling frequency should be greater than twice the highest frequency of the signal. The value is set to 4 times the response bandwidth to retain sufficient sampling margin. The response bandwidth of the steel belt drive pulley is expressed in Hertz (Hz). In this embodiment... The value is set to 20Hz. The values ​​are 4 and Substituting a value of 20Hz into the above formula, the calculated equidistant sampling time interval is 0.0125 seconds.

[0034] In practice, the number of motion trajectory reference points extracted within each interpolation segment is calculated by dividing the interpolation segment duration by the equidistant sampling time interval. In this embodiment, the total duration of the motion time axis is 2.5 seconds, the total number of interpolation segments is 20, the duration of each interpolation segment is 0.125 seconds, and the equidistant sampling time interval is 0.0125 seconds. Therefore, 10 motion trajectory reference points are extracted within each interpolation segment. The number of motion trajectory reference points extracted within each interpolation segment is proportional to the total number of interpolation segments in the current cycle, with the proportionality coefficient being the number of motion trajectory reference points within each interpolation segment. Within the first interpolation segment, starting from the start time of the first interpolation segment, a path point is extracted every 0.0125 seconds until the end time of the first interpolation segment, obtaining 10 motion trajectory reference points for the first interpolation segment; within the second interpolation segment, starting from the start time of the second interpolation segment, a path point is extracted every 0.0125 seconds until the end time of the second interpolation segment, obtaining 10 motion trajectory reference points for the second interpolation segment; and so on, obtaining motion trajectory reference points for all 20 interpolation segments.

[0035] In practice, the velocity and acceleration constraint intervals corresponding to the current interpolation segment are read from the motion parameter constraint matrix. The row vector dimension of the motion parameter constraint matrix corresponds to the lifting, rotation, and telescopic axes in the robot's joint space, while the column vector dimension corresponds to the velocity constraint intervals for each axis during acceleration and deceleration. The motion parameter constraint matrix also contains the storage locations for acceleration constraint intervals. The current interpolation segment corresponds to the movement of the robot's telescopic axis; therefore, the velocity and acceleration constraint intervals corresponding to the telescopic axis row in the motion parameter constraint matrix are read. The velocity constraint interval includes the lower and upper limits of the acceleration and deceleration velocity constraints. The acceleration constraint interval includes the lower and upper limits of the acceleration constraints.

[0036] In practice, it is determined whether the instantaneous velocity and instantaneous acceleration of each motion trajectory reference point fall within the velocity and acceleration constraint intervals. The instantaneous velocity of the motion trajectory reference point is obtained by differentiating the tangent direction of the wafer transmission path at the corresponding moment, and the instantaneous acceleration is obtained by differentiating the instantaneous velocity of the motion trajectory reference point with respect to time at the corresponding moment. If the instantaneous velocity of the motion trajectory reference point is between the lower limit and the upper limit of the velocity constraint during the acceleration phase, or if the instantaneous velocity of the motion trajectory reference point is between the lower limit and the upper limit of the velocity constraint during the deceleration phase, and the instantaneous acceleration of the motion trajectory reference point is between the lower limit and the upper limit of the acceleration constraint, then the motion trajectory reference point is retained; if the instantaneous velocity of the motion trajectory reference point is not within the velocity and acceleration constraint intervals, or if the instantaneous acceleration of the motion trajectory reference point is not within the acceleration constraint intervals, then the motion trajectory reference point is discarded. Perform the above judgment operation on all motion trajectory reference points in all 20 interpolation segments one by one, retain the motion trajectory reference points that meet the constraints, and remove the motion trajectory reference points that do not meet the constraints.

