WO2021153228A1 - Stage orientation estimation device, conveyance device, and stage orientation estimation method - Google Patents

Stage orientation estimation device, conveyance device, and stage orientation estimation method Download PDF

Info

Publication number
WO2021153228A1
WO2021153228A1 PCT/JP2021/000818 JP2021000818W WO2021153228A1 WO 2021153228 A1 WO2021153228 A1 WO 2021153228A1 JP 2021000818 W JP2021000818 W JP 2021000818W WO 2021153228 A1 WO2021153228 A1 WO 2021153228A1
Authority
WO
WIPO (PCT)
Prior art keywords
stage
estimation
input data
unit
posture estimation
Prior art date
Application number
PCT/JP2021/000818
Other languages
French (fr)
Japanese (ja)
Inventor
宏昭 臼本
Original Assignee
株式会社Ｓｃｒｅｅｎホールディングス
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 株式会社Ｓｃｒｅｅｎホールディングス filed Critical 株式会社Ｓｃｒｅｅｎホールディングス
Priority to KR1020227025709A priority Critical patent/KR20220120645A/en
Priority to CN202180011089.XA priority patent/CN115023660A/en
Publication of WO2021153228A1 publication Critical patent/WO2021153228A1/en

Links

Images

Classifications

    • G—PHYSICS
    • G03—PHOTOGRAPHY; CINEMATOGRAPHY; ANALOGOUS TECHNIQUES USING WAVES OTHER THAN OPTICAL WAVES; ELECTROGRAPHY; HOLOGRAPHY
    • G03F—PHOTOMECHANICAL PRODUCTION OF TEXTURED OR PATTERNED SURFACES, e.g. FOR PRINTING, FOR PROCESSING OF SEMICONDUCTOR DEVICES; MATERIALS THEREFOR; ORIGINALS THEREFOR; APPARATUS SPECIALLY ADAPTED THEREFOR
    • G03F7/00—Photomechanical, e.g. photolithographic, production of textured or patterned surfaces, e.g. printing surfaces; Materials therefor, e.g. comprising photoresists; Apparatus specially adapted therefor
    • G03F7/70—Microphotolithographic exposure; Apparatus therefor
    • G03F7/70691—Handling of masks or workpieces
    • G03F7/70783—Handling stress or warp of chucks, masks or workpieces, e.g. to compensate for imaging errors or considerations related to warpage of masks or workpieces due to their own weight
    • G—PHYSICS
    • G03—PHOTOGRAPHY; CINEMATOGRAPHY; ANALOGOUS TECHNIQUES USING WAVES OTHER THAN OPTICAL WAVES; ELECTROGRAPHY; HOLOGRAPHY
    • G03F—PHOTOMECHANICAL PRODUCTION OF TEXTURED OR PATTERNED SURFACES, e.g. FOR PRINTING, FOR PROCESSING OF SEMICONDUCTOR DEVICES; MATERIALS THEREFOR; ORIGINALS THEREFOR; APPARATUS SPECIALLY ADAPTED THEREFOR
    • G03F7/00—Photomechanical, e.g. photolithographic, production of textured or patterned surfaces, e.g. printing surfaces; Materials therefor, e.g. comprising photoresists; Apparatus specially adapted therefor
    • G03F7/20—Exposure; Apparatus therefor
    • G—PHYSICS
    • G03—PHOTOGRAPHY; CINEMATOGRAPHY; ANALOGOUS TECHNIQUES USING WAVES OTHER THAN OPTICAL WAVES; ELECTROGRAPHY; HOLOGRAPHY
    • G03F—PHOTOMECHANICAL PRODUCTION OF TEXTURED OR PATTERNED SURFACES, e.g. FOR PRINTING, FOR PROCESSING OF SEMICONDUCTOR DEVICES; MATERIALS THEREFOR; ORIGINALS THEREFOR; APPARATUS SPECIALLY ADAPTED THEREFOR
    • G03F7/00—Photomechanical, e.g. photolithographic, production of textured or patterned surfaces, e.g. printing surfaces; Materials therefor, e.g. comprising photoresists; Apparatus specially adapted therefor
    • G03F7/70—Microphotolithographic exposure; Apparatus therefor
    • G03F7/70216—Mask projection systems
    • G03F7/70283—Mask effects on the imaging process
    • G03F7/70291—Addressable masks, e.g. spatial light modulators [SLMs], digital micro-mirror devices [DMDs] or liquid crystal display [LCD] patterning devices
    • G—PHYSICS
    • G03—PHOTOGRAPHY; CINEMATOGRAPHY; ANALOGOUS TECHNIQUES USING WAVES OTHER THAN OPTICAL WAVES; ELECTROGRAPHY; HOLOGRAPHY
    • G03F—PHOTOMECHANICAL PRODUCTION OF TEXTURED OR PATTERNED SURFACES, e.g. FOR PRINTING, FOR PROCESSING OF SEMICONDUCTOR DEVICES; MATERIALS THEREFOR; ORIGINALS THEREFOR; APPARATUS SPECIALLY ADAPTED THEREFOR
    • G03F7/00—Photomechanical, e.g. photolithographic, production of textured or patterned surfaces, e.g. printing surfaces; Materials therefor, e.g. comprising photoresists; Apparatus specially adapted therefor
    • G03F7/70—Microphotolithographic exposure; Apparatus therefor
    • G03F7/70216—Mask projection systems
    • G03F7/70358—Scanning exposure, i.e. relative movement of patterned beam and workpiece during imaging
    • G—PHYSICS
    • G03—PHOTOGRAPHY; CINEMATOGRAPHY; ANALOGOUS TECHNIQUES USING WAVES OTHER THAN OPTICAL WAVES; ELECTROGRAPHY; HOLOGRAPHY
    • G03F—PHOTOMECHANICAL PRODUCTION OF TEXTURED OR PATTERNED SURFACES, e.g. FOR PRINTING, FOR PROCESSING OF SEMICONDUCTOR DEVICES; MATERIALS THEREFOR; ORIGINALS THEREFOR; APPARATUS SPECIALLY ADAPTED THEREFOR
    • G03F7/00—Photomechanical, e.g. photolithographic, production of textured or patterned surfaces, e.g. printing surfaces; Materials therefor, e.g. comprising photoresists; Apparatus specially adapted therefor
    • G03F7/70—Microphotolithographic exposure; Apparatus therefor
    • G03F7/70691—Handling of masks or workpieces
    • G03F7/70716—Stages
    • H—ELECTRICITY
    • H01—ELECTRIC ELEMENTS
    • H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
    • H01L21/00—Processes or apparatus adapted for the manufacture or treatment of semiconductor or solid state devices or of parts thereof
    • H01L21/67—Apparatus specially adapted for handling semiconductor or electric solid state devices during manufacture or treatment thereof; Apparatus specially adapted for handling wafers during manufacture or treatment of semiconductor or electric solid state devices or components ; Apparatus not specifically provided for elsewhere
    • H01L21/67005—Apparatus not specifically provided for elsewhere
    • H01L21/67242—Apparatus for monitoring, sorting or marking
    • H01L21/67288—Monitoring of warpage, curvature, damage, defects or the like
    • H—ELECTRICITY
    • H01—ELECTRIC ELEMENTS
    • H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
    • H01L21/00—Processes or apparatus adapted for the manufacture or treatment of semiconductor or solid state devices or of parts thereof
    • H01L21/67—Apparatus specially adapted for handling semiconductor or electric solid state devices during manufacture or treatment thereof; Apparatus specially adapted for handling wafers during manufacture or treatment of semiconductor or electric solid state devices or components ; Apparatus not specifically provided for elsewhere
    • H01L21/68—Apparatus specially adapted for handling semiconductor or electric solid state devices during manufacture or treatment thereof; Apparatus specially adapted for handling wafers during manufacture or treatment of semiconductor or electric solid state devices or components ; Apparatus not specifically provided for elsewhere for positioning, orientation or alignment
    • H—ELECTRICITY
    • H01—ELECTRIC ELEMENTS
    • H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
    • H01L21/00—Processes or apparatus adapted for the manufacture or treatment of semiconductor or solid state devices or of parts thereof
    • H01L21/67—Apparatus specially adapted for handling semiconductor or electric solid state devices during manufacture or treatment thereof; Apparatus specially adapted for handling wafers during manufacture or treatment of semiconductor or electric solid state devices or components ; Apparatus not specifically provided for elsewhere
    • H01L21/683—Apparatus specially adapted for handling semiconductor or electric solid state devices during manufacture or treatment thereof; Apparatus specially adapted for handling wafers during manufacture or treatment of semiconductor or electric solid state devices or components ; Apparatus not specifically provided for elsewhere for supporting or gripping
    • H01L21/687—Apparatus specially adapted for handling semiconductor or electric solid state devices during manufacture or treatment thereof; Apparatus specially adapted for handling wafers during manufacture or treatment of semiconductor or electric solid state devices or components ; Apparatus not specifically provided for elsewhere for supporting or gripping using mechanical means, e.g. chucks, clamps or pinches
    • H01L21/68714—Apparatus specially adapted for handling semiconductor or electric solid state devices during manufacture or treatment thereof; Apparatus specially adapted for handling wafers during manufacture or treatment of semiconductor or electric solid state devices or components ; Apparatus not specifically provided for elsewhere for supporting or gripping using mechanical means, e.g. chucks, clamps or pinches the wafers being placed on a susceptor, stage or support
    • H01L21/68764—Apparatus specially adapted for handling semiconductor or electric solid state devices during manufacture or treatment thereof; Apparatus specially adapted for handling wafers during manufacture or treatment of semiconductor or electric solid state devices or components ; Apparatus not specifically provided for elsewhere for supporting or gripping using mechanical means, e.g. chucks, clamps or pinches the wafers being placed on a susceptor, stage or support characterised by a movable susceptor, stage or support, others than those only rotating on their own vertical axis, e.g. susceptors on a rotating caroussel

