JP2014016470A - Position prediction device, position prediction method and substrate treatment apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a technique capable of predicting the position of an object upon practicing of treatment at sufficient precision.SOLUTION: The position prediction device 1 comprises: a measurement part, with each of a plurality of times within a time after the relative movement start of an object 9 and a treatment head 102 and before the treatment time at which a treatment head 102 practices treatment to the object 9 as measurement time, the measurement part measuring the position of the object 9 in each of a plurality of the measurement times; and a prediction part which predicts the position of the object 9 of the treatment time based on the measurement position of the object 9 in each of a plurality of the measurement times.

Description

  The present invention relates to a technique for predicting the position of an object relative to a processing head in a processing apparatus including a processing head that performs processing on the object while moving relative to the object.

  There are various apparatuses that perform processing on an object by emitting light rays or droplets from the processing head to the object while relatively moving the object and the processing head. For example, a substrate coated with a photosensitive material such as a photoresist is moved relative to an optical head that is a processing head, and light is irradiated from the optical head to the substrate to form a pattern (circuit pattern) on the substrate. There is known a substrate processing apparatus for forming a substrate. In addition, for example, an inkjet image forming apparatus that forms an image or the like on a print sheet by ejecting ink droplets from the print head to the print sheet while moving the print sheet relative to a print head that is a processing head. It has been known.

  In many cases, this type of apparatus is required to have a function capable of accurately irradiating or discharging light or droplets onto a target position on an object. For example, the above-described substrate processing apparatus is required to have a function capable of irradiating light in a state of being accurately aligned with a target position (for example, a formation position of a lower layer pattern) on the substrate. Since the pattern to be formed on the substrate is continually miniaturized, the processing accuracy required for the substrate processing apparatus is becoming stricter year by year.

  However, when processing is performed, the position of the object relative to the processing head may be slightly deviated from the ideal position. When such a positional deviation occurs, it is impossible to accurately reach the target position on the object such as a light beam or a droplet emitted from the processing head. Therefore, for example, in Patent Document 1, in a process processing apparatus including a process processing unit that applies a spacer material on a surface of a substrate to form a spacer pattern, the position of the substrate based on image data captured by a measurement camera. It is disclosed that the amount of deviation is specified, and further, the amount of positional deviation is corrected in consideration of the relative distance between the measurement camera and the process processing unit. Further, for example, Patent Document 2 discloses a technique for detecting a misregistration detection pattern for each color and calculating a misregistration amount of each color pattern in a tandem image forming apparatus.

JP 2010-79797 A JP 2009-157056 A

  By the way, misalignment of the object with respect to the processing head is caused by various factors such as drive system vibration, mechanical error, and state of the object acting in a complex manner. For this reason, the positional deviation amount of the object with respect to the processing head may change with time after the relative movement is started, and the mode of the change is also different for each transport operation. There is a possibility. In such a situation, it is not easy to predict the positional deviation amount of the object when the process is executed with high accuracy. For example, even in the configurations of Patent Documents 1 and 2, the predicted positional deviation amount However, there is a possibility that the actual positional deviation amount of the object when the process is executed is greatly deviated. In this case, the process is performed at a position shifted from the target position, and the processing accuracy is lowered.

  The present invention has been made in view of the above-described problems, and the target when the processing head executes processing on the target object when performing processing on the target object while relatively moving the target object and the processing head. The object is to provide a technique capable of predicting the position of an object with sufficient accuracy.

  In a processing apparatus including a processing head that performs processing on the target while moving relative to the target, the first aspect is a processing time at which the processing head executes processing on the target A position predicting device for predicting the position of the object relative to the processing head, wherein the object and the processing head start moving relatively and before the processing time. Each of the plurality of times as a measurement time, based on the measurement unit for measuring the position of the object at each of the plurality of measurement times, and the measurement position of the object at each of the plurality of measurement times, A prediction unit that predicts the position of the object at the processing time.