[0037] In practical implementation, cubic spline interpolation is used to fit all retained motion trajectory reference points. The cubic spline interpolation method constructs a cubic polynomial between every two adjacent retained motion trajectory reference points. The cubic polynomial is a function of position with respect to time, and its coefficients are solved using position continuity conditions, first derivative continuity conditions, and second derivative continuity conditions at adjacent motion trajectory reference points. A first boundary condition is set at the starting point of the path's motion trajectory reference point, and a second boundary condition is set at the ending point. Both the first and second boundary conditions have a second derivative of zero. By solving the cubic spline interpolation linear equation system, the coefficients of each segment of the cubic polynomial are obtained, resulting in a continuous and smooth wafer transport reference path. The wafer transport reference path is represented in a time-parameterized form, where each position vector on the wafer transport reference path is expressed as a function of time parameters. The robot controller reads the position coordinate values ​​at corresponding times on the wafer transport reference path according to the time parameters.

[0038] See Figure 4 In the figure, the horizontal axis represents time in seconds, ranging from 0 to 2.5 seconds; the left vertical axis represents the linear displacement of the telescopic shaft in millimeters, ranging from 0 to 200 millimeters; and the right vertical axis represents the velocity of the telescopic shaft in meters per second, ranging from approximately 18 meters per second to 160 meters per second. The blue dots in the figure represent motion trajectory reference points, the red solid line represents the wafer transfer reference path fitted through these reference points, the green dashed line represents the upper limit of the velocity constraint, and the purple dotted line represents the lower limit of the velocity constraint.

[0039] As shown in the figure, the displacement of the telescopic axis of the wafer transport reference path smoothly increases over time, starting from 0 mm and eventually reaching approximately 200 mm, demonstrating the orderly movement trajectory of the robot's telescopic axis from the pick-up position to the placement position. The upper limit of the speed constraint remains constant at approximately 160 meters per second, corresponding to the upper limit of the allowable linear velocity of the steel strip set in the system parameters of this embodiment. The lower limit of the speed constraint remains close to zero, reflecting the minimum constraint boundary of the speed.

[0040] During the period from 0 to approximately 0.5 seconds, the slope of the telescopic shaft displacement curve gradually increases, indicating that the speed of the robot's telescopic shaft gradually increases during the acceleration phase. A certain gap is maintained between the displacement curve and the upper limit of the speed constraint to ensure a safety margin in motion. From approximately 0.5 to 1.8 seconds, the speed curve approaches the upper limit of the speed constraint, and the telescopic shaft displacement curve shows a linear increase, indicating that the robot's telescopic shaft moves at a constant speed close to the maximum allowable speed during this phase. From 1.8 to 2.5 seconds, the slope of the telescopic shaft displacement curve gradually decreases, indicating that the robot is entering the deceleration phase. The speed curve gradually decreases away from the upper limit, ensuring that the end effector smoothly reaches the placement position.

[0041] Furthermore, the blue scatter points of the motion trajectory reference points highly overlap with the fitting curve of the red wafer transmission reference path, indicating that the motion trajectory reference points extracted by the equidistant sampling time interval and the cubic spline interpolation fitting method effectively realize the continuous and smooth motion trajectory of the robot, meet the requirements of the upper and lower limits of speed constraints, and avoid overspeed or underspeed phenomena.

[0042] In practice, when the robotic arm's end effector reaches the preset deceleration zone near the placement position, it triggers a laser displacement sensor mounted on the wafer stage base. The preset deceleration zone consists of a path segment on the wafer transport reference path from 0.15m away from the placement position to the placement position itself. The 0.15m distance is set based on the deceleration distance required for the robotic arm's telescopic axis to decelerate from the upper speed constraint limit to the placement speed. The laser displacement sensor is mounted on the side of the wafer stage base, with its optical axis pointing vertically upwards. The sensor emits a laser beam to the upper surface of the wafer stage and receives the reflected laser beam. The height difference between the upper surface of the wafer stage and the reference plane of the robotic arm's end effector is calculated using the laser's time-of-flight. The measurement reference for the wafer stage height is the lower surface plane of the robotic arm's end effector when it is in a horizontal position. The laser displacement sensor's sampling frequency is set to 500Hz, continuously collecting 20 height difference values. The average of these 20 height differences is taken as the wafer stage height value.