Definitions

  • the present invention relates to a technique for estimating the yawing angle of a stage in a transfer device that conveys a flat plate-shaped stage by a transfer mechanism.
  • Patent Document 1 describes a device that draws an exposure pattern on the upper surface of a substrate (W) while moving a stage (10) on which a substrate (W) is placed by a stage moving mechanism (20). ..
  • the stage transfer device mounted on this type of device may be equipped with a pair of straight-ahead mechanisms. Specifically, there is known a mechanism for transporting a stage in a predetermined direction by a pair of linear motors provided in parallel with each other.
  • the conventional transfer device was equipped with a large-scale measuring device in order to grasp the fluctuation of the yawing angle described above. Then, the operation of the transport device was corrected based on the measurement result of the measuring device. However, if a large-scale measuring device is mounted, it becomes difficult to miniaturize the transport device. In addition, by mounting the measuring device, the manufacturing cost of the transport device also increases.
  • each of the plurality of machine learning algorithms has strengths and weaknesses, and the estimation accuracy of each machine learning algorithm varies depending on the operating condition of the transfer device. Therefore, depending on only one trained model generated by one machine learning algorithm, it may not be possible to perform accurate estimation depending on the operating condition of the transport device.
  • the present invention has been made in view of such circumstances, and can estimate the yawing angle of the stage without constantly installing a large-scale measuring device, and realizes high estimation accuracy in many operating conditions of the transport device.
  • the purpose is to provide the technology that can be used.
  • the first invention of the present application is a stage attitude estimation device that estimates the yawing angle of the stage in a transfer device that conveys a flat plate-shaped stage by a transfer mechanism, and is output from the transfer mechanism.
  • An input data acquisition unit that acquires the measured value or a value calculated based on the measured value as input data, and an attitude estimation that estimates the yawing angle of the stage based on the input data and outputs the estimation result.
  • the posture estimation unit includes a unit, a plurality of trained models that output a tentative estimated value of the yawing angle of the stage based on the input data, and a plurality of the trained models that are output from the plurality of trained models. It has an estimated value determining unit that determines the estimated result based on the tentative estimated value.
  • the second invention of the present application is the stage posture estimation device of the first invention, and the plurality of trained models are generated by algorithms different from each other.
  • the third invention of the present application is the stage posture estimation device of the first invention or the second invention, and the estimated value determining unit uses the average value of the plurality of tentative estimated values as the estimated result.
  • the fourth invention of the present application is the stage posture estimation device of the first invention or the second invention, in which weight ratios are set in the plurality of trained models, and the estimated value determining unit determines the weight ratios.
  • the weighted average value of the plurality of tentative estimated values used is used as the estimated result.
  • the fifth invention of the present application is the stage posture estimation device of the fourth invention, and the estimated value determining unit changes the weighting ratio based on a state variable representing an operating state of the transport mechanism.
  • the sixth invention of the present application is the stage posture estimation device of the first invention or the second invention, and the estimated value determining unit is based on a state variable representing an operating state of the transport mechanism, and has a plurality of tentative estimated values. One of them is selected, and the selected tentative estimated value is used as the estimated result.
  • the seventh invention of the present application is the stage attitude estimation device of any one of the first to sixth inventions, wherein the transport mechanism transports the stage by a pair of straight-moving mechanisms, and the input data acquisition unit. Generates the input data based on the difference between the torque values of the pair of straight-ahead mechanisms.
  • the eighth invention of the present application is a transport device, and includes the stage posture estimation device of any one of the first to seventh inventions, the stage, and the transport mechanism.
  • the ninth invention of the present application is a stage attitude estimation method for estimating the yawing angle of the stage in a transport device that transports a flat plate-shaped stage by a transport mechanism, and a) a measured value output from the transport mechanism or the said.
  • the step includes a step of acquiring a value calculated based on a measured value as input data, and b) a step of estimating a yawing angle of the stage based on the input data and outputting an estimation result.
  • the input data is input to the plurality of trained models, and the estimation result is determined based on the plurality of tentative estimation values output from the plurality of trained models.
  • the yawing angle of the stage is estimated based on the measured value output from the transport mechanism or the value calculated based on the measured value. This makes it possible to estimate the yawing angle of the stage without constantly installing a large-scale measuring device.
  • one estimation result is output based on a plurality of tentative estimates output from the plurality of trained models. As a result, high estimation accuracy can be realized in many operating conditions of the transport device.
  • main scanning direction the direction in which the stage moves by the pair of straight-moving mechanisms
  • secondary scanning direction the direction orthogonal to the main scanning direction
  • FIG. 1 is a perspective view of a drawing device 1 provided with a transfer device 10 according to an embodiment of the present invention.
  • FIG. 2 is a schematic top view of the drawing apparatus 1.
  • the drawing device 1 is a device that irradiates the upper surface of a substrate W such as a semiconductor substrate or a glass substrate coated with a photosensitive material with spatially modulated light to draw an exposure pattern on the upper surface of the substrate W.
  • the drawing device 1 includes a transfer device 10, a frame 20, a drawing processing unit 30, and a control unit 40.
  • the transport device 10 is a device that transports the flat plate-shaped stage 12 in a substantially constant posture in the horizontal direction on the upper surface of the base 11.
  • the transport device 10 has a transport mechanism including a main scanning mechanism 13 and a sub scanning mechanism 14.
  • the main scanning mechanism 13 is a mechanism for transporting the stage 12 in the main scanning direction.
  • the sub-scanning mechanism 14 is a mechanism for transporting the stage 12 in the sub-scanning direction.
  • the substrate W is held on the upper surface of the stage 12 in a horizontal posture, and moves together with the stage 12 in the main scanning direction and the sub-scanning direction.
  • the frame 20 has a structure for holding the drawing processing unit 30 above the base 11.
  • the frame 20 has a pair of strut portions 21 and a cross-linking portion 22.
  • the pair of support columns 21 are erected at intervals in the sub-scanning direction.
  • Each support column 21 extends upward from the upper surface of the base 11.
  • the cross-linked portion 22 extends in the sub-scanning direction between the upper ends of the two strut portions 21.
  • the stage 12 holding the substrate W passes between the pair of support columns 21 and below the cross-linked portion 22.
  • the drawing processing unit 30 has two optical heads 31, an illumination optical system 32, a laser oscillator 33, and a laser drive unit 34.
  • the two optical heads 31 are fixed to the cross-linked portion 22 at intervals in the sub-scanning direction.
  • the illumination optical system 32, the laser oscillator 33, and the laser drive unit 34 are housed in, for example, the internal space of the cross-linking unit 22.
  • the laser drive unit 34 is electrically connected to the laser oscillator 33. When the laser drive unit 34 is operated, pulsed light is emitted from the laser oscillator 33. Then, the pulsed light emitted from the laser oscillator 33 is introduced into the optical head 31 via the illumination optical system 32.
  • An optical system including a spatial modulator is provided inside the optical head 31.
  • the spatial modulator for example, GLV (Grating Light Valve) (registered trademark), which is a diffraction grating type spatial light modulator, is used.
  • the pulsed light introduced into the optical head 31 is modulated into a predetermined pattern by the spatial modulator and is irradiated on the upper surface of the substrate W. As a result, a photosensitive material such as a resist coated on the upper surface of the substrate W is exposed.
  • the drawing device 1 When the drawing device 1 is in operation, the exposure by the optical head 31 and the transfer of the substrate W by the transfer device 10 are repeatedly executed. Specifically, the sub-scanning mechanism 14 conveys the stage 12 in the sub-scanning direction, and irradiates pulsed light from the optical head 31 to expose a band-shaped region (swath) extending in the sub-scanning direction. After that, the main scanning mechanism 13 conveys the stage 12 in the main scanning direction by one swath. The drawing apparatus 1 draws a pattern on the entire upper surface of the substrate W by repeating such exposure in the sub-scanning direction and transfer of the stage 12 in the main scanning direction.
  • the sub-scanning mechanism 14 conveys the stage 12 in the sub-scanning direction, and irradiates pulsed light from the optical head 31 to expose a band-shaped region (swath) extending in the sub-scanning direction.
  • the main scanning mechanism 13 conveys the stage 12 in the main scanning direction by one swa
  • the control unit 40 is a means for controlling the operation of each unit of the drawing device 1.
  • FIG. 3 is a block diagram showing an electrical connection between the control unit 40 and each unit in the drawing device 1.
  • the control unit 40 is composed of a computer having a processor 41 such as a CPU, a memory 42 such as a RAM, and a storage unit 43 such as a hard disk drive.
  • the storage unit 43 stores a computer program P for controlling the operation of the drawing device 1.
  • the control unit 40 includes a drawing processing unit 30 (including the optical head 31 and the laser driving unit 34 described above) and a main scanning mechanism 13 (including a linear motor 61 and an air guide 62 described later). , A sub-scanning mechanism 14 (including a linear motor 71 described later), a rotating mechanism 15 described later, and various sensors 50 are electrically connected.
  • the control unit 40 reads the computer program P and data D stored in the storage unit 43 into the memory 42, and the processor 41 performs arithmetic processing based on the computer program P and data D in the drawing device 1. The operation of each of the above parts is controlled. As a result, the drawing process in the drawing device 1 proceeds.
  • FIG. 4 is a partial cross-sectional view of a portion of the transport device 10 cut along a plane perpendicular to the main scanning direction.
  • the transport device 10 includes a base 11, a stage 12, a main scanning mechanism 13, a sub-scanning mechanism 14, a rotating mechanism 15, a lower support plate 16, a middle support plate 17, and a stage posture estimation. It has a device 18.
  • the base 11 is a support base that supports each part of the transport device 10.
  • the base 11 has a flat plate-like outer shape that extends in the main scanning direction and the sub-scanning direction.
  • Four legs 111 and two dampers 112 are provided on the lower surface of the base 11. The lengths of the legs 111 and the damper 112 can be adjusted individually. Therefore, the posture of the base 11 can be adjusted horizontally by adjusting the lengths of the legs 111 and the damper 112.
  • the lower support plate 16, the middle support plate 17, and the stage 12 each have a flat outer shape.
  • the lower support plate 16 is movably supported in the main scanning direction by the main scanning mechanism 13 on the base 11.
  • the middle-stage support plate 17 is movably supported on the lower-stage support plate 16 by the sub-scanning mechanism 14 in the sub-scanning direction.
  • the stage 12 is rotatably supported around a vertical axis by a rotation mechanism 15 on the middle support plate 17.
  • the stage 12 has an upper surface on which the substrate W can be placed. Further, on the upper surface of the stage 12, a chuck pin for holding the substrate W or a plurality of suction holes for sucking the substrate W are provided.
  • the main scanning mechanism 13 is a mechanism for moving the lower support plate 16 with respect to the base 11 in the main scanning direction.
  • the main scanning mechanism 13 has a pair of straight traveling mechanisms 60.
  • the pair of straight-moving mechanisms 60 are provided at both ends of the upper surface of the base 11 in the sub-scanning direction. As shown in FIGS. 2 and 4, each of the pair of straight-moving mechanisms 60 has a linear motor 61 and an air guide 62, respectively.
  • the linear motor 61 has a stator 611 and a mover 612.
  • the stator 611 is laid on the upper surface of the base 11 along the main scanning direction. That is, the pair of stators 611 are arranged parallel to each other.
  • the mover 612 is fixed to the lower support plate 16 via an air bearing 622 described later.
  • the main scanning mechanism 13 has a control board 63 for controlling the operation of the linear motor 61.
  • a control board 63 for example, a servo pack (registered trademark) is used.
  • the control board 63 is electrically connected to the control unit 40.
  • the control board 63 calculates the torque to be generated in the linear motor 61 according to the command from the control unit 40.
  • a drive signal corresponding to the calculated torque is supplied to the stator 611 of each linear motor 61.
  • the mover 612 moves in the main scanning direction along the stator 611 due to the magnetic attraction and repulsion generated between the stator 611 and the mover 612.
  • the air guide 62 has a guide rail 621 and an air bearing 622.
  • the guide rail 621 is laid on the upper surface of the base 11 along the main scanning direction. That is, the stator 611 of the linear motor 61 and the guide rail 621 of the air guide 62 are arranged in parallel with each other.
  • the air bearing 622 is fixed to the lower support plate 16 and the mover 612. Further, the air bearing 622 is arranged above the guide rail 621.
  • a gas outlet 623 is provided on the lower surface of the air bearing 622.
  • gas is constantly supplied from the utility of the factory to the air bearing 622, and the pressurized gas is blown out from the gas outlet 623 toward the upper surface of the guide rail 621.
  • the air bearing 622 is floated and supported on the guide rail 621 in a non-contact manner. Therefore, when the linear motor 61 is driven, the lower support plate 16 moves smoothly along the main scanning direction with low friction in a state of being floated and supported by the air guide 62.
  • the sub-scanning mechanism 14 is a mechanism for moving the middle-stage support plate 17 with respect to the lower-stage support plate 16 in the sub-scanning direction.
  • the sub-scanning mechanism 14 includes a linear motor 71 and a pair of guide mechanisms 72.
  • the linear motor 71 is provided substantially in the center of the upper surface of the lower support plate 16 in the main scanning direction.
  • the linear motor 71 has a stator 711 and a mover 712.
  • the stator 711 is laid on the upper surface of the lower support plate 16 along the sub-scanning direction.
  • the mover 712 is fixed to the middle support plate 17. When the linear motor 71 is driven, the mover 712 moves in the sub-scanning direction along the stator 711 due to the magnetic attraction and repulsion generated between the stator 711 and the mover 712.
  • the pair of guide mechanisms 72 are provided at both ends of the upper surface of the lower support plate 16 in the main scanning direction.
  • the pair of guide mechanisms 72 each have a guide rail 721 and a ball bearing 722.
  • the guide rail 721 is laid on the upper surface of the lower support plate 16 along the sub-scanning direction.
  • the ball bearing 722 is fixed to the lower surface of the middle support plate 17. Further, the ball bearing 722 can move in the sub-scanning direction along the guide rail 721. Therefore, when the linear motor 71 is driven, the middle support plate 17 moves in the sub-scanning direction with respect to the lower support plate 16.
  • the rotation mechanism 15 is a mechanism for adjusting the angle around the vertical axis of the stage 12 with respect to the middle support plate 17.
  • a motor is used for the rotation mechanism 15.
  • the stage 12 rotates about the vertical axis with respect to the middle support plate 17.
  • the angle (yaw angle) ⁇ around the vertical axis of the stage 12 can be adjusted.
  • the stage 12 can be moved in the main scanning direction and the sub-scanning direction with respect to the base 11 by the main scanning mechanism 13, the sub-scanning mechanism 14, and the rotation mechanism 15, and the yawing angle ⁇ is adjusted. It is possible to do.
  • the posture measuring device 80 can be installed in the transport device 10.
  • the posture measuring device 80 is a device for measuring the yawing angle ⁇ of the stage 12.
  • the posture measuring device 80 includes a mirror 81 fixed to the stage 12 and a laser interferometer 82.
  • the mirror 81 is fixed to the edge portion of the stage 12 in the main scanning direction.
  • the laser interferometer 82 is fixed to the upper surface of the base 11.
  • the laser interferometer 82 irradiates the mirror 81 with two laser beams. Then, the optical path difference between the two laser beams is detected by the interference of the two laser beams reflected from the mirror 81. Then, the yawing angle ⁇ of the stage 12 is measured based on the optical path difference.
  • the posture measuring device 80 is installed when performing the pre-learning process described later. After the pre-learning is completed, the posture measuring device 80 can be removed and the transport device 10 can be used.
  • FIG. 5 is a block diagram showing the configuration of the stage posture estimation device 18.
  • the stage posture estimation device 18 is a device that estimates the yawing angle ⁇ of the stage 12 based on the measured values output from the main scanning mechanism 13.
  • the stage posture estimation device 18 includes an input data acquisition unit 91 and a posture estimation unit 92.
  • the input data acquisition unit 91 has a measurement value input unit 911 and an input data generation unit 912.
  • the posture estimation unit 92 has a plurality of trained models M1, M2, M3, ..., And an estimation value determination unit 921.
  • the stage attitude estimation device 18 is composed of a computer having a processor such as a CPU, a memory such as a RAM, and a storage unit such as a hard disk drive. Each function of the measurement value input unit 911, the input data generation unit 912, and the estimation value determination unit 921 is realized by operating the processor according to the computer program stored in the storage unit.
  • the trained models M1, M2, M3, ... Are inference programs whose parameters have been adjusted by pre-learning using a machine learning algorithm.
  • Machine learning algorithms for obtaining trained models M1, M2, M3, ... include, for example, a neural network including a layer neural network and deep learning, a decision tree algorithm including a random forest, gradient boosting, and the like. So-called supervised machine learning algorithms, such as support vector machines, are used.
  • the plurality of trained models M1, M2, M3, ... Are generated by different machine learning algorithms, respectively.
  • the stage posture estimation device 18 may be configured by the same computer as the control unit 40 described above, or may be configured by a computer different from the control unit 40.
  • FIG. 6 is a flowchart showing the flow of the pre-learning process.
  • the posture measuring device 80 described above is installed in the transport device 10 (step S11). Then, the next steps S12 to S15 are repeatedly executed while operating the main scanning mechanism 13.
  • the measured value output from the control board 63 is input to the measured value input unit 911 (step S12).
  • the measured value is, for example, the torque value of the pair of linear motors 61 of the main scanning mechanism 13.
  • the measured value input to the measured value input unit 911 includes other items such as the air pressure of the air bearing 622, the temperature of the guide rail 621, the driving sound of the transport device 10, the vibration of the stage 12, and the position of the stage 12. It may be measured by various sensors 50.
  • the input data generation unit 912 generates the input data d based on the measurement value input to the measurement value input unit 911 (step S13). For example, when the measured values are the torque values of the pair of linear motors 61, the input data generation unit 912 calculates the difference between the torque values. Then, the input data d is generated by removing unnecessary frequencies from the calculated difference time series data. Even when the measured value input to the measured value input unit 911 is other than the torque value, the input data generation unit 912 generates input data d suitable for machine learning by performing a predetermined calculation and filtering process. ..
  • the input data generation unit 912 may use the measurement value itself input to the measurement value input unit 911 as the input data d. That is, the input data generation unit 912 may use the measured value output from the transport mechanism or the value calculated by performing a predetermined calculation and the filtering process on the measured value as the input data d.
  • the posture estimation unit 92 performs machine learning using the input data d generated by the input data generation unit 912 as input and the measurement result ⁇ m of the posture measuring device 80 as teacher data (step S14). That is, the posture estimation unit 92 learns the relationship between the input data d and the measurement result ⁇ m of the posture measuring device 80 by the machine learning algorithm described above.
  • the posture estimation unit 92 of the present embodiment performs the machine learning of step S14 in parallel by a plurality of different machine learning algorithms. Therefore, the machine learning in step S14 generates a plurality of different trained models M1, M2, M3, ...
  • the posture estimation unit 92 compares the output values of the trained models M1, M2, M3, ... Generated by the machine learning in step S14 with the measurement result ⁇ m which is the teacher data. Then, when the difference between the output values of the trained models M1, M2, M3, ... And the measurement result ⁇ m is not equal to or less than the preset threshold value, the posture estimation unit 92 determines the trained models M1, M2. , M3, ... It is determined that the estimation accuracy has not reached a desired level (step S15: no). In this case, the stage posture estimation device 18 repeats the processes of steps S12 to S14 described above. By repeating machine learning in this way, the estimation accuracy of the trained models M1, M2, M3, ... Is gradually improved.
  • the posture estimation unit 92 causes each trained model M1, It is determined that the estimation accuracy of M2, M3, ... Has reached a desired level. In this case, the posture estimation unit 92 ends machine learning (step S15: yes). Then, the posture measuring device 80 is removed from the transport device 10 (step S16). The stage posture estimation device 18 may end machine learning in step S15 when the number of repetitions of steps S12 to S14 reaches a preset upper limit value.
  • FIG. 7 is a flowchart showing the flow of the estimation process.
  • the measured value output from the control board 63 is input to the measured value input unit 911 (step S21).
  • the same type of measurement value as in step S12 described above is input.
  • the measured value input in step S12 is the torque value of the pair of linear motors 61
  • the measured value input in step S21 is also the torque value of the pair of linear motors 61.
  • the input data generation unit 912 generates the input data d based on the measurement value input to the measurement value input unit 911 (step S22).
  • the same process as in step S13 described above is executed. That is, when the process executed in step S13 is the difference calculation and the filter process, the input data d is generated by executing the difference calculation and the filter process in the step S22 as well.
  • the posture estimation unit 92 inputs the generated input data d into the plurality of trained models M1, M2, M3, ... (Step S23). Then, the trained models M1, M2, M3, ... Output the tentative estimated values ⁇ 1, ⁇ 2, ⁇ 3, ... Of the yawing angle ⁇ of the stage 12 corresponding to the input data d. As a result, a plurality of tentative estimated values ⁇ 1, ⁇ 2, ⁇ 3, ... Are obtained for one input data d (step S24).
  • the estimation value determination unit 921 of the posture estimation unit 92 determines one estimation result ⁇ r based on the plurality of provisional estimation values ⁇ 1, ⁇ 2, ⁇ 3, ... (Step S25). Specifically, for example, the estimation value determination unit 921 calculates the average value of a plurality of provisional estimation values ⁇ 1, ⁇ 2, ⁇ 3, ..., And determines the calculated average value as the estimation result ⁇ r. However, the estimated value determination unit 921 may determine the estimated result ⁇ r by another method.
  • FIG. 8 is a flowchart showing a second example of a method of determining the estimation result by the estimation value determination unit 921.
  • the weighting ratios w1, w2, w3, ... are set in advance for each of the trained models M1, M2, M3, ...
  • the weighting ratios w1, w2, w3, ... Are stored in advance in the storage unit of the computer constituting the stage posture estimation device 18.
  • the estimated value determining unit 921 first reads the weighted ratios w1, w2, w3, ... From the storage unit (step S31). Then, the estimated value determination unit 921 calculates and calculates the weighted average value of a plurality of tentative estimated values ⁇ 1, ⁇ 2, ⁇ 3, ... Using the read weighted ratios w1, w2, w3, ... Let the weighted average value be the estimation result ⁇ r (step S32).
  • the estimation result ⁇ r using the weighted average in step S32 can be calculated by the following equation (1).
  • ⁇ r (w1, ⁇ 1 + w2, ⁇ 2 + w3, ⁇ 3) / (w1 + w2 + w3) (1)
  • the weighting ratio of the trained model is relatively set. It is good to set it high. Then, by calculating the weighted average value as described above, a more preferable estimation result ⁇ r can be obtained.
  • FIG. 9 is a flowchart showing a third example of a method of determining the estimation result by the estimation value determination unit 921.
  • the weighting ratios w1, w2, w3, ... Corresponding to each of the trained models M1, M2, M3, ... are not fixed values but depend on the operating state of the main scanning mechanism 13. Change.
  • the estimated value determining unit 921 first acquires a state variable representing the operating state of the main scanning mechanism 13 (step S41).
  • the state variable may be a measured value input to the measured value input unit 911 described above, a variable acquired by another sensor, or an operation set by the user for the control unit 40. It may be a mode or the like.
  • the estimated value determination unit 921 changes the weighting ratios w1, w2, w3, ... Based on the acquired state variables (step S42). As a result, the weighting ratio of the trained model capable of exhibiting high estimation accuracy is increased according to the operating state of the main scanning mechanism 13. For example, in the operation state represented by a certain state variable, when the estimation accuracy of the trained model M2 is particularly high among the plurality of trained models M1, M2, M3, ... , The plurality of weighting ratios w1, w2, w3, ... Are changed so that the value of the weighting ratio w2 becomes relatively high.
  • the estimated value determination unit 921 calculates and calculates the weighted average value of the plurality of tentative estimated values ⁇ 1, ⁇ 2, ⁇ 3, ... Using the changed weighted ratios w1, w2, w3, ... Let the weighted average value be the estimation result ⁇ r (step S43).
  • the trained model to be emphasized can be changed for each operating state. Therefore, it is possible to obtain a more accurate estimation result ⁇ r by increasing the weighting ratio of the trained model that can exhibit high estimation accuracy for each operating state.
  • FIG. 10 is a flowchart showing a fourth example of a method of determining an estimation result by the estimation value determination unit 921.
  • the estimated value determination unit 921 first acquires a state variable representing the operating state of the main scanning mechanism 13 (step S51).
  • the state variable may be a measured value input to the measured value input unit 911 described above, a variable acquired by another sensor, or an operation set by the user for the control unit 40. It may be a mode or the like.
  • the estimated value determination unit 921 selects one of a plurality of tentative estimated values ⁇ 1, ⁇ 2, ⁇ 3, ... Based on the acquired state variables, and estimates the selected tentative estimated values ⁇ r. (Step S52). For example, in the operation state represented by a certain state variable, when the estimation accuracy of the trained model M2 is particularly high among the plurality of trained models M1, M2, M3, ... , The tentative estimated value ⁇ 2 output from the trained model M2 is used as the estimation result ⁇ r.
  • the estimation value determination unit 921 outputs the estimation result ⁇ r to the control unit 40 (step S26).
  • the control unit 40 corrects the yawing angle ⁇ of the stage 12 based on the estimation result ⁇ r of the yawing angle ⁇ output from the posture estimation unit 92 (step S27). Specifically, the control unit 40 operates the rotation mechanism 15 or adjusts the torque value of either of the pair of linear motors 61. As a result, the yawing angle ⁇ of the stage 12 is corrected so as to approach a desired value.
  • the transport device 10 estimates the yawing angle ⁇ of the stage 12 based on the measured value output from the main scanning mechanism 13 or the input data d which is a value calculated based on the measured value.
  • the yawing angle ⁇ of the stage 12 can be estimated without constantly installing the large-scale posture measuring device 80.
  • one estimation result ⁇ r is obtained based on a plurality of tentative estimation values ⁇ 1, ⁇ 2, ⁇ 3, ... Output from the plurality of trained models M1, M2, M3, ... Output. Therefore, high estimation accuracy can be realized in many operating conditions of the transport device 10. Further, in the present embodiment, a plurality of trained models M1, M2, M3, ... Are generated by different machine learning algorithms. Therefore, in an operating situation where it is difficult for a certain machine learning algorithm to exhibit high estimation accuracy, another machine learning algorithm can complement it. As a result, the yawing angle ⁇ of the stage 12 can be estimated accurately in more operating conditions.
  • the measurement value input unit 911 acquires the torque value of the linear motor 61 from the control board 63.
  • the measured value input unit 911 may acquire the torque value of the linear motor 61 by another method.
  • a torque sensor may be attached to each linear motor 61 of the straight-ahead mechanism 60, and a torque value may be acquired from the torque sensor.
  • the same input data d is input to all the trained models M1, M2, M3, ...
  • the input data generation unit 912 may perform different processing for each trained model on the measurement value input to the measurement value input unit 911. Then, different input data generated by different processes may be input to the plurality of trained models M1, M2, M3, .... In this way, more suitable input data can be input for each trained model.
  • the transport device 10 of the above embodiment includes not only the main scanning mechanism 13 but also the sub-scanning mechanism 14 and the rotation mechanism 15.
  • the present invention may be intended for a transfer device that does not include a sub-scanning mechanism 14 and a rotating mechanism 15.
  • the transport device 10 of the above embodiment was mounted on the drawing device 1.
  • the present invention may be intended for a transfer device mounted on a device other than the drawing device 1.
  • the transfer device may be mounted on a device that applies a processing liquid to a substrate held on a stage.
  • the transfer device may be mounted on a device that prints on a recording medium held on the stage.
  • the yawing angle ⁇ of the stage 12 is actually measured by the laser interferometer 82 in the pre-learning process.
  • the yawing angle ⁇ of the stage 12 may be actually measured by another method.
  • the yawing angle ⁇ of the stage 12 may be actually measured based on the image of the stage 12 acquired by the camera.
  • the straight-ahead mechanism 60 of the above-described embodiment has a linear motor 61.
  • a mechanism that converts the rotary motion output from the rotary motor into a linear motion by a ball screw may be used.