  A 2nd aspect is a position prediction apparatus which concerns on a 1st aspect, Comprising: The said prediction part makes 1 or more measurement time of the said several measurement time and the said process time into object time, and sets a process parameter. A predicted position calculation unit that predicts the position of the target object at the target time based on the measurement position of the target object at each of two or more measurement times before the target time, using a prediction model including And when the predicted position calculation unit predicts the position of the target object using any one of the plurality of measurement times as a target time, and the target at the measurement time determined as the target time. A prediction error acquisition unit that calculates a difference from the measurement position of the object and acquires it as a prediction error; a parameter adjustment unit that adjusts the processing parameter so that the prediction error approaches zero; and the prediction position When the calculation unit predicts the position of the target object using the prediction model including the processing parameter adjusted by the parameter adjustment unit with the processing time as the target time, the predicted position obtained is the final An output unit that outputs the predicted position.

  A third aspect is a position prediction apparatus according to the second aspect, wherein the prediction unit includes a plurality of the prediction position calculation units, and the plurality of prediction position calculation units use the prediction models different from each other. The position of the object is predicted, and the output unit evaluates each of the plurality of prediction models based on the prediction error calculated for the prediction position acquired using each of the plurality of prediction models. An evaluation unit, the prediction position of the object at the processing time obtained using a prediction model given a positive evaluation by the evaluation unit among the plurality of prediction models, Output as predicted position.

  A fourth aspect is the position prediction apparatus according to any one of the first to third aspects, wherein the measurement unit is fixed to the processing head, and the imaging unit captures the object. At each of a plurality of measurement times, an imaging control unit that causes the imaging unit to image the object, and the imaging data acquired by the imaging unit are analyzed, and the object at the measurement time at which the imaging data is acquired is analyzed. An analysis processing unit that specifies a position and acquires the position as a measurement position of the object.

  A 5th aspect is a position prediction apparatus which concerns on a 4th aspect, Comprising: The said imaging unit is provided with the some imaging part, and the said several imaging part is the said processing head relative movement with respect to the said target object. Are arranged along the movement direction on the downstream side of the processing head, and the imaging control unit is configured so that each of the plurality of imaging units is above a characteristic portion on the object. At the time of arrival, the image capturing unit is caused to image the characteristic portion.

  According to a sixth aspect of the present invention, in the processing apparatus including a processing head that performs processing on the target while moving relative to the target, processing time at which the processing head executes processing on the target A position prediction method for predicting the position of the object with respect to the processing head, wherein a) the object and the processing head are started to move relatively and before the processing time. Each of a plurality of times within the time is set as a measurement time, and the step of measuring the position of the object at each of the plurality of measurement times, and b) each of the plurality of measurement times acquired in step a) Predicting the position of the object at the processing time based on the measurement position of the object.

  A seventh aspect is a position prediction method according to the sixth aspect, wherein the step b) includes b1) one or more of the plurality of measurement times and the processing time as a target time. Predicting the position of the object at the target time based on the measurement position of the object at each of two or more measurement times before the target time, using a prediction model including a processing parameter. And b2) In the step b1), when the position of the object is predicted with any one of the plurality of measurement times as the target time, the obtained predicted position and the measurement time at the target time are Calculating a difference from the measurement position of the object and obtaining it as a prediction error; b3) adjusting the processing parameter so that the prediction error approaches zero; and b4) b1). , When the position of the target object is predicted using the prediction model including the processing parameter adjusted in the step b3) with the processing time as the target time, the obtained predicted position is used as the final predicted position. And a step of outputting.

  An eighth aspect is a position prediction method according to the seventh aspect, wherein in the step b1), the position of the object at the target time is predicted using each of a plurality of different prediction models, The step b) further comprises: b5) evaluating each of the plurality of prediction models based on the prediction error calculated for the prediction position acquired using each of the plurality of prediction models. In the step b4), the predicted position of the object at the processing time obtained using the prediction model given a positive evaluation in the step b5) is output as the final predicted position.

  A ninth aspect is a position prediction method according to any one of the sixth to eighth aspects, wherein the step a) includes: a1) the plurality of measurements on an imaging unit fixed to the processing head. A step of imaging the object at each of the times; a2) analyzing the imaging data acquired by the imaging unit, identifying the position of the object at the measurement time at which the imaging data was acquired, and Obtaining as a measurement position of the object.