[0043] In practice, the angle sensor installed on the side of the substrate receiving stage is simultaneously triggered. The angle sensor is an absolute angle encoder, mounted on the central axis of the substrate receiving stage. It measures the angle between the central axis of the stage and the feed direction of the robot's end effector. The horizontal deflection angle is measured based on the theoretical feed direction vector of the end effector in the horizontal plane. The angle sensor sampling frequency is set to 500Hz, continuously collecting 20 angle values, and the average of these 20 values ​​is taken as the horizontal deflection angle value.

[0044] In practical implementation, the vertical height compensation of the end effector is calculated based on the height of the receiving platform, and the horizontal rotation compensation of the end effector is calculated based on the horizontal deflection angle. The height compensation and rotation compensation are calculated using the following formulas: in, This indicates the vertical height compensation of the end effector, in millimeters, with the positive direction being vertically upward. This indicates the rotational compensation of the end effector in the horizontal plane, expressed in radians, with the positive direction being counterclockwise. This represents the high-compensation gain coefficient, which is dimensionless and ranges from 0.8 to 1.2. In this embodiment... The value is set to 1.0, based on the linear correspondence between the height of the support stage measured by the laser displacement sensor and the actual height difference. The height compensation gain coefficient is set to 1.0 to make the height compensation amount equal to the height of the support stage. This represents the rotational compensation gain coefficient, which is dimensionless and ranges from 0.8 to 1.2. In this embodiment... The value is set to 1.0, based on the linear correspondence between the horizontal deflection angle measured by the angle sensor and the actual angle deviation. The rotation compensation gain coefficient is set to 1.0 to make the rotation compensation amount equal to the horizontal deflection angle value. This indicates the height of the support platform, in millimeters. This indicates the horizontal deflection angle, expressed in radians.

[0045] In practice, height compensation is superimposed on the vertical coordinate component of the end-effector pose parameters of the wafer transport reference path, and rotation compensation is superimposed on the horizontal rotation angle component of the end-effector pose parameters of the wafer transport reference path to generate the target placement approach path. The end-effector pose parameters of the wafer transport reference path include the X-axis, Y-axis, and Z-axis position coordinates of the robot end effector at the placement position, as well as the X-axis, Y-axis, and Z-axis attitude angles. The target placement approach path is the wafer transport reference path after end-effector pose compensation.

[0046] In specific implementation, please refer to Figure 3According to the time parameter sequence of the target part approach path, the robot controller sequentially outputs position and speed commands for each interpolation point to the servo driver. The time parameter sequence of the target part approach path is generated from the wafer transfer reference path after end-effector attitude compensation. Each time point in the time parameter sequence corresponds to a position command value and a speed command value for the end-effector. Starting from the beginning of the time parameter sequence, the robot controller sequentially reads the position and speed command values ​​for the corresponding time points in the time parameter sequence at fixed control cycles of 0.002 seconds. The fixed control cycle of 0.002 seconds is based on the servo driver's current loop control cycle of 0.0001 seconds, speed loop control cycle of 0.0005 seconds, and position loop control cycle of 0.002 seconds. The robot controller uses the position loop control cycle as the command output cycle. The robot controller converts each position command value into an angular displacement command value for the steel belt drive wheel and each speed command value into an angular velocity command value for the steel belt drive wheel. The angular displacement command value of the steel belt drive pulley is obtained by the ratio of the linear displacement of the telescopic shaft to the radius of the steel belt drive pulley. The radius of the steel belt drive pulley is stored in the system parameters. In this embodiment, the radius of the steel belt drive pulley is 0.025 meters. The robot controller sends the angular displacement command value and angular velocity command value of the steel belt drive pulley to the servo driver. The servo driver drives the steel belt drive pulley to rotate according to the command. The steel belt drive pulley drives the steel belt driven pulley to rotate through the steel belt. The steel belt driven pulley drives the end effector to move along the target placement approach path.