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Condensed Matter Physics & Semiconductors (AREA)
  • Manufacturing & Machinery (AREA)
  • Computer Hardware Design (AREA)
  • Microelectronics & Electronic Packaging (AREA)
  • Power Engineering (AREA)
  • Exposure And Positioning Against Photoresist Photosensitive Materials (AREA)
  • Container, Conveyance, Adherence, Positioning, Of Wafer (AREA)

Abstract

A stage orientation estimation device (18) comprises: an input data acquisition unit (91); and an orientation estimation unit (92) that estimates a yaw angle of a stage on the basis of input data. Due to this configuration, it is possible to estimate the yaw angle of the stage without the need for a large-scale measurement device to be installed at all times. Furthermore, the orientation estimation unit 92 determines one estimation result (θr) on the basis of a plurality of provisional estimation values (θ1, θ2, θ3,...) outputted from a plurality of trained models (M1, M2, M3,...). Due to this configuration, high estimation accuracy can be achieved when a conveyance device operates under various conditions.

Description

ステージ姿勢推定装置、搬送装置、およびステージ姿勢推定方法Stage attitude estimation device, transfer device, and stage attitude estimation method
　本発明は、平板状のステージを搬送機構により搬送する搬送装置において、ステージのヨーイング角度を推定する技術に関する。 The present invention relates to a technique for estimating the yawing angle of a stage in a transfer device that conveys a flat plate-shaped stage by a transfer mechanism.
　従来、平板状のステージを搬送しつつ、ステージに保持された基板に対して種々の処理を行う装置が知られている。例えば、特許文献１には、基板（Ｗ）を載置したステージ（１０）をステージ移動機構（２０）により移動させつつ、基板（Ｗ）の上面に露光パターンを描画する装置が記載されている。 Conventionally, there are known devices that perform various processes on a substrate held on a stage while transporting a flat plate-shaped stage. For example, Patent Document 1 describes a device that draws an exposure pattern on the upper surface of a substrate (W) while moving a stage (10) on which a substrate (W) is placed by a stage moving mechanism (20). ..
特開２０１６－７２４３４号公報Japanese Unexamined Patent Publication No. 2016-72434
　この種の装置に搭載されるステージの搬送装置は、一対の直進機構を備えている場合がある。具体的には、互いに平行に設けられた一対のリニアモータにより、ステージを所定の方向に搬送する機構が知られている。 The stage transfer device mounted on this type of device may be equipped with a pair of straight-ahead mechanisms. Specifically, there is known a mechanism for transporting a stage in a predetermined direction by a pair of linear motors provided in parallel with each other.
　当該搬送装置では、ステージを一定の姿勢で移動させるために、一対の直進機構を均等に動作させる必要がある。しかしながら、一対の直進機構の微小な駆動誤差、リニアモータのガイドの間隙内のエア圧変動、加工誤差等によって、ステージの鉛直軸周りの回転角度（いわゆる「ヨーイング角度」）が僅かに変動する場合がある。このようなヨーイング角度の変動が生じると、ステージに保持された基板に対して、精密な処理を行うことが困難となる。 In the transfer device, it is necessary to operate the pair of straight-moving mechanisms evenly in order to move the stage in a constant posture. However, when the rotation angle around the vertical axis of the stage (so-called "yaw angle") fluctuates slightly due to minute drive errors of the pair of straight-moving mechanisms, air pressure fluctuations in the gaps of the linear motor guides, machining errors, etc. There is. When such a variation in yawing angle occurs, it becomes difficult to perform precise processing on the substrate held on the stage.
　従来の搬送装置には、上記のヨーイング角度の変動を把握するために、大掛かりな計測装置が搭載されていた。そして、計測装置の計測結果に基づいて、搬送装置の動作が補正されていた。しかしながら、大掛かりな計測装置を搭載すると、搬送装置の小型化が困難となる。また、計測装置を搭載することにより、搬送装置の製造コストも上昇する。 The conventional transfer device was equipped with a large-scale measuring device in order to grasp the fluctuation of the yawing angle described above. Then, the operation of the transport device was corrected based on the measurement result of the measuring device. However, if a large-scale measuring device is mounted, it becomes difficult to miniaturize the transport device. In addition, by mounting the measuring device, the manufacturing cost of the transport device also increases.
　このような大掛かりな計測装置を常時設置することなく、ステージのヨーイング角度を把握するためには、例えば、機械学習を利用することが考えられる。具体的には、搬送装置から出力されるトルク値等の計測値と、ステージのヨーイング角度との関係を学習した学習済みモデルを用意し、当該学習済みモデルに計測値を入力して、ステージのヨーイング角度を出力することが考えられる。 In order to grasp the yawing angle of the stage without constantly installing such a large-scale measuring device, it is conceivable to use machine learning, for example. Specifically, a trained model in which the relationship between the measured value such as the torque value output from the transport device and the yawing angle of the stage is learned is prepared, and the measured value is input to the trained model to perform the stage. It is conceivable to output the yawing angle.
　しかしながら、学習済みモデルを生成するための機械学習アルゴリズムには、多くの種類がある。また、複数の機械学習アルゴリズムには、それぞれ得手・不得手があり、搬送装置の動作状況によって、各機械学習アルゴリズムの推定精度が変動する。このため、１つの機械学習アルゴリズムにより生成される１つの学習済みモデルのみに依存すると、搬送装置の動作状況によっては、精度のよい推定を行うことができない場合がある。 However, there are many types of machine learning algorithms for generating trained models. In addition, each of the plurality of machine learning algorithms has strengths and weaknesses, and the estimation accuracy of each machine learning algorithm varies depending on the operating condition of the transfer device. Therefore, depending on only one trained model generated by one machine learning algorithm, it may not be possible to perform accurate estimation depending on the operating condition of the transport device.
　本発明は、このような事情に鑑みなされたものであり、大掛かりな計測装置を常時設置することなく、ステージのヨーイング角度を推定でき、かつ、搬送装置の多くの動作状況において高い推定精度を実現できる技術を提供することを目的とする。 The present invention has been made in view of such circumstances, and can estimate the yawing angle of the stage without constantly installing a large-scale measuring device, and realizes high estimation accuracy in many operating conditions of the transport device. The purpose is to provide the technology that can be used.
　上記課題を解決するため、本願の第１発明は、平板状のステージを搬送機構により搬送する搬送装置において、前記ステージのヨーイング角度を推定するステージ姿勢推定装置であって、前記搬送機構から出力される計測値または前記計測値に基づいて算出される値を、入力データとして取得する入力データ取得部と、前記入力データに基づいて前記ステージのヨーイング角度を推定して、推定結果を出力する姿勢推定部と、を備え、前記姿勢推定部は、前記入力データに基づいて前記ステージのヨーイング角度の仮推定値を出力する複数の学習済みモデルと、前記複数の学習済みモデルから出力される複数の前記仮推定値に基づいて、前記推定結果を決定する推定値決定部と、を有する。 In order to solve the above problems, the first invention of the present application is a stage attitude estimation device that estimates the yawing angle of the stage in a transfer device that conveys a flat plate-shaped stage by a transfer mechanism, and is output from the transfer mechanism. An input data acquisition unit that acquires the measured value or a value calculated based on the measured value as input data, and an attitude estimation that estimates the yawing angle of the stage based on the input data and outputs the estimation result. The posture estimation unit includes a unit, a plurality of trained models that output a tentative estimated value of the yawing angle of the stage based on the input data, and a plurality of the trained models that are output from the plurality of trained models. It has an estimated value determining unit that determines the estimated result based on the tentative estimated value.
　本願の第２発明は、第１発明のステージ姿勢推定装置であって、前記複数の学習済みモデルは、互いに異なるアルゴリズムにより生成されたものである。 The second invention of the present application is the stage posture estimation device of the first invention, and the plurality of trained models are generated by algorithms different from each other.
　本願の第３発明は、第１発明または第２発明のステージ姿勢推定装置であって、前記推定値決定部は、前記複数の仮推定値の平均値を、前記推定結果とする。 The third invention of the present application is the stage posture estimation device of the first invention or the second invention, and the estimated value determining unit uses the average value of the plurality of tentative estimated values as the estimated result.
　本願の第４発明は、第１発明または第２発明のステージ姿勢推定装置であって、前記複数の学習済みモデルに、加重比率が設定されており、前記推定値決定部は、前記加重比率を用いた前記複数の仮推定値の加重平均値を、前記推定結果とする。 The fourth invention of the present application is the stage posture estimation device of the first invention or the second invention, in which weight ratios are set in the plurality of trained models, and the estimated value determining unit determines the weight ratios. The weighted average value of the plurality of tentative estimated values used is used as the estimated result.
　本願の第５発明は、第４発明のステージ姿勢推定装置であって、前記推定値決定部は、前記搬送機構の動作状態を表す状態変数に基づき、前記加重比率を変更する。 The fifth invention of the present application is the stage posture estimation device of the fourth invention, and the estimated value determining unit changes the weighting ratio based on a state variable representing an operating state of the transport mechanism.
　本願の第６発明は、第１発明または第２発明のステージ姿勢推定装置であって、前記推定値決定部は、前記搬送機構の動作状態を表す状態変数に基づき、前記複数の仮推定値のうちのいずれか１つを選択し、選択された前記仮推定値を、前記推定結果とする。 The sixth invention of the present application is the stage posture estimation device of the first invention or the second invention, and the estimated value determining unit is based on a state variable representing an operating state of the transport mechanism, and has a plurality of tentative estimated values. One of them is selected, and the selected tentative estimated value is used as the estimated result.
　本願の第７発明は、第１発明から第６発明までのいずれか１発明のステージ姿勢推定装置であって、前記搬送機構は、前記ステージを一対の直進機構により搬送し、前記入力データ取得部は、前記一対の直進機構のトルク値の差分に基づいて、前記入力データを生成する。 The seventh invention of the present application is the stage attitude estimation device of any one of the first to sixth inventions, wherein the transport mechanism transports the stage by a pair of straight-moving mechanisms, and the input data acquisition unit. Generates the input data based on the difference between the torque values of the pair of straight-ahead mechanisms.
　本願の第８発明は、搬送装置であって、第１発明から第７発明までのいずれか１発明のステージ姿勢推定装置と、前記ステージと、前記搬送機構と、を備える。 The eighth invention of the present application is a transport device, and includes the stage posture estimation device of any one of the first to seventh inventions, the stage, and the transport mechanism.
　本願の第９発明は、平板状のステージを搬送機構により搬送する搬送装置において、前記ステージのヨーイング角度を推定するステージ姿勢推定方法であって、ａ）前記搬送機構から出力される計測値または前記計測値に基づいて算出される値を、入力データとして取得する工程と、ｂ）前記入力データに基づいて前記ステージのヨーイング角度を推定して、推定結果を出力する工程と、を備え、前記工程ｂ）では、複数の学習済みモデルに前記入力データを入力し、前記複数の学習済みモデルから出力される複数の仮推定値に基づいて、前記推定結果を決定する。 The ninth invention of the present application is a stage attitude estimation method for estimating the yawing angle of the stage in a transport device that transports a flat plate-shaped stage by a transport mechanism, and a) a measured value output from the transport mechanism or the said. The step includes a step of acquiring a value calculated based on a measured value as input data, and b) a step of estimating a yawing angle of the stage based on the input data and outputting an estimation result. In b), the input data is input to the plurality of trained models, and the estimation result is determined based on the plurality of tentative estimation values output from the plurality of trained models.
　本願の第１発明から第９発明によれば、搬送機構から出力される計測値または計測値に基づいて算出される値に基づいて、ステージのヨーイング角度を推定する。これにより、大掛かりな計測装置を常時設置することなく、ステージのヨーイング角度を推定できる。また、複数の学習済みモデルから出力される複数の仮推定値に基づいて、１つの推定結果を出力する。これにより、搬送装置の多くの動作状況において高い推定精度を実現できる。 According to the first to ninth inventions of the present application, the yawing angle of the stage is estimated based on the measured value output from the transport mechanism or the value calculated based on the measured value. This makes it possible to estimate the yawing angle of the stage without constantly installing a large-scale measuring device. In addition, one estimation result is output based on a plurality of tentative estimates output from the plurality of trained models. As a result, high estimation accuracy can be realized in many operating conditions of the transport device.
搬送装置を備えた描画装置の斜視図である。It is a perspective view of the drawing apparatus provided with the transfer apparatus. 搬送装置を備えた描画装置の概略上面図である。It is a schematic top view of the drawing apparatus provided with the transfer apparatus. 制御部と描画装置内の各部との電気的接続を示したブロック図である。It is a block diagram which showed the electrical connection between a control part and each part in a drawing apparatus. 搬送装置の一部分を、主走査方向に対して垂直な面で切断したときの部分断面図である。It is a partial cross-sectional view when a part of a transport device is cut in a plane perpendicular to the main scanning direction. ステージ姿勢推定装置の構成を示したブロック図である。It is a block diagram which showed the structure of the stage posture estimation apparatus. 事前学習処理の流れを示すフローチャートである。It is a flowchart which shows the flow of the pre-learning process. ステージ姿勢推定装置による推定処理の流れを示したフローチャートである。It is a flowchart which showed the flow of the estimation process by a stage posture estimation apparatus. 推定値決定部による推定結果の決定方法の第２の例を示すフローチャートである。It is a flowchart which shows the 2nd example of the method of determining the estimation result by the estimation value determination part. 推定値決定部による推定結果の決定方法の第３の例を示すフローチャートである。It is a flowchart which shows the 3rd example of the method of determining the estimation result by the estimation value determination unit. 推定値決定部による推定結果の決定方法の第４の例を示すフローチャートである。It is a flowchart which shows the 4th example of the method of determining the estimation result by the estimation value determination unit.
　以下、本発明の実施形態について、図面を参照しつつ説明する。 Hereinafter, embodiments of the present invention will be described with reference to the drawings.
　なお、以下では、水平方向のうち、一対の直進機構によりステージが移動する方向を「主走査方向」と称し、主走査方向に直交する方向を「副走査方向」と称する。 In the following, among the horizontal directions, the direction in which the stage moves by the pair of straight-moving mechanisms is referred to as the "main scanning direction", and the direction orthogonal to the main scanning direction is referred to as the "secondary scanning direction".
　＜１．描画装置の構成＞
　図１は、本発明の一実施形態に係る搬送装置１０を備えた描画装置１の斜視図である。図２は、描画装置１の概略上面図である。この描画装置１は、感光材料が塗布された半導体基板またはガラス基板等の基板Ｗの上面に、空間変調された光を照射して、基板Ｗの上面に露光パターンを描画する装置である。図１および図２に示すように、描画装置１は、搬送装置１０、フレーム２０、描画処理部３０、および制御部４０を備える。 <1. Configuration of drawing device>
FIG. 1 is a perspective view of a drawing device 1 provided with a transfer device 10 according to an embodiment of the present invention. FIG. 2 is a schematic top view of the drawing apparatus 1. The drawing device 1 is a device that irradiates the upper surface of a substrate W such as a semiconductor substrate or a glass substrate coated with a photosensitive material with spatially modulated light to draw an exposure pattern on the upper surface of the substrate W. As shown in FIGS. 1 and 2, the drawing device 1 includes a transfer device 10, a frame 20, a drawing processing unit 30, and a control unit 40.