  A tenth aspect is a position prediction method according to the ninth aspect, wherein the imaging unit includes a plurality of imaging units, and the plurality of imaging units are configured to move the processing head relative to the object. Are arranged along the movement direction on the downstream side of the processing head, and in the step a1), each of the plurality of imaging units is located above the characteristic portion on the object. At the time of arrival, the image capturing unit is caused to image the characteristic portion.

  An eleventh aspect is a substrate processing apparatus, a processing head that irradiates a substrate with light to expose a pattern, a drive mechanism that relatively moves the substrate and the processing head, and the processing head A position prediction unit that predicts the position of the substrate relative to the processing head at a processing time when the substrate is irradiated with light, and the processing based on the position of the substrate at the processing time predicted by the position prediction unit. A correction unit that corrects the irradiation position of light from the head, and the position prediction unit is a time after the substrate and the processing head start to move relatively and before the processing time. Each of a plurality of times is a measurement time, a measurement unit that measures the position of the object at each of a plurality of measurement times, and a measurement position of the object at each of the plurality of measurement times Based on, and a prediction unit for predicting a position of the object of the processing time.

  According to the first and sixth aspects, the position of the object at the processing time based on the measurement position of the object at each of the plurality of measurement times after the object and the processing head start moving relatively. Predict. According to this configuration, the position of the object can be predicted in consideration of the displacement state of the object after the start of relative movement. Therefore, it is possible to predict the position of the object when the process is executed with sufficient accuracy.

  According to the second and seventh aspects, the position of the object at the measurement time is predicted using the prediction model. Then, the processing parameter included in the prediction model is adjusted so that the difference between the obtained predicted position and the actual measurement position approaches zero. According to this configuration, the processing parameters included in the prediction model are adjusted according to the displacement state of the object after the start of relative movement, so that the prediction accuracy can be further improved.

  According to the third and eighth aspects, the position of the object is predicted using each of a plurality of different prediction models, and obtained using a prediction model given a positive evaluation among the plurality of prediction models. The predicted position is output as the final predicted position. According to this configuration, it is possible to widely cope with various displacement situations of the object, and high versatility is realized. In addition, the predicted position obtained using the prediction model that best matches the displacement status of the object after the start of relative movement is output as the final predicted position, so high prediction accuracy can be realized stably. can do.

  According to the fourth, fifth, ninth, and tenth aspects, the imaging unit images the object at each of the plurality of measurement times, and the measurement of the object at each measurement time is performed based on the acquired imaging data. A location is identified. According to this configuration, the measurement position of the object at each of a plurality of measurement times can be acquired with a simple configuration.

  According to the eleventh aspect, the position of the substrate at the processing time is predicted based on the measurement position of the substrate at each of the plurality of measurement times after the movement of the substrate and the processing head is relatively started. According to this configuration, the position of the substrate can be predicted in consideration of the displacement state of the substrate after the start of relative movement. Therefore, the position of the substrate when processing is performed can be predicted with sufficient accuracy. As a result, it is possible to accurately irradiate the target position on the substrate from the processing head.

It is a figure which shows typically schematic structure of the processing apparatus carrying a position prediction apparatus. It is a block diagram which shows the hardware constitutions of a control part. It is a block diagram which shows the structure with which the position prediction apparatus which concerns on 1st Embodiment is provided. It is a schematic diagram for demonstrating the flow of the position prediction process which concerns on 1st Embodiment. It is a schematic diagram for demonstrating the flow of the position prediction process which concerns on 1st Embodiment. It is a schematic diagram for demonstrating the flow of the position prediction process which concerns on 1st Embodiment. It is a schematic diagram for demonstrating the flow of the position prediction process which concerns on 1st Embodiment. It is a schematic diagram for demonstrating the flow of the position prediction process which concerns on 1st Embodiment. It is a schematic diagram for demonstrating the flow of the position prediction process which concerns on 1st Embodiment. It is a schematic diagram for demonstrating the flow of the position prediction process which concerns on 1st Embodiment. It is a block diagram which shows the structure with which the position prediction apparatus which concerns on 2nd Embodiment is provided. It is a schematic diagram for