[0047] In practice, during the placement action, the number of pulses from the rotary encoder on the steel belt drive pulley is continuously collected at a fixed sampling period. The fixed sampling period is set to 0.001 seconds, based on the fact that the highest response frequency of the rotary encoder on the steel belt drive pulley is 10kHz, and the sampling frequency corresponding to the fixed sampling period of 0.001 seconds is 1kHz. The sampling frequency is less than one-tenth of the highest response frequency of the rotary encoder to retain sampling margin. The rotary encoder is installed at the end of the shaft of the steel belt drive pulley. The rotary encoder outputs 4096 pulses per revolution, and the pulse signal output by the rotary encoder is received by the encoder interface module of the robot controller. The robot controller reads the pulse count value of the rotary encoder at the end of each fixed sampling period. The pulse count value is the cumulative number of pulses relative to the start time of the placement action. The pulse count value at each sampling time is divided by the number of pulses per revolution of the rotary encoder (4096), and then multiplied by 2π radians to obtain the instantaneous angular displacement value at each sampling time. The instantaneous angular displacement values ​​at each sampling time are arranged in the order of sampling time to form an angular displacement sequence.

[0048] In practice, the instantaneous rotational speed of the steel belt drive pulley is simultaneously collected. The instantaneous rotational speed of the steel belt drive pulley is obtained by dividing the difference between the instantaneous angular displacement values ​​of two adjacent sampling moments by a fixed sampling period. Specifically, the calculation method is to subtract the instantaneous angular displacement value of the previous sampling moment from the instantaneous angular displacement value of the later sampling moment, and then divide the difference by the fixed sampling period of 0.001 seconds. The instantaneous rotational speeds at each sampling moment are arranged in chronological order to form an angular velocity sequence. The instantaneous rotational speed of the first sampling moment in the angular velocity sequence is set to the commanded speed value at the start of the placement action, because the first sampling moment lacks the instantaneous angular displacement value of the previous sampling moment.

[0049] In practice, the placement position deviation is obtained by comparing the angular displacement value at the end of the angular displacement sequence with the expected angular displacement value corresponding to the theoretical placement position. The expected angular displacement value corresponding to the theoretical placement position is obtained by converting the end position command value of the target placement path. The end position command value is the linear displacement command value of the telescopic shaft when the robot's end effector reaches the placement position. The expected angular displacement value is equal to the end position command value divided by the radius of the steel belt drive pulley, which is 0.025 meters. The placement position deviation is obtained by subtracting the expected angular displacement value from the angular displacement value at the end of the sequence. The unit of the placement position deviation is radians, and the positive direction of the placement position deviation is the same as the positive rotation direction of the steel belt drive pulley.

[0050] In practical implementation, the speed tracking deviation is obtained by comparing the angular velocity values ​​at each sampling moment in the angular velocity sequence with the corresponding commanded speed values. The commanded speed value at the corresponding moment is determined by the angular velocity command value of the steel belt drive pulley at each time point in the time parameter sequence of the target workpiece approach path. When the sampling moment in the angular velocity sequence is inconsistent with the time point in the time parameter sequence, the commanded speed value at the corresponding moment is calculated by linear interpolation from the angular velocity command values ​​of adjacent time points in the time parameter sequence. The speed tracking deviation is calculated according to the following formula: in, This represents the root mean square value of the speed tracking deviation, in radians per second. The square root operator is represented by the symbol. Represents a fraction, with the numerator being... The denominator is ; This represents the total number of sampling points in the angular velocity sequence during the placement action, in this embodiment. Equals the execution time of the placement action divided by the fixed sampling period Second; Indicates to from arrive Summing all possible values; Indicates the sampling point number in the angular velocity sequence; Represents the angular velocity sequence of the th The instantaneous value of angular velocity at each sampling moment, in radians per second; Indicates the first The command velocity value corresponding to each sampling moment, in radians per second; superscript This represents the squaring operation.

[0051] In practice, the root mean square values ​​of the placement position deviation and the speed tracking deviation are combined to form a placement deviation vector. This placement deviation vector is a two-dimensional column vector, with the first row representing the placement position deviation and the second row representing the root mean square value of the speed tracking deviation. The placement deviation vector is stored in the system's memory area as the placement deviation vector for the current cycle.