　搬送装置１０は、基台１１の上面において、平板状のステージ１２を、略一定の姿勢で水平方向に搬送する装置である。搬送装置１０は、主走査機構１３と副走査機構１４とを含む搬送機構を有する。主走査機構１３は、ステージ１２を主走査方向に搬送するための機構である。副走査機構１４は、ステージ１２を副走査方向に搬送するための機構である。基板Ｗは、ステージ１２の上面に水平姿勢で保持され、ステージ１２とともに主走査方向および副走査方向に移動する。 The transport device 10 is a device that transports the flat plate-shaped stage 12 in a substantially constant posture in the horizontal direction on the upper surface of the base 11. The transport device 10 has a transport mechanism including a main scanning mechanism 13 and a sub scanning mechanism 14. The main scanning mechanism 13 is a mechanism for transporting the stage 12 in the main scanning direction. The sub-scanning mechanism 14 is a mechanism for transporting the stage 12 in the sub-scanning direction. The substrate W is held on the upper surface of the stage 12 in a horizontal posture, and moves together with the stage 12 in the main scanning direction and the sub-scanning direction.
　搬送装置１０のより詳細な構成については、後述する。 A more detailed configuration of the transport device 10 will be described later.
　フレーム２０は、基台１１の上方において、描画処理部３０を保持するための構造である。フレーム２０は、一対の支柱部２１と架橋部２２とを有する。一対の支柱部２１は、副走査方向に間隔をあけて立設されている。各支柱部２１は、基台１１の上面から上方へ向けて延びる。架橋部２２は、２本の支柱部２１の上端部の間において、副走査方向に延びる。基板Ｗを保持したステージ１２は、一対の支柱部２１の間、かつ、架橋部２２の下方を通過する。 The frame 20 has a structure for holding the drawing processing unit 30 above the base 11. The frame 20 has a pair of strut portions 21 and a cross-linking portion 22. The pair of support columns 21 are erected at intervals in the sub-scanning direction. Each support column 21 extends upward from the upper surface of the base 11. The cross-linked portion 22 extends in the sub-scanning direction between the upper ends of the two strut portions 21. The stage 12 holding the substrate W passes between the pair of support columns 21 and below the cross-linked portion 22.
　描画処理部３０は、２つの光学ヘッド３１、照明光学系３２、レーザ発振器３３、およびレーザ駆動部３４を有する。２つの光学ヘッド３１は、副走査方向に間隔をあけて、架橋部２２に固定される。照明光学系３２、レーザ発振器３３、およびレーザ駆動部３４は、例えば、架橋部２２の内部空間に収容される。レーザ駆動部３４は、レーザ発振器３３と電気的に接続されている。レーザ駆動部３４を動作させると、レーザ発振器３３からパルス光が出射される。そして、レーザ発振器３３から出射されたパルス光が、照明光学系３２を介して光学ヘッド３１へ導入される。 The drawing processing unit 30 has two optical heads 31, an illumination optical system 32, a laser oscillator 33, and a laser drive unit 34. The two optical heads 31 are fixed to the cross-linked portion 22 at intervals in the sub-scanning direction. The illumination optical system 32, the laser oscillator 33, and the laser drive unit 34 are housed in, for example, the internal space of the cross-linking unit 22. The laser drive unit 34 is electrically connected to the laser oscillator 33. When the laser drive unit 34 is operated, pulsed light is emitted from the laser oscillator 33. Then, the pulsed light emitted from the laser oscillator 33 is introduced into the optical head 31 via the illumination optical system 32.
　光学ヘッド３１の内部には、空間変調器を含む光学系が設けられている。空間変調器には、例えば、回折格子型の空間光変調器であるＧＬＶ（Grating Light Valve）（登録商標）が用いられる。光学ヘッド３１へ導入されたパルス光は、空間変調器により所定のパターンに変調されて、基板Ｗの上面に照射される。これにより、基板Ｗの上面に塗布されたレジスト等の感光材料が露光される。 An optical system including a spatial modulator is provided inside the optical head 31. As the spatial modulator, for example, GLV (Grating Light Valve) (registered trademark), which is a diffraction grating type spatial light modulator, is used. The pulsed light introduced into the optical head 31 is modulated into a predetermined pattern by the spatial modulator and is irradiated on the upper surface of the substrate W. As a result, a photosensitive material such as a resist coated on the upper surface of the substrate W is exposed.
　描画装置１の稼働時には、光学ヘッド３１による露光と、搬送装置１０による基板Ｗの搬送とが、繰り返し実行される。具体的には、副走査機構１４によりステージ１２を副走査方向に搬送しつつ、光学ヘッド３１からのパルス光の照射を行うことにより、副走査方向に延びる帯状の領域（スワス）に露光を行った後、主走査機構１３によりステージ１２を主走査方向に１スワス分だけ搬送する。描画装置１は、このような副走査方向の露光と、主走査方向のステージ１２の搬送とを繰り返すことにより、基板Ｗの上面全体にパターンを描画する。 When the drawing device 1 is in operation, the exposure by the optical head 31 and the transfer of the substrate W by the transfer device 10 are repeatedly executed. Specifically, the sub-scanning mechanism 14 conveys the stage 12 in the sub-scanning direction, and irradiates pulsed light from the optical head 31 to expose a band-shaped region (swath) extending in the sub-scanning direction. After that, the main scanning mechanism 13 conveys the stage 12 in the main scanning direction by one swath. The drawing apparatus 1 draws a pattern on the entire upper surface of the substrate W by repeating such exposure in the sub-scanning direction and transfer of the stage 12 in the main scanning direction.
　制御部４０は、描画装置１の各部を動作制御するための手段である。図３は、制御部４０と、描画装置１内の各部との電気的接続を示したブロック図である。図３中に概念的に示すように、制御部４０は、ＣＰＵ等のプロセッサ４１、ＲＡＭ等のメモリ４２、およびハードディスクドライブ等の記憶部４３を有するコンピュータにより構成されている。記憶部４３には、描画装置１を動作制御するためのコンピュータプログラムＰが記憶されている。 The control unit 40 is a means for controlling the operation of each unit of the drawing device 1. FIG. 3 is a block diagram showing an electrical connection between the control unit 40 and each unit in the drawing device 1. As conceptually shown in FIG. 3, the control unit 40 is composed of a computer having a processor 41 such as a CPU, a memory 42 such as a RAM, and a storage unit 43 such as a hard disk drive. The storage unit 43 stores a computer program P for controlling the operation of the drawing device 1.
　また、図３に示すように、制御部４０は、描画処理部３０（上述した光学ヘッド３１およびレーザ駆動部３４を含む）、主走査機構１３（後述するリニアモータ６１およびエアガイド６２を含む）、副走査機構１４（後述するリニアモータ７１を含む）、後述する回転機構１５、および各種センサ５０と、電気的に接続されている。制御部４０は、記憶部４３に記憶されたコンピュータプログラムＰおよびデータＤをメモリ４２に読み出し、当該コンピュータプログラムＰおよびデータＤに基づいて、プロセッサ４１が演算処理を行うことにより、描画装置１内の上記各部を動作制御する。これにより、描画装置１における描画処理が進行する。 Further, as shown in FIG. 3, the control unit 40 includes a drawing processing unit 30 (including the optical head 31 and the laser driving unit 34 described above) and a main scanning mechanism 13 (including a linear motor 61 and an air guide 62 described later). , A sub-scanning mechanism 14 (including a linear motor 71 described later), a rotating mechanism 15 described later, and various sensors 50 are electrically connected. The control unit 40 reads the computer program P and data D stored in the storage unit 43 into the memory 42, and the processor 41 performs arithmetic processing based on the computer program P and data D in the drawing device 1. The operation of each of the above parts is controlled. As a result, the drawing process in the drawing device 1 proceeds.
　＜２．搬送装置の構成＞
　次に、搬送装置１０の詳細な構成について、説明する。図４は、搬送装置１０の一部分を、主走査方向に対して垂直な面で切断したときの部分断面図である。図１～図４に示すように、搬送装置１０は、基台１１、ステージ１２、主走査機構１３、副走査機構１４、回転機構１５、下段支持プレート１６、中段支持プレート１７、およびステージ姿勢推定装置１８を有する。 <2. Conveyor device configuration>
Next, the detailed configuration of the transport device 10 will be described. FIG. 4 is a partial cross-sectional view of a portion of the transport device 10 cut along a plane perpendicular to the main scanning direction. As shown in FIGS. 1 to 4, the transport device 10 includes a base 11, a stage 12, a main scanning mechanism 13, a sub-scanning mechanism 14, a rotating mechanism 15, a lower support plate 16, a middle support plate 17, and a stage posture estimation. It has a device 18.
　基台１１は、搬送装置１０の各部を支持する支持台である。基台１１は、主走査方向および副走査方向に広がる平板状の外形を有する。基台１１の下面には、４つの脚部１１１および２つのダンパ１１２が設けられている。脚部１１１およびダンパ１１２の長さは、個別に調節可能である。したがって、脚部１１１およびダンパ１１２の長さを調節することにより、基台１１の姿勢を水平に調整できる。 The base 11 is a support base that supports each part of the transport device 10. The base 11 has a flat plate-like outer shape that extends in the main scanning direction and the sub-scanning direction. Four legs 111 and two dampers 112 are provided on the lower surface of the base 11. The lengths of the legs 111 and the damper 112 can be adjusted individually. Therefore, the posture of the base 11 can be adjusted horizontally by adjusting the lengths of the legs 111 and the damper 112.
　下段支持プレート１６、中段支持プレート１７、およびステージ１２は、それぞれ、平板状の外形を有する。下段支持プレート１６は、基台１１上において、主走査機構１３により主走査方向に移動可能に支持されている。中段支持プレート１７は、下段支持プレート１６上において、副走査機構１４により副走査方向に移動可能に支持されている。ステージ１２は、中段支持プレート１７上において、回転機構１５により鉛直軸周りに回転可能に支持されている。ステージ１２は、基板Ｗを載置可能な上面を有する。また、ステージ１２の上面には、基板Ｗを保持するためのチャックピンまたは基板Ｗを吸着する複数の吸着孔が設けられている。 The lower support plate 16, the middle support plate 17, and the stage 12 each have a flat outer shape. The lower support plate 16 is movably supported in the main scanning direction by the main scanning mechanism 13 on the base 11. The middle-stage support plate 17 is movably supported on the lower-stage support plate 16 by the sub-scanning mechanism 14 in the sub-scanning direction. The stage 12 is rotatably supported around a vertical axis by a rotation mechanism 15 on the middle support plate 17. The stage 12 has an upper surface on which the substrate W can be placed. Further, on the upper surface of the stage 12, a chuck pin for holding the substrate W or a plurality of suction holes for sucking the substrate W are provided.
　主走査機構１３は、基台１１に対して下段支持プレート１６を、主走査方向に移動させる機構である。主走査機構１３は、一対の直進機構６０を有する。一対の直進機構６０は、基台１１の上面の副走査方向の両端部に設けられている。図２および図４に示すように、一対の直進機構６０は、それぞれ、リニアモータ６１とエアガイド６２とを有する。 The main scanning mechanism 13 is a mechanism for moving the lower support plate 16 with respect to the base 11 in the main scanning direction. The main scanning mechanism 13 has a pair of straight traveling mechanisms 60. The pair of straight-moving mechanisms 60 are provided at both ends of the upper surface of the base 11 in the sub-scanning direction. As shown in FIGS. 2 and 4, each of the pair of straight-moving mechanisms 60 has a linear motor 61 and an air guide 62, respectively.
　リニアモータ６１は、固定子６１１および移動子６１２を有する。固定子６１１は、基台１１の上面に、主走査方向に沿って敷設されている。すなわち、一対の固定子６１１は、互いに平行に配置されている。移動子６１２は、後述するエアベアリング６２２を介して、下段支持プレート１６に固定されている。 The linear motor 61 has a stator 611 and a mover 612. The stator 611 is laid on the upper surface of the base 11 along the main scanning direction. That is, the pair of stators 611 are arranged parallel to each other. The mover 612 is fixed to the lower support plate 16 via an air bearing 622 described later.
　また、主走査機構１３は、リニアモータ６１の動作を制御するための制御基板６３を有する。制御基板６３には、例えばサーボパック（登録商標）が用いられる。制御基板６３は、制御部４０と電気的に接続されている。リニアモータ６１の駆動時には、制御部４０からの指令に従って、制御基板６３が、リニアモータ６１において発生させるべきトルクを算出する。そして、算出されたトルクに応じた駆動信号を、各リニアモータ６１の固定子６１１へ供給する。そうすると、固定子６１１と移動子６１２との間に生じる磁気的な吸引力および反発力によって、移動子６１２が、固定子６１１に沿って主走査方向に移動する。 Further, the main scanning mechanism 13 has a control board 63 for controlling the operation of the linear motor 61. For the control board 63, for example, a servo pack (registered trademark) is used. The control board 63 is electrically connected to the control unit 40. When the linear motor 61 is driven, the control board 63 calculates the torque to be generated in the linear motor 61 according to the command from the control unit 40. Then, a drive signal corresponding to the calculated torque is supplied to the stator 611 of each linear motor 61. Then, the mover 612 moves in the main scanning direction along the stator 611 due to the magnetic attraction and repulsion generated between the stator 611 and the mover 612.
　エアガイド６２は、ガイドレール６２１およびエアベアリング６２２を有する。ガイドレール６２１は、基台１１の上面に、主走査方向に沿って敷設されている。すなわち、リニアモータ６１の固定子６１１と、エアガイド６２のガイドレール６２１とは、互いに平行に配置されている。エアベアリング６２２は、下段支持プレート１６と移動子６１２とに、固定されている。また、エアベアリング６２２は、ガイドレール６２１の上方に配置されている。 The air guide 62 has a guide rail 621 and an air bearing 622. The guide rail 621 is laid on the upper surface of the base 11 along the main scanning direction. That is, the stator 611 of the linear motor 61 and the guide rail 621 of the air guide 62 are arranged in parallel with each other. The air bearing 622 is fixed to the lower support plate 16 and the mover 612. Further, the air bearing 622 is arranged above the guide rail 621.
　図４に示すように、エアベアリング６２２の下面には、気体吹出口６２３が設けられている。搬送装置１０の稼働時には、工場のユーティリティからエアベアリング６２２に常に気体が供給され、気体吹出口６２３からガイドレール６２１の上面に向けて、加圧された気体が吹き出される。これにより、エアベアリング６２２は、ガイドレール６２１上に非接触で浮上支持される。したがって、リニアモータ６１を駆動させると、下段支持プレート１６は、エアガイド６２により浮上支持された状態で、主走査方向に沿って低摩擦で滑らかに移動する。 As shown in FIG. 4, a gas outlet 623 is provided on the lower surface of the air bearing 622. When the transfer device 10 is in operation, gas is constantly supplied from the utility of the factory to the air bearing 622, and the pressurized gas is blown out from the gas outlet 623 toward the upper surface of the guide rail 621. As a result, the air bearing 622 is floated and supported on the guide rail 621 in a non-contact manner. Therefore, when the linear motor 61 is driven, the lower support plate 16 moves smoothly along the main scanning direction with low friction in a state of being floated and supported by the air guide 62.
　副走査機構１４は、下段支持プレート１６に対して中段支持プレート１７を、副走査方向に移動させる機構である。副走査機構１４は、リニアモータ７１と、一対のガイド機構７２とを有する。 The sub-scanning mechanism 14 is a mechanism for moving the middle-stage support plate 17 with respect to the lower-stage support plate 16 in the sub-scanning direction. The sub-scanning mechanism 14 includes a linear motor 71 and a pair of guide mechanisms 72.
　リニアモータ７１は、下段支持プレート１６の上面の主走査方向の略中央に設けられている。リニアモータ７１は、固定子７１１および移動子７１２を有する。固定子７１１は、下段支持プレート１６の上面に、副走査方向に沿って敷設されている。移動子７１２は、中段支持プレート１７に対して固定されている。リニアモータ７１の駆動時には、固定子７１１と移動子７１２との間に生じる磁気的な吸引力および反発力により、移動子７１２が、固定子７１１に沿って副走査方向に移動する。 The linear motor 71 is provided substantially in the center of the upper surface of the lower support plate 16 in the main scanning direction. The linear motor 71 has a stator 711 and a mover 712. The stator 711 is laid on the upper surface of the lower support plate 16 along the sub-scanning direction. The mover 712 is fixed to the middle support plate 17. When the linear motor 71 is driven, the mover 712 moves in the sub-scanning direction along the stator 711 due to the magnetic attraction and repulsion generated between the stator 711 and the mover 712.
　一対のガイド機構７２は、下段支持プレート１６の上面の主走査方向の両端部に設けられている。一対のガイド機構７２は、それぞれ、ガイドレール７２１とボールベアリング７２２とを有する。ガイドレール７２１は、下段支持プレート１６の上面に、副走査方向に沿って敷設されている。ボールベアリング７２２は、中段支持プレート１７の下面に固定されている。また、ボールベアリング７２２は、ガイドレール７２１に沿って、副走査方向に移動可能である。したがって、リニアモータ７１を駆動させると、中段支持プレート１７は、下段支持プレート１６に対して、副走査方向に移動する。 The pair of guide mechanisms 72 are provided at both ends of the upper surface of the lower support plate 16 in the main scanning direction. The pair of guide mechanisms 72 each have a guide rail 721 and a ball bearing 722. The guide rail 721 is laid on the upper surface of the lower support plate 16 along the sub-scanning direction. The ball bearing 722 is fixed to the lower surface of the middle support plate 17. Further, the ball bearing 722 can move in the sub-scanning direction along the guide rail 721. Therefore, when the linear motor 71 is driven, the middle support plate 17 moves in the sub-scanning direction with respect to the lower support plate 16.
　回転機構１５は、中段支持プレート１７に対するステージ１２の鉛直軸周りの角度を調整する機構である。回転機構１５には、例えばモータが用いられる。モータを動作させると、中段支持プレート１７に対してステージ１２が、鉛直軸周りに回転する。これにより、ステージ１２の鉛直軸周りの角度（ヨーイング角度）θを調整することができる。 The rotation mechanism 15 is a mechanism for adjusting the angle around the vertical axis of the stage 12 with respect to the middle support plate 17. For example, a motor is used for the rotation mechanism 15. When the motor is operated, the stage 12 rotates about the vertical axis with respect to the middle support plate 17. Thereby, the angle (yaw angle) θ around the vertical axis of the stage 12 can be adjusted.
　このように、ステージ１２は、主走査機構１３、副走査機構１４、および回転機構１５により、基台１１に対して、主走査方向および副走査方向に移動可能であるとともに、ヨーイング角度θを調整することが可能となっている。 In this way, the stage 12 can be moved in the main scanning direction and the sub-scanning direction with respect to the base 11 by the main scanning mechanism 13, the sub-scanning mechanism 14, and the rotation mechanism 15, and the yawing angle θ is adjusted. It is possible to do.
　図１および図２に示すように、この搬送装置１０には、姿勢計測装置８０を設置することが可能である。姿勢計測装置８０は、ステージ１２のヨーイング角度θを計測するための装置である。姿勢計測装置８０は、ステージ１２に固定されるミラー８１と、レーザ干渉計８２とを有する。ミラー８１は、ステージ１２の主走査方向の端縁部に固定される。レーザ干渉計８２は、基台１１の上面に固定される。レーザ干渉計８２は、ミラー８１へ向けて２本のレーザ光を照射する。そして、ミラー８１から反射する２本のレーザ光の干渉により、２本のレーザ光の光路差を検出する。そして、当該光路差に基づいて、ステージ１２のヨーイング角度θを計測する。 As shown in FIGS. 1 and 2, the posture measuring device 80 can be installed in the transport device 10. The posture measuring device 80 is a device for measuring the yawing angle θ of the stage 12. The posture measuring device 80 includes a mirror 81 fixed to the stage 12 and a laser interferometer 82. The mirror 81 is fixed to the edge portion of the stage 12 in the main scanning direction. The laser interferometer 82 is fixed to the upper surface of the base 11. The laser interferometer 82 irradiates the mirror 81 with two laser beams. Then, the optical path difference between the two laser beams is detected by the interference of the two laser beams reflected from the mirror 81. Then, the yawing angle θ of the stage 12 is measured based on the optical path difference.
　姿勢計測装置８０は、後述する事前学習処理を行う場合に設置される。事前学習が完了した後は、姿勢計測装置８０を取り外して、搬送装置１０を使用することができる。 The posture measuring device 80 is installed when performing the pre-learning process described later. After the pre-learning is completed, the posture measuring device 80 can be removed and the transport device 10 can be used.
　＜３．ステージ姿勢推定装置について＞
　＜３－１．ステージ姿勢推定装置の構成＞
　続いて、搬送装置１０に搭載されるステージ姿勢推定装置１８について、説明する。図５は、ステージ姿勢推定装置１８の構成を示したブロック図である。ステージ姿勢推定装置１８は、主走査機構１３から出力される計測値に基づいて、ステージ１２のヨーイング角度θを推定する装置である。図５に示すように、ステージ姿勢推定装置１８は、入力データ取得部９１と、姿勢推定部９２とを備える。入力データ取得部９１は、計測値入力部９１１と、入力データ生成部９１２とを有する。姿勢推定部９２は、複数の学習済みモデルＭ１，Ｍ２，Ｍ３，・・・と、推定値決定部９２１とを有する。 <3. About stage posture estimation device ＞
<3-1. Configuration of stage posture estimation device>
Subsequently, the stage posture estimation device 18 mounted on the transport device 10 will be described. FIG. 5 is a block diagram showing the configuration of the stage posture estimation device 18. The stage posture estimation device 18 is a device that estimates the yawing angle θ of the stage 12 based on the measured values output from the main scanning mechanism 13. As shown in FIG. 5, the stage posture estimation device 18 includes an input data acquisition unit 91 and a posture estimation unit 92. The input data acquisition unit 91 has a measurement value input unit 911 and an input data generation unit 912. The posture estimation unit 92 has a plurality of trained models M1, M2, M3, ..., And an estimation value determination unit 921.
　ステージ姿勢推定装置１８は、ＣＰＵ等のプロセッサ、ＲＡＭ等のメモリ、およびハードディスクドライブ等の記憶部を有するコンピュータにより構成される。計測値入力部９１１、入力データ生成部９１２、および推定値決定部９２１の各機能は、記憶部に記憶されたコンピュータプログラムに従ってプロセッサが動作することにより、実現される。 The stage attitude estimation device 18 is composed of a computer having a processor such as a CPU, a memory such as a RAM, and a storage unit such as a hard disk drive. Each function of the measurement value input unit 911, the input data generation unit 912, and the estimation value determination unit 921 is realized by operating the processor according to the computer program stored in the storage unit.
　学習済みモデルＭ１，Ｍ２，Ｍ３，・・・は、機械学習アルゴリズムを用いた事前学習により、パラメータが調整された推論プログラムである。学習済みモデルＭ１，Ｍ２，Ｍ３，・・・を得るための機械学習アルゴリズムとしては、例えば、一層ニューラルネットワークやディープラーニング等を含むニューラルネットワーク、ランダムフォレストや勾配ブースティング等を含む決定木系アルゴリズム、サポートベクトルマシンといった、いわゆる教師あり機械学習アルゴリズムが用られる。複数の学習済みモデルＭ１，Ｍ２，Ｍ３，・・・は、それぞれ、異なる機械学習アルゴリズムにより生成されたものである。 The trained models M1, M2, M3, ... Are inference programs whose parameters have been adjusted by pre-learning using a machine learning algorithm. Machine learning algorithms for obtaining trained models M1, M2, M3, ... include, for example, a neural network including a layer neural network and deep learning, a decision tree algorithm including a random forest, gradient boosting, and the like. So-called supervised machine learning algorithms, such as support vector machines, are used. The plurality of trained models M1, M2, M3, ... Are generated by different machine learning algorithms, respectively.
　なお、ステージ姿勢推定装置１８は、上述した制御部４０と同一のコンピュータにより構成されていてもよく、制御部４０とは別のコンピュータにより構成されていてもよい。 The stage posture estimation device 18 may be configured by the same computer as the control unit 40 described above, or may be configured by a computer different from the control unit 40.
　＜３－２．ステージ姿勢推定装置の事前学習＞
　続いて、ステージ姿勢推定装置１８において予め実行される事前学習処理について、説明する。図６は、事前学習処理の流れを示すフローチャートである。 <3-2. Pre-learning of stage posture estimation device ＞
Next, the pre-learning process executed in advance in the stage posture estimation device 18 will be described. FIG. 6 is a flowchart showing the flow of the pre-learning process.
　事前学習処理を行うときには、搬送装置１０に、上述した姿勢計測装置８０を設置する（ステップＳ１１）。そして、主走査機構１３を動作させつつ、次のステップＳ１２～Ｓ１５を繰り返し実行する。 When performing the pre-learning process, the posture measuring device 80 described above is installed in the transport device 10 (step S11). Then, the next steps S12 to S15 are repeatedly executed while operating the main scanning mechanism 13.
　まず、計測値入力部９１１に、制御基板６３から出力される計測値が入力される（ステップＳ１２）。計測値は、例えば、主走査機構１３の一対のリニアモータ６１のトルク値とされる。ただし、計測値入力部９１１へ入力される計測値は、エアベアリング６２２の空気圧、ガイドレール６２１の温度、搬送装置１０の駆動音、ステージ１２の振動、ステージ１２の位置などの他の項目を、各種センサ５０により計測したものであってもよい。 First, the measured value output from the control board 63 is input to the measured value input unit 911 (step S12). The measured value is, for example, the torque value of the pair of linear motors 61 of the main scanning mechanism 13. However, the measured value input to the measured value input unit 911 includes other items such as the air pressure of the air bearing 622, the temperature of the guide rail 621, the driving sound of the transport device 10, the vibration of the stage 12, and the position of the stage 12. It may be measured by various sensors 50.
　次に、入力データ生成部９１２が、計測値入力部９１１に入力された計測値に基づいて、入力データｄを生成する（ステップＳ１３）。例えば、計測値が一対のリニアモータ６１のトルク値である場合には、入力データ生成部９１２は、それらのトルク値の差分を算出する。そして、算出された差分の時系列データから、不要な周波数を除去することにより、入力データｄを生成する。計測値入力部９１１に入力される計測値が、トルク値以外の場合にも、入力データ生成部９１２は、所定の演算およびフィルタ処理を行うことにより、機械学習に適した入力データｄを生成する。 Next, the input data generation unit 912 generates the input data d based on the measurement value input to the measurement value input unit 911 (step S13). For example, when the measured values are the torque values of the pair of linear motors 61, the input data generation unit 912 calculates the difference between the torque values. Then, the input data d is generated by removing unnecessary frequencies from the calculated difference time series data. Even when the measured value input to the measured value input unit 911 is other than the torque value, the input data generation unit 912 generates input data d suitable for machine learning by performing a predetermined calculation and filtering process. ..
　なお、入力データ生成部９１２は、計測値入力部９１１に入力された計測値そのものを、入力データｄとしてもよい。すなわち、入力データ生成部９１２は、搬送機構から出力される計測値、または、計測値に所定の演算およびフィルタ処理を行うことにより算出される値を、入力データｄとすればよい。 Note that the input data generation unit 912 may use the measurement value itself input to the measurement value input unit 911 as the input data d. That is, the input data generation unit 912 may use the measured value output from the transport mechanism or the value calculated by performing a predetermined calculation and the filtering process on the measured value as the input data d.
　続いて、姿勢推定部９２は、入力データ生成部９１２により生成された入力データｄを入力とし、姿勢計測装置８０の計測結果θｍを教師データとして、機械学習を行う（ステップＳ１４）。すなわち、姿勢推定部９２は、上述した機械学習アルゴリズムにより、入力データｄと、姿勢計測装置８０の計測結果θｍとの関係を学習する。本実施形態の姿勢推定部９２は、このステップＳ１４の機械学習を、複数の異なる機械学習アルゴリズムにより、並行して行う。したがって、ステップＳ１４の機械学習により、複数の異なる学習済みモデルＭ１，Ｍ２，Ｍ３，・・・が生成される。 Subsequently, the posture estimation unit 92 performs machine learning using the input data d generated by the input data generation unit 912 as input and the measurement result θm of the posture measuring device 80 as teacher data (step S14). That is, the posture estimation unit 92 learns the relationship between the input data d and the measurement result θm of the posture measuring device 80 by the machine learning algorithm described above. The posture estimation unit 92 of the present embodiment performs the machine learning of step S14 in parallel by a plurality of different machine learning algorithms. Therefore, the machine learning in step S14 generates a plurality of different trained models M1, M2, M3, ...
　姿勢推定部９２は、ステップＳ１４の機械学習により生成される学習済みモデルＭ１，Ｍ２，Ｍ３，・・・の出力値と、教師データである計測結果θｍとを比較する。そして、姿勢推定部９２は、学習済みモデルＭ１，Ｍ２，Ｍ３，・・・の出力値と、計測結果θｍとの差分が、予め設定された閾値以下でない場合には、学習済みモデルＭ１，Ｍ２，Ｍ３，・・・の推定精度が、所望のレベルに達していないと判断する（ステップＳ１５：ｎｏ）。この場合、ステージ姿勢推定装置１８は、上述したステップＳ１２～Ｓ１４の処理を繰り返す。このように、機械学習を繰り返すことにより、学習済みモデルＭ１，Ｍ２，Ｍ３，・・・の推定精度が、次第に向上する。 The posture estimation unit 92 compares the output values of the trained models M1, M2, M3, ... Generated by the machine learning in step S14 with the measurement result θm which is the teacher data. Then, when the difference between the output values of the trained models M1, M2, M3, ... And the measurement result θm is not equal to or less than the preset threshold value, the posture estimation unit 92 determines the trained models M1, M2. , M3, ... It is determined that the estimation accuracy has not reached a desired level (step S15: no). In this case, the stage posture estimation device 18 repeats the processes of steps S12 to S14 described above. By repeating machine learning in this way, the estimation accuracy of the trained models M1, M2, M3, ... Is gradually improved.
　やがて、学習済みモデルＭ１，Ｍ２，Ｍ３，・・・の出力値と、計測結果θｍとの差分が、予め設定された閾値以下となった場合、姿勢推定部９２は、各学習済みモデルＭ１，Ｍ２，Ｍ３，・・・の推定精度が所望のレベルに達したと判断する。この場合、姿勢推定部９２は、機械学習を終了する（ステップＳ１５：ｙｅｓ）。そして、搬送装置１０から姿勢計測装置８０が取り外される（ステップＳ１６）。なお、ステージ姿勢推定装置１８は、ステップＳ１２～Ｓ１４の繰り返し回数が、予め設定された上限値に達した場合に、ステップＳ１５において、機械学習を終了してもよい。 Eventually, when the difference between the output values of the trained models M1, M2, M3, ... And the measurement result θm becomes equal to or less than a preset threshold value, the posture estimation unit 92 causes each trained model M1, It is determined that the estimation accuracy of M2, M3, ... Has reached a desired level. In this case, the posture estimation unit 92 ends machine learning (step S15: yes). Then, the posture measuring device 80 is removed from the transport device 10 (step S16). The stage posture estimation device 18 may end machine learning in step S15 when the number of repetitions of steps S12 to S14 reaches a preset upper limit value.
　＜３－３．ステージ姿勢推定装置の推定処理＞
　続いて、ステージ姿勢推定装置１８によるヨーイング角度θの推定処理について、説明する。この推定処理は、上述した事前学習処理が完了した後、搬送装置１０を動作させるときに実行される。図７は、推定処理の流れを示したフローチャートである。 <3-3. Estimating process of stage attitude estimation device>
Subsequently, the estimation process of the yawing angle θ by the stage posture estimation device 18 will be described. This estimation process is executed when the transfer device 10 is operated after the above-mentioned pre-learning process is completed. FIG. 7 is a flowchart showing the flow of the estimation process.
　推定処理を行うときには、まず、計測値入力部９１１に、制御基板６３から出力される計測値が入力される（ステップＳ２１）。ここでは、上述したステップＳ１２と同種の計測値が入力される。例えば、ステップＳ１２において入力される計測値が、一対のリニアモータ６１のトルク値である場合には、ステップＳ２１において入力される計測値も、一対のリニアモータ６１のトルク値とされる。 When performing the estimation process, first, the measured value output from the control board 63 is input to the measured value input unit 911 (step S21). Here, the same type of measurement value as in step S12 described above is input. For example, when the measured value input in step S12 is the torque value of the pair of linear motors 61, the measured value input in step S21 is also the torque value of the pair of linear motors 61.
　次に、入力データ生成部９１２が、計測値入力部９１１に入力された計測値に基づいて、入力データｄを生成する（ステップＳ２２）。ここでは、上述したステップＳ１３と同一の処理が実行される。すなわち、ステップＳ１３において実行される処理が、差分の算出およびフィルタ処理である場合には、ステップＳ２２においても、差分の算出およびフィルタ処理を実行することにより、入力データｄが生成される。 Next, the input data generation unit 912 generates the input data d based on the measurement value input to the measurement value input unit 911 (step S22). Here, the same process as in step S13 described above is executed. That is, when the process executed in step S13 is the difference calculation and the filter process, the input data d is generated by executing the difference calculation and the filter process in the step S22 as well.
　続いて、姿勢推定部９２は、生成された入力データｄを、複数の学習済みモデルＭ１，Ｍ２，Ｍ３，・・・に、それぞれ入力する（ステップＳ２３）。そうすると、各学習済みモデルＭ１，Ｍ２，Ｍ３，・・・は、入力データｄに対応するステージ１２のヨーイング角度θの仮推定値θ１，θ２，θ３，・・・を出力する。これにより、１つの入力データｄに対して、複数の仮推定値θ１，θ２，θ３，・・・が得られる（ステップＳ２４）。 Subsequently, the posture estimation unit 92 inputs the generated input data d into the plurality of trained models M1, M2, M3, ... (Step S23). Then, the trained models M1, M2, M3, ... Output the tentative estimated values θ1, θ2, θ3, ... Of the yawing angle θ of the stage 12 corresponding to the input data d. As a result, a plurality of tentative estimated values θ1, θ2, θ3, ... Are obtained for one input data d (step S24).
　その後、姿勢推定部９２の推定値決定部９２１が、複数の仮推定値θ１，θ２，θ３，・・・に基づいて、１つの推定結果θｒを決定する（ステップＳ２５）。具体的には、例えば、推定値決定部９２１は、複数の仮推定値θ１，θ２，θ３，・・・の平均値を算出し、算出された平均値を推定結果θｒとして決定する。ただし、推定値決定部９２１は、他の方法で、推定結果θｒを決定してもよい。 After that, the estimation value determination unit 921 of the posture estimation unit 92 determines one estimation result θr based on the plurality of provisional estimation values θ1, θ2, θ3, ... (Step S25). Specifically, for example, the estimation value determination unit 921 calculates the average value of a plurality of provisional estimation values θ1, θ2, θ3, ..., And determines the calculated average value as the estimation result θr. However, the estimated value determination unit 921 may determine the estimated result θr by another method.