[0052] In practice, the historical placement deviation vector, recorded at the end of the previous cycle, is read from the system storage area. The system storage area is a non-volatile memory; the historical placement deviation vector is written to the system storage area by the robot controller at the end of the previous cycle. When reading the historical placement deviation vector, it is retrieved from a preset address space in the system storage area. The historical placement deviation vector is a two-dimensional column vector; the first row represents the historical placement position deviation, and the second row represents the root mean square value of the historical speed tracking deviation. The placement deviation vector for the current cycle is a calculated two-dimensional column vector; the first row represents the placement position deviation for the current cycle, and the second row represents the root mean square value of the speed tracking deviation for the current cycle.

[0053] In practice, the difference vector between the current cycle's placement deviation vector and the historical placement deviation vector is calculated as the deviation change. The deviation change is calculated using the following formula: in, This represents the vector of deviation changes, which is a two-dimensional column vector. This represents the placement deviation vector for the current cycle. The placement deviation vector for the current cycle is a two-dimensional column vector. The first row of elements represents the placement position deviation in radians, and the second row of elements represents the root mean square value of the speed tracking deviation in radians per second. This represents the historical placement deviation vector of the previous cycle. The historical placement deviation vector is a two-dimensional column vector. The first row of elements represents the historical placement position deviation in radians, and the second row of elements represents the root mean square value of the historical speed tracking deviation in radians per second. The minus sign indicates vector subtraction. Vector subtraction is performed on the elements with the same row number in the two two-dimensional column vectors. Subtracting the elements in the first row yields the first row of the deviation change vector, and subtracting the elements in the second row yields the second row of the deviation change vector.

[0054] In practical implementation, the position deviation change component and the velocity deviation change component are extracted from the deviation change vector. The position deviation change component is the first element of the deviation change vector, and its unit is the same as that of the placement position deviation, both being radians. The velocity deviation change component is the second element of the deviation change vector, and its unit is the same as that of the root mean square value of the velocity tracking deviation, both being radians per second. The position deviation change component is the difference between the placement position deviation of the current cycle and the historical placement position deviation, while the velocity deviation change component is the difference between the root mean square value of the velocity tracking deviation of the current cycle and the historical root mean square value of the velocity tracking deviation.

[0055] In some embodiments, if the position deviation change component is greater than zero, the first adjustment step size is subtracted from the upper limit of the velocity constraint in the current motion parameter constraint matrix; if the position deviation change component is less than zero, the first adjustment step size is added to the upper limit of the velocity constraint in the current motion parameter constraint matrix; if the position deviation change component is equal to zero, the upper limit of the velocity constraint in the current motion parameter constraint matrix remains unchanged. The first adjustment step size is stored in the system parameters, and its value is 0.005 meters per second. The first adjustment step size is one-hundredth of the upper limit of the velocity constraint in the current motion parameter constraint matrix. The first adjustment step size is determined based on the principle that the adjustment amount in each cycle does not exceed one-hundredth of the current upper limit of the velocity constraint, thus avoiding drastic changes in the upper limit of the velocity constraint between adjacent cycles. The upper limit of the velocity constraint in the current motion parameter constraint matrix is ​​the upper limit of the velocity constraint in the constructed motion parameter constraint matrix.

[0056] In some embodiments, if the velocity deviation change component is greater than zero, the second adjustment step size is subtracted from the upper limit of the acceleration constraint in the current motion parameter constraint matrix; if the velocity deviation change component is less than zero, the second adjustment step size is added to the upper limit of the acceleration constraint in the current motion parameter constraint matrix; if the velocity deviation change component is equal to zero, the upper limit of the acceleration constraint in the current motion parameter constraint matrix remains unchanged. The second adjustment step size is stored in the system parameters, and its value is 0.01 meters per second squared. The second adjustment step size is one-hundredth of the upper limit of the acceleration constraint in the current motion parameter constraint matrix. The second adjustment step size is determined based on the principle that the adjustment amount in each cycle does not exceed one-hundredth of the upper limit of the current acceleration constraint, thus avoiding drastic changes in the upper limit of the acceleration constraint between adjacent cycles. The upper limit of the acceleration constraint in the current motion parameter constraint matrix is ​​the upper limit of the acceleration constraint in the constructed motion parameter constraint matrix.