　図８は、推定値決定部９２１による推定結果の決定方法の第２の例を示すフローチャートである。図８の例では、学習済みモデルＭ１，Ｍ２，Ｍ３，・・・のそれぞれに対して、予め加重比率ｗ１，ｗ２，ｗ３，・・・が設定されている。加重比率ｗ１，ｗ２，ｗ３，・・・は、ステージ姿勢推定装置１８を構成するコンピュータの記憶部に、予め記憶されている。推定値決定部９２１は、まず、記憶部から加重比率ｗ１，ｗ２，ｗ３，・・・を読み出す（ステップＳ３１）。そして、推定値決定部９２１は、読み出した加重比率ｗ１，ｗ２，ｗ３，・・・を用いて、複数の仮推定値θ１，θ２，θ３，・・・の加重平均値を算出し、算出された加重平均値を、推定結果θｒとする（ステップＳ３２）。 FIG. 8 is a flowchart showing a second example of a method of determining the estimation result by the estimation value determination unit 921. In the example of FIG. 8, the weighting ratios w1, w2, w3, ... Are set in advance for each of the trained models M1, M2, M3, ... The weighting ratios w1, w2, w3, ... Are stored in advance in the storage unit of the computer constituting the stage posture estimation device 18. The estimated value determining unit 921 first reads the weighted ratios w1, w2, w3, ... From the storage unit (step S31). Then, the estimated value determination unit 921 calculates and calculates the weighted average value of a plurality of tentative estimated values θ1, θ2, θ3, ... Using the read weighted ratios w1, w2, w3, ... Let the weighted average value be the estimation result θr (step S32).
　例えば、３つの学習済みモデルＭ１，Ｍ２，Ｍ３を使用する場合、ステップＳ３２において、加重平均を用いた推定結果θｒは、次の式（１）により算出することができる。
　θｒ＝（ｗ１・θ１＋ｗ２・θ２＋ｗ３・θ３）／（ｗ１＋ｗ２＋ｗ３）　　（１） For example, when three trained models M1, M2, and M3 are used, the estimation result θr using the weighted average in step S32 can be calculated by the following equation (1).
θr = (w1, θ1 + w2, θ2 + w3, θ3) / (w1 + w2 + w3) (1)
　複数の学習済みモデルＭ１，Ｍ２，Ｍ３，・・・の中で、特に推定精度の高い学習済みモデルまたは重視すべき学習済みモデルが存在する場合、その学習済みモデルの加重比率を、相対的に高く設定しておくとよい。そうすれば、上記のように加重平均値を算出することで、より好ましい推定結果θｒを得ることができる。 If there is a trained model with particularly high estimation accuracy or a trained model that should be emphasized among a plurality of trained models M1, M2, M3, ..., The weighting ratio of the trained model is relatively set. It is good to set it high. Then, by calculating the weighted average value as described above, a more preferable estimation result θr can be obtained.
　図９は、推定値決定部９２１による推定結果の決定方法の第３の例を示すフローチャートである。図９の例では、学習済みモデルＭ１，Ｍ２，Ｍ３，・・・のそれぞれに対応する加重比率ｗ１，ｗ２，ｗ３，・・・が、固定値ではなく、主走査機構１３の動作状態に応じて変化する。推定値決定部９２１は、まず、主走査機構１３の動作状態を表す状態変数を取得する（ステップＳ４１）。状態変数は、上述した計測値入力部９１１に入力される計測値であってもよく、他のセンサにより取得される変数であってもよく、あるいは、制御部４０に対してユーザが設定する動作モードなどであってもよい。 FIG. 9 is a flowchart showing a third example of a method of determining the estimation result by the estimation value determination unit 921. In the example of FIG. 9, the weighting ratios w1, w2, w3, ... Corresponding to each of the trained models M1, M2, M3, ... Are not fixed values but depend on the operating state of the main scanning mechanism 13. Change. The estimated value determining unit 921 first acquires a state variable representing the operating state of the main scanning mechanism 13 (step S41). The state variable may be a measured value input to the measured value input unit 911 described above, a variable acquired by another sensor, or an operation set by the user for the control unit 40. It may be a mode or the like.
　推定値決定部９２１は、取得した状態変数に基づいて、加重比率ｗ１，ｗ２，ｗ３，・・・を変更する（ステップＳ４２）。これにより、主走査機構１３の動作状態に応じて、高い推定精度を発揮できる学習済みモデルの加重比率を高くする。例えば、ある状態変数により表される動作状態において、複数の学習済みモデルＭ１，Ｍ２，Ｍ３，・・・のうち、特に学習済みモデルＭ２の推定精度が高い場合には、推定値決定部９２１は、加重比率ｗ２の値が相対的に高くなるように、複数の加重比率ｗ１，ｗ２，ｗ３，・・・を変更する。 The estimated value determination unit 921 changes the weighting ratios w1, w2, w3, ... Based on the acquired state variables (step S42). As a result, the weighting ratio of the trained model capable of exhibiting high estimation accuracy is increased according to the operating state of the main scanning mechanism 13. For example, in the operation state represented by a certain state variable, when the estimation accuracy of the trained model M2 is particularly high among the plurality of trained models M1, M2, M3, ... , The plurality of weighting ratios w1, w2, w3, ... Are changed so that the value of the weighting ratio w2 becomes relatively high.
　その後、推定値決定部９２１は、変更後の加重比率ｗ１，ｗ２，ｗ３，・・・を用いて、複数の仮推定値θ１，θ２，θ３，・・・の加重平均値を算出し、算出された加重平均値を、推定結果θｒとする（ステップＳ４３）。 After that, the estimated value determination unit 921 calculates and calculates the weighted average value of the plurality of tentative estimated values θ1, θ2, θ3, ... Using the changed weighted ratios w1, w2, w3, ... Let the weighted average value be the estimation result θr (step S43).
　このように、動作状態に応じて加重比率ｗ１，ｗ２，ｗ３，・・・を変更すれば、動作状態毎に、重視する学習済みモデルを変更できる。したがって、動作状態毎に、高い推定精度を発揮できる学習済みモデルの加重比率を高くして、より精度のよい推定結果θｒを得ることができる。 In this way, by changing the weighting ratios w1, w2, w3, ... According to the operating state, the trained model to be emphasized can be changed for each operating state. Therefore, it is possible to obtain a more accurate estimation result θr by increasing the weighting ratio of the trained model that can exhibit high estimation accuracy for each operating state.
　図１０は、推定値決定部９２１による推定結果の決定方法の第４の例を示すフローチャートである。図１０の例では、推定値決定部９２１は、まず、主走査機構１３の動作状態を表す状態変数を取得する（ステップＳ５１）。状態変数は、上述した計測値入力部９１１に入力される計測値であってもよく、他のセンサにより取得される変数であってもよく、あるいは、制御部４０に対してユーザが設定する動作モードなどであってもよい。 FIG. 10 is a flowchart showing a fourth example of a method of determining an estimation result by the estimation value determination unit 921. In the example of FIG. 10, the estimated value determination unit 921 first acquires a state variable representing the operating state of the main scanning mechanism 13 (step S51). The state variable may be a measured value input to the measured value input unit 911 described above, a variable acquired by another sensor, or an operation set by the user for the control unit 40. It may be a mode or the like.
　推定値決定部９２１は、取得した状態変数に基づいて、複数の仮推定値θ１，θ２，θ３，・・・のうちのいずれか１つを選択し、選択された仮推定値を推定結果θｒとする（ステップＳ５２）。例えば、ある状態変数により表される動作状態において、複数の学習済みモデルＭ１，Ｍ２，Ｍ３，・・・のうち、特に学習済みモデルＭ２の推定精度が高い場合には、推定値決定部９２１は、当該学習済みモデルＭ２から出力された仮推定値θ２を、推定結果θｒとする。 The estimated value determination unit 921 selects one of a plurality of tentative estimated values θ1, θ2, θ3, ... Based on the acquired state variables, and estimates the selected tentative estimated values θr. (Step S52). For example, in the operation state represented by a certain state variable, when the estimation accuracy of the trained model M2 is particularly high among the plurality of trained models M1, M2, M3, ... , The tentative estimated value θ2 output from the trained model M2 is used as the estimation result θr.
　すなわち、この例では、動作状態に応じて、複数の学習済みモデルＭ１，Ｍ２，Ｍ３，・・・から出力される複数の仮推定値θ１，θ２，θ３，・・・のうちのいずれか１つを採用する。このようにすれば、動作状態毎に、高い推定精度を発揮できる学習済みモデルを選択して、精度のよい推定結果θｒを得ることができる。 That is, in this example, one of a plurality of tentative estimated values θ1, θ2, θ3, ... Output from the plurality of trained models M1, M2, M3, ... Depending on the operating state. Adopt one. In this way, it is possible to select a trained model capable of exhibiting high estimation accuracy for each operating state and obtain an accurate estimation result θr.
　図７に戻る。ヨーイング角度θの推定結果θｒが決定されると、推定値決定部９２１は、その推定結果θｒを、制御部４０へ出力する（ステップＳ２６）。その後、制御部４０は、姿勢推定部９２から出力されるヨーイング角度θの推定結果θｒに基づいて、ステージ１２のヨーイング角度θを補正する（ステップＳ２７）。具体的には、制御部４０は、回転機構１５を動作させるか、あるいは、一対のリニアモータ６１のいずれかのトルク値を調整する。これにより、ステージ１２のヨーイング角度θが、所望の値に近づくように補正される。 Return to Fig. 7. When the estimation result θr of the yawing angle θ is determined, the estimation value determination unit 921 outputs the estimation result θr to the control unit 40 (step S26). After that, the control unit 40 corrects the yawing angle θ of the stage 12 based on the estimation result θr of the yawing angle θ output from the posture estimation unit 92 (step S27). Specifically, the control unit 40 operates the rotation mechanism 15 or adjusts the torque value of either of the pair of linear motors 61. As a result, the yawing angle θ of the stage 12 is corrected so as to approach a desired value.
　以上のように、この搬送装置１０では、主走査機構１３から出力される計測値または計測値に基づいて算出される値である入力データｄに基づいて、ステージ１２のヨーイング角度θを推定する。これにより、大掛かりな姿勢計測装置８０を常時設置することなく、ステージ１２のヨーイング角度θを推定できる。 As described above, the transport device 10 estimates the yawing angle θ of the stage 12 based on the measured value output from the main scanning mechanism 13 or the input data d which is a value calculated based on the measured value. As a result, the yawing angle θ of the stage 12 can be estimated without constantly installing the large-scale posture measuring device 80.
　また、この搬送装置１０では、複数の学習済みモデルＭ１，Ｍ２，Ｍ３，・・・から出力される複数の仮推定値θ１，θ２，θ３，・・・に基づいて、１つの推定結果θｒを出力する。このため、搬送装置１０の多くの動作状況において高い推定精度を実現できる。また、本実施形態では、複数の学習済みモデルＭ１，Ｍ２，Ｍ３，・・・が、それぞれ異なる機械学習アルゴリズムにより生成されたものである。このため、ある機械学習アルゴリズムにおいて高い推定精度を発揮しにくい動作状況において、他の機械学習アルゴリズムがそれを補完することができる。これにより、より多くの動作状況において、ステージ１２のヨーイング角度θを精度よく推定できる。 Further, in the transfer device 10, one estimation result θr is obtained based on a plurality of tentative estimation values θ1, θ2, θ3, ... Output from the plurality of trained models M1, M2, M3, ... Output. Therefore, high estimation accuracy can be realized in many operating conditions of the transport device 10. Further, in the present embodiment, a plurality of trained models M1, M2, M3, ... Are generated by different machine learning algorithms. Therefore, in an operating situation where it is difficult for a certain machine learning algorithm to exhibit high estimation accuracy, another machine learning algorithm can complement it. As a result, the yawing angle θ of the stage 12 can be estimated accurately in more operating conditions.
　＜４．変形例＞
　以上、本発明の一実施形態について説明したが、本発明は、上記の実施形態に限定されるものではない。 <4. Modification example>
Although one embodiment of the present invention has been described above, the present invention is not limited to the above embodiment.
　上記の実施形態では、計測値入力部９１１は、リニアモータ６１のトルク値を、制御基板６３から取得していた。しかしながら、計測値入力部９１１は、他の方法で、リニアモータ６１のトルク値を取得してもよい。例えば、直進機構６０の各リニアモータ６１にトルクセンサを取り付け、そのトルクセンサからトルク値を取得してもよい。 In the above embodiment, the measurement value input unit 911 acquires the torque value of the linear motor 61 from the control board 63. However, the measured value input unit 911 may acquire the torque value of the linear motor 61 by another method. For example, a torque sensor may be attached to each linear motor 61 of the straight-ahead mechanism 60, and a torque value may be acquired from the torque sensor.
　また、上記の実施形態では、全ての学習済みモデルＭ１，Ｍ２，Ｍ３，・・・に、同一の入力データｄが入力されていた。しかしながら、入力データ生成部９１２は、計測値入力部９１１に入力された計測値に対して、学習済みモデル毎に異なる処理を施してもよい。そして、複数の学習済みモデルＭ１，Ｍ２，Ｍ３，・・・に、異なる処理により生成された異なる入力データを入力してもよい。このようにすれば、各学習済みモデルに対して、より適した入力データを入力できる。 Further, in the above embodiment, the same input data d is input to all the trained models M1, M2, M3, ... However, the input data generation unit 912 may perform different processing for each trained model on the measurement value input to the measurement value input unit 911. Then, different input data generated by different processes may be input to the plurality of trained models M1, M2, M3, .... In this way, more suitable input data can be input for each trained model.
　また、上記実施形態の搬送装置１０は、主走査機構１３だけではなく、副走査機構１４および回転機構１５を備えていた。しかしながら、本発明は、副走査機構１４および回転機構１５を備えていない搬送装置を対象とするものであってもよい。 Further, the transport device 10 of the above embodiment includes not only the main scanning mechanism 13 but also the sub-scanning mechanism 14 and the rotation mechanism 15. However, the present invention may be intended for a transfer device that does not include a sub-scanning mechanism 14 and a rotating mechanism 15.
　また、上記実施形態の搬送装置１０は、描画装置１に搭載されていた。しかしながら、本発明は、描画装置１以外の装置に搭載される搬送装置を対象とするものであってもよい。例えば、搬送装置は、ステージ上に保持された基板に対して処理液を塗布する装置に搭載されるものであってもよい。また、搬送装置は、ステージ上に保持された記録媒体に対して印刷を行う装置に搭載されるものであってもよい。 Further, the transport device 10 of the above embodiment was mounted on the drawing device 1. However, the present invention may be intended for a transfer device mounted on a device other than the drawing device 1. For example, the transfer device may be mounted on a device that applies a processing liquid to a substrate held on a stage. Further, the transfer device may be mounted on a device that prints on a recording medium held on the stage.
　また、上記の実施形態では、事前学習処理において、レーザ干渉計８２により、ステージ１２のヨーイング角度θを実測していた。しかしながら、ステージ１２のヨーイング角度θは、他の方法で実測してもよい。例えば、カメラにより取得したステージ１２の画像に基づいて、ステージ１２のヨーイング角度θを実測してもよい。 Further, in the above embodiment, the yawing angle θ of the stage 12 is actually measured by the laser interferometer 82 in the pre-learning process. However, the yawing angle θ of the stage 12 may be actually measured by another method. For example, the yawing angle θ of the stage 12 may be actually measured based on the image of the stage 12 acquired by the camera.
　また、上記の実施形態の直進機構６０は、リニアモータ６１を有するものであった。しかしながら、リニアモータ６１に代えて、回転モータから出力される回転運動をボールネジにより直進運動に変換する機構を用いてもよい。 Further, the straight-ahead mechanism 60 of the above-described embodiment has a linear motor 61. However, instead of the linear motor 61, a mechanism that converts the rotary motion output from the rotary motor into a linear motion by a ball screw may be used.
　また、上記の実施形態および変形例に登場した各要素を、矛盾が生じない範囲で、適宜に組み合わせてもよい。 Further, each element appearing in the above-described embodiment and modification may be appropriately combined as long as there is no contradiction.
　１　描画装置
　１０　搬送装置
　１１　基台
　１２　ステージ
　１３　主走査機構
　１４　副走査機構
　１５　回転機構
　１６　下段支持プレート
　１７　中段支持プレート
　１８　ステージ姿勢推定装置
　２０　フレーム
　３０　描画処理部
　４０　制御部
　５０　センサ
　６０　直進機構
　６１　リニアモータ
　６２　エアガイド
　６３　制御基板
　８０　姿勢計測装置
　９１　入力データ取得部
　９２　姿勢推定部
　９１１　計測値入力部
　９１２　入力データ生成部
　９２１　推定値決定部
　Ｍ１，Ｍ２，Ｍ３　学習済みモデル
　Ｗ　基板
　ｄ　入力データ
　θ　ヨーイング角度
　θ１，θ２，θ３　仮推定値
　θｒ　推定結果 1 Drawing device 10 Conveyor device 11 Base 12 Stage 13 Main scanning mechanism 14 Sub-scanning mechanism 15 Rotation mechanism 16 Lower support plate 17 Middle support plate 18 Stage posture estimation device 20 Frame 30 Drawing processing unit 40 Control unit 50 Sensor 60 Straight-moving mechanism 61 Linear motor 62 Air guide 63 Control board 80 Attitude measurement device 91 Input data acquisition unit 92 Attitude estimation unit 911 Measurement value input unit 912 Input data generation unit 921 Estimated value determination unit M1, M2, M3 Learned model W board d Input data θ Yawing angle θ1, θ2, θ3 Temporary estimation value θr estimation result