[0057] In practice, the constraint boundary values ​​of the motion parameter constraint matrix for the next cycle are obtained. The adjusted upper limit value of the velocity constraint is used as the upper limit value of the velocity constraint in the motion parameter constraint matrix for the next cycle, and the adjusted upper limit value of the acceleration constraint is used as the upper limit value of the acceleration constraint in the motion parameter constraint matrix for the next cycle. The lower limit value of the velocity constraint in the motion parameter constraint matrix for the next cycle remains unchanged from the lower limit value of the current motion parameter constraint matrix, and the lower limit value of the acceleration constraint in the motion parameter constraint matrix for the next cycle also remains unchanged from the lower limit value of the current motion parameter constraint matrix.

[0058] It can be understood that a single-disc vacuum transfer system includes a memory, a processor, and a computer program stored in the memory and running on the processor. The memory is a random access memory, the processor is a central processing unit, and the computer program contains all the instructions for implementing the cyclic operation control method of the single-disc vacuum transfer system. When the processor executes the computer program, it implements the steps of the cyclic operation control method of the single-disc vacuum transfer system, including acquiring the initial vacuum degree sequence and the initial steel strip tension value sequence in response to the initialization completion signal, generating a motion parameter constraint matrix, extracting motion trajectory reference points to fit the wafer transfer reference path, acquiring the stage height value and horizontal deflection angle value to perform end-effector attitude compensation to obtain the target placement approach path, controlling the robot to perform the placement action and calculating the placement deviation vector, and adjusting the constraint boundary values ​​of the motion parameter constraint matrix for the next cycle according to the deviation change.

[0059] See Figure 5In the graph, the horizontal axis represents the cycle number, ranging from 0 to 600. The left side of the vertical axis represents the placement position deviation in radians, and the right side represents the root mean square value of the speed tracking deviation in radians per second. The solid blue curve represents the placement position deviation as the cycle number changes, and the dashed red curve represents the root mean square value of the speed tracking deviation as the cycle number changes.

[0060] From the image trend, the placement position deviation shows an overall decreasing trend as the cycle number increases. The maximum placement position deviation in the initial cycle is approximately 0.02 radians, gradually decreasing and approaching zero thereafter, with the final fluctuation range controlled within ±0.005 radians. The root mean square value of the speed tracking deviation also shows a similar decreasing trend, with a maximum value of approximately 0.1 rad / s in the initial cycle, gradually decreasing and eventually stabilizing in a near-zero range, with a fluctuation range of approximately ±0.02 rad / s.

[0061] This change reflects the process of adjusting the upper limits of velocity and acceleration constraints in the motion parameter constraint matrix for the next cycle by using the difference between the current cycle's placement deviation vector and the historical placement deviation vector from the previous cycle as the deviation change amount. Specifically, when the position deviation change component is large, the system correspondingly reduces the upper limit of the velocity constraint, thus decreasing the position deviation; when the velocity deviation change component is large, the system correspondingly reduces the upper limit of the acceleration constraint, thus reducing the velocity tracking deviation. The figure shows that after multiple cycles of adjustment, the positioning accuracy and velocity tracking performance of the placement action are significantly improved and tend to stabilize.

[0062] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A single-disc vacuum transfer system, comprising a robotic arm, characterized in that, It also includes a sealed robotic arm chamber, with an automatically opening and closing door above the chamber. A valve controlling the opening and closing of the chamber is located on one side of the robotic arm chamber. The front end of the robotic arm can pass laterally through the valve. The rear end of the robotic arm is fixed to a steel belt. The two ends of the steel belt are respectively connected to the outer edges of the driving pulley and the driven pulley. The steel belt is fixedly connected to the output belt of the synchronous belt drive mechanism through a first connecting block. The input end of the synchronous belt drive mechanism is connected to the drive mechanism, which drives the rotation.

2. A method for controlling the cyclic operation of a single-disc vacuum transfer system, characterized in that, The method, applied to the single-disc vacuum transfer system of claim 1, comprises: In response to the initialization completion signal, the initial vacuum degree sequence and the initial steel strip tension value sequence of the current cycle are obtained, and the motion parameter constraint matrix of the current cycle is generated based on the initial vacuum degree sequence and the initial steel strip tension value sequence. In each dimension of the motion parameter constraint matrix, multiple motion trajectory reference points are extracted using an equidistant sampling method. Based on these multiple motion trajectory reference points, the wafer transfer reference path for the current cycle is fitted and obtained. Collect the height and horizontal deflection angle of the wafer stage at the placement position within the current cycle. Based on the height and horizontal deflection angle, perform end-position compensation on the wafer transport reference path to obtain the target placement approach path. Based on the target placement approach path, the robot arm is controlled to perform the placement action of the current cycle. The angular displacement sequence and angular velocity sequence of the steel belt drive wheel are collected during the placement action. Based on the angular displacement sequence and angular velocity sequence, the placement deviation vector of the current cycle is calculated. Obtain the historical placement deviation vector of the previous cycle, and adjust the constraint boundary values ​​of the motion parameter constraint matrix for the next cycle based on the deviation change between the placement deviation vector of the current cycle and the historical placement deviation vector.

3. The method for cyclic operation control of a single-disc vacuum transfer system according to claim 2, characterized in that, The process of generating the motion parameter constraint matrix for the current cycle includes: After the robotic arm completes the return-to-origin operation and all axes are ready, the initialization completion signal is received, triggering the vacuum sensor group to continuously collect multiple instantaneous vacuum values ​​in the robotic arm cavity at a preset sampling frequency, forming the initial vacuum sequence. Simultaneously, the tension sensor group is triggered to continuously collect multiple instantaneous tension values ​​on the driving pulley side and the driven pulley side of the steel belt, forming the initial steel belt tension value sequence; The maximum and minimum vacuum values ​​are extracted from the initial vacuum value sequence, and the maximum and minimum tension values ​​are extracted from the initial steel strip tension value sequence, to form a boundary array of motion parameters; Based on the boundary array of motion parameters, a motion parameter constraint matrix is ​​constructed with the upper limit values ​​of the linear velocity and acceleration of the steel strip as variables. The row vectors of the motion parameter constraint matrix correspond to the acceleration constraint intervals of each motion axis, and the column vectors correspond to the velocity constraint intervals of each motion axis.

4. The method for cyclic operation control of a single-disc vacuum transfer system according to claim 3, characterized in that, In each dimension of the motion parameter constraint matrix, multiple motion trajectory reference points are extracted using an equidistant sampling method. Based on these multiple motion trajectory reference points, a wafer transfer reference path for the current cycle is fitted, including: The movement space of the wafer from the pick-up position to the placement position is divided into multiple interpolation segments. During the start and end time of each interpolation segment, multiple path points are extracted at equal time intervals to obtain multiple motion trajectory reference points for each interpolation segment. Read the velocity and acceleration constraint intervals corresponding to the current interpolation segment in the motion parameter constraint matrix, and determine whether the instantaneous velocity and instantaneous acceleration of each motion trajectory reference point fall within the velocity and acceleration constraint intervals. If so, retain the motion trajectory reference point; otherwise, discard the motion trajectory reference point. All retained motion trajectory reference points are fitted using cubic spline interpolation to obtain a continuous and smooth wafer transfer reference path, which is represented in a time-parameterized form.