Claims (9)

  1. 　平板状のステージを搬送機構により搬送する搬送装置において、前記ステージのヨーイング角度を推定するステージ姿勢推定装置であって、
    　前記搬送機構から出力される計測値または前記計測値に基づいて算出される値を、入力データとして取得する入力データ取得部と、
    　前記入力データに基づいて前記ステージのヨーイング角度を推定して、推定結果を出力する姿勢推定部と、
    を備え、
    　前記姿勢推定部は、
    　　前記入力データに基づいて前記ステージのヨーイング角度の仮推定値を出力する複数の学習済みモデルと、
    　　前記複数の学習済みモデルから出力される複数の前記仮推定値に基づいて、前記推定結果を決定する推定値決定部と、
    を有する、ステージ姿勢推定装置。     A stage attitude estimation device that estimates the yawing angle of the stage in a transfer device that conveys a flat plate-shaped stage by a transfer mechanism.
    An input data acquisition unit that acquires a measured value output from the transport mechanism or a value calculated based on the measured value as input data.
    A posture estimation unit that estimates the yawing angle of the stage based on the input data and outputs the estimation result.
    With
    The posture estimation unit
    A plurality of trained models that output a tentative estimate of the yawing angle of the stage based on the input data,
    An estimation value determination unit that determines the estimation result based on the plurality of provisional estimates output from the plurality of trained models.
    A stage attitude estimation device.
  2. 　請求項１に記載のステージ姿勢推定装置であって、
    　前記複数の学習済みモデルは、互いに異なるアルゴリズムにより生成されたものである、ステージ姿勢推定装置。     The stage posture estimation device according to claim 1.
    The plurality of trained models are generated by algorithms different from each other, and are stage attitude estimation devices.
  3. 　請求項１または請求項２に記載のステージ姿勢推定装置であって、
    　前記推定値決定部は、前記複数の仮推定値の平均値を、前記推定結果とする、ステージ姿勢推定装置。     The stage posture estimation device according to claim 1 or 2.
    The estimated value determining unit is a stage posture estimation device that uses the average value of the plurality of tentative estimated values as the estimated result.
  4. 　請求項１または請求項２に記載のステージ姿勢推定装置であって、
    　前記複数の学習済みモデルに、加重比率が設定されており、
    　前記推定値決定部は、前記加重比率を用いた前記複数の仮推定値の加重平均値を、前記推定結果とする、ステージ姿勢推定装置。     The stage posture estimation device according to claim 1 or 2.
    Weight ratios are set for the plurality of trained models.
    The estimated value determining unit is a stage posture estimation device that uses the weighted average value of the plurality of tentative estimated values using the weighted ratio as the estimated result.
  5. 　請求項４に記載のステージ姿勢推定装置であって、
    　前記推定値決定部は、前記搬送機構の動作状態を表す状態変数に基づき、前記加重比率を変更する、ステージ姿勢推定装置。     The stage posture estimation device according to claim 4.
    The estimated value determining unit is a stage posture estimation device that changes the weighting ratio based on a state variable representing an operating state of the transport mechanism.
  6. 　請求項１または請求項２に記載のステージ姿勢推定装置であって、
    　前記推定値決定部は、前記搬送機構の動作状態を表す状態変数に基づき、前記複数の仮推定値のうちのいずれか１つを選択し、選択された前記仮推定値を、前記推定結果とする、ステージ姿勢推定装置。     The stage posture estimation device according to claim 1 or 2.
    The estimated value determining unit selects one of the plurality of tentative estimated values based on a state variable representing the operating state of the transport mechanism, and uses the selected tentative estimated value as the estimated result. Stage attitude estimation device.
  7. 　請求項１から請求項６までのいずれか１項に記載のステージ姿勢推定装置であって、
    　前記搬送機構は、前記ステージを一対の直進機構により搬送し、
    　前記入力データ取得部は、前記一対の直進機構のトルク値の差分に基づいて、前記入力データを生成する、ステージ姿勢推定装置。     The stage posture estimation device according to any one of claims 1 to 6.
    The transport mechanism transports the stage by a pair of straight-moving mechanisms.
    The input data acquisition unit is a stage posture estimation device that generates the input data based on the difference between the torque values of the pair of straight-ahead mechanisms.
  8. 　請求項１から請求項７までのいずれか１項に記載のステージ姿勢推定装置と、
    　前記ステージと、
    　前記搬送機構と、
    を備えた、搬送装置。     The stage posture estimation device according to any one of claims 1 to 7.
    With the stage
    With the transfer mechanism
    A transport device equipped with.
  9. 　平板状のステージを搬送機構により搬送する搬送装置において、前記ステージのヨーイング角度を推定するステージ姿勢推定方法であって、
    　ａ）前記搬送機構から出力される計測値または前記計測値に基づいて算出される値を、入力データとして取得する工程と、
    　ｂ）前記入力データに基づいて前記ステージのヨーイング角度を推定して、推定結果を出力する工程と、
    を備え、
    　前記工程ｂ）では、複数の学習済みモデルに前記入力データを入力し、前記複数の学習済みモデルから出力される複数の仮推定値に基づいて、前記推定結果を決定する、ステージ姿勢推定方法。     A stage posture estimation method for estimating the yawing angle of the stage in a transport device that transports a flat stage by a transport mechanism.
    a) A step of acquiring a measured value output from the transport mechanism or a value calculated based on the measured value as input data, and
    b) A step of estimating the yawing angle of the stage based on the input data and outputting the estimation result.
    With
    In step b), a stage attitude estimation method in which the input data is input to a plurality of trained models, and the estimation result is determined based on a plurality of tentative estimation values output from the plurality of trained models.
PCT/JP2021/000818 2020-01-28 2021-01-13 Stage orientation estimation device, conveyance device, and stage orientation estimation method WO2021153228A1 (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
KR1020227025709A KR20220120645A (en) 2020-01-28 2021-01-13 Stage attitude estimation apparatus, conveyance apparatus, and stage attitude estimation method
CN202180011089.XA CN115023660A (en) 2020-01-28 2021-01-13 Stage attitude estimation device, conveyance device, and stage attitude estimation method

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP2020-011638 2020-01-28
JP2020011638A JP7379183B2 (en) 2020-01-28 2020-01-28 Stage posture estimation device, transport device, and stage posture estimation method

Publications (1)

Publication Number Publication Date
WO2021153228A1 true WO2021153228A1 (en) 2021-08-05

Family

ID=77078783

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2021/000818 WO2021153228A1 (en) 2020-01-28 2021-01-13 Stage orientation estimation device, conveyance device, and stage orientation estimation method

Country Status (4)

Country Link
JP (1) JP7379183B2 (en)
KR (1) KR20220120645A (en)
CN (1) CN115023660A (en)
WO (1) WO2021153228A1 (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TW202310131A (en) * 2021-08-24 2023-03-01 日商東京威力科創股份有限公司 Substrate-processing apparatus, model data generation apparatus, substrate-processing method, and model generation method

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2003264133A (en) * 2002-03-08 2003-09-19 Nikon Corp Stage controller, exposure apparatus, device manufacturing method and stage control method
JP2016072434A (en) * 2014-09-30 2016-05-09 株式会社Ｓｃｒｅｅｎホールディングス Pattern formation device and pattern formation method
WO2019189661A1 (en) * 2018-03-29 2019-10-03 国立大学法人奈良先端科学技術大学院大学 Learning dataset creation method and device
JP2020001127A (en) * 2018-06-28 2020-01-09 勇貴 高橋 Picking system, picking processing equipment, and program

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP7330816B2 (en) 2019-08-26 2023-08-22 株式会社Ｓｃｒｅｅｎホールディングス STAGE POSTURE ESTIMATION DEVICE, TRANSPORT DEVICE, STAGE POSTURE ESTIMATION METHOD, AND TRANSPORT METHOD

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2003264133A (en) * 2002-03-08 2003-09-19 Nikon Corp Stage controller, exposure apparatus, device manufacturing method and stage control method
JP2016072434A (en) * 2014-09-30 2016-05-09 株式会社Ｓｃｒｅｅｎホールディングス Pattern formation device and pattern formation method
WO2019189661A1 (en) * 2018-03-29 2019-10-03 国立大学法人奈良先端科学技術大学院大学 Learning dataset creation method and device
JP2020001127A (en) * 2018-06-28 2020-01-09 勇貴 高橋 Picking system, picking processing equipment, and program

Also Published As

Publication number Publication date
JP7379183B2 (en) 2023-11-14
KR20220120645A (en) 2022-08-30
JP2021117394A (en) 2021-08-10
CN115023660A (en) 2022-09-06

Similar Documents

Publication Publication Date Title
JP7470295B2 (en) Technique for aligned printing of permanent layers with improved speed and accuracy.
JP5279299B2 (en) POSITION CONTROL APPARATUS, EXPOSURE APPARATUS, AND DEVICE MANUFACTURING METHOD EQUIPPED WITH AN ITS
KR960016176B1 (en) Method to determine tool paths for thinning and correcting errors in thickness profiles of films
US7291999B2 (en) System and method for diagnosing a controller in a limited rotation motor system
WO2021153228A1 (en) Stage orientation estimation device, conveyance device, and stage orientation estimation method
Miermeister et al. Auto-calibration method for overconstrained cable-driven parallel robots
JP2008139035A (en) 3-d environment measurement device and mobile robot provided with the same
US7342645B2 (en) Stage control apparatus and method, stage apparatus and exposure apparatus
US11223303B2 (en) Motor with force constant modeling and identification for flexible mode control
Spacek et al. Digital control of Ball & Plate model using LQ controller
JP2006226862A (en) Xy stage
Oomen et al. Model-based control for high-tech mechatronic systems
JP2021033068A (en) Stage posture estimation device, transport device, stage posture estimation method and transport method
JP7263121B2 (en) DATA COLLECTION METHOD, LEARNING METHOD, DATA COLLECTION SYSTEM AND LEARNING SYSTEM
WO2021153229A1 (en) Transport device and transport method
CN108563123A (en) The adaptive iteration study robust control system and design method of linear motor X-Y
KR20220046483A (en) Positioning apparatus, lithography apparatus and article manufacturing method
WO2020162200A1 (en) Control device and control program
JP7242422B2 (en) Stage posture grasping device, stage moving device, stage posture grasping method, and stage posture correcting method
JP4580600B2 (en) Galvano scanner control method, apparatus, and galvano scanner
JP6234170B2 (en) Program, generating method, generating apparatus, driving apparatus, processing apparatus, lithographic apparatus, and article manufacturing method
TWI808448B (en) Motion control using an artificial neural network
US11774865B2 (en) Method of controlling a position of a first object relative to a second object, control unit, lithographic apparatus and apparatus
Oomen et al. 4 Model-Based Control
Vignesh et al. Implementation of Indoor Navigation Control for Two-Wheeled Self-balancing Robot

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21748203

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 20227025709

Country of ref document: KR

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21748203

Country of ref document: EP

Kind code of ref document: A1