5. The method for cyclic operation control of a single-disc vacuum transfer system according to claim 2, characterized in that, The method of obtaining the target placement approach path includes: When the end effector of the robot reaches the preset deceleration zone near the placement position, it triggers the laser displacement sensor installed on the substrate of the substrate to measure the height difference between the upper surface of the substrate and the reference plane of the robot end effector, which is used as the height value of the substrate. Simultaneously, the angle sensor installed on the side of the plate-bearing stage is triggered to measure the angle between the central axis of the plate-bearing stage and the feed direction of the robot end effector, which is used as the horizontal deflection angle value. Based on the height of the support platform, calculate the height compensation of the end effector in the vertical direction, and based on the horizontal deflection angle, calculate the rotation compensation of the end effector in the horizontal plane. The height compensation and rotation compensation are superimposed on the end pose parameters of the wafer transfer reference path to generate the target placement approach path.

6. The method for cyclic operation control of a single-disc vacuum transfer system according to claim 2, characterized in that, Based on the target placement approach path, the robot arm is controlled to execute the placement action for the current cycle. The angular displacement and angular velocity sequences of the steel belt drive pulley are collected during the placement action. Based on these sequences, the placement deviation vector for the current cycle is calculated, including: According to the time parameter sequence of the target placement path, the position and speed commands of each interpolation point are output to the servo driver in turn, so that the steel belt drive wheel rotates according to the command and drives the end effector to move along the target placement path. During the placement action, the number of rotary encoder pulses on the steel belt drive pulley is continuously collected at a fixed sampling period. The number of pulses at each sampling moment is converted into the instantaneous value of angular displacement to form the angular displacement sequence. At the same time, the instantaneous rotational speed of the steel belt drive pulley is collected to form the angular velocity sequence. The placement position deviation is obtained by the difference between the angular displacement value at the end of the angular displacement sequence and the expected angular displacement value corresponding to the theoretical placement position. Based on the difference between the angular velocity value at each sampling moment in the angular velocity sequence and the command velocity value at the corresponding moment, the speed tracking deviation is obtained, and the placement position deviation and the speed tracking deviation are combined into the placement position deviation vector.

7. The method for cyclic operation control of a single-disc vacuum transfer system according to claim 2, characterized in that, Obtain the historical placement deviation vector from the previous cycle. Based on the change in deviation between the current cycle's placement deviation vector and the historical placement deviation vector, adjust the constraint boundary values ​​of the motion parameter constraint matrix for the next cycle, including: Read the historical placement deviation vector recorded at the end of the previous cycle from the system storage area, and calculate the difference vector between the placement deviation vector of the current cycle and the historical placement deviation vector as the deviation change. Extract the position deviation change component and the velocity deviation change component from the deviation change amount. If the position deviation change component is greater than zero, subtract the first adjustment step size from the velocity constraint upper limit value of the current motion parameter constraint matrix. If the position deviation change component is less than zero, increase the first adjustment step size from the velocity constraint upper limit value of the current motion parameter constraint matrix. If the velocity deviation change component is greater than zero, then the second adjustment step size is subtracted from the upper limit value of the acceleration constraint of the current motion parameter constraint matrix. If the velocity deviation change component is less than zero, then the second adjustment step size is added to the upper limit value of the acceleration constraint of the current motion parameter constraint matrix to obtain the constraint boundary value of the motion parameter constraint matrix for the next cycle.

8. The method for controlling the cyclic operation of a single-disc vacuum transfer system according to claim 3, characterized in that, The row vector dimension of the motion parameter constraint matrix corresponds to the lifting axis, rotation axis, and telescopic axis in the joint space of the robot hand, and the column vector dimension corresponds to the speed constraint range of each axis during the acceleration and deceleration phases.

9. A method for controlling the cyclic operation of a single-disc vacuum transfer system according to claim 4, characterized in that, During the start and end times of each interpolation segment, the equal time intervals are adaptively determined based on the response bandwidth of the steel belt drive wheel, and the number of motion trajectory reference points extracted in each interpolation segment is proportional to the total number of interpolation segments in the current cycle.

10. A method for controlling the cyclic operation of a single-disc vacuum transfer system according to claim 5, characterized in that, The measurement reference for the height of the plate bearing stage is the lower surface plane of the robot end effector when it is in a horizontal posture, and the measurement reference for the horizontal deflection angle is the theoretical feed direction vector of the end effector in the horizontal plane.

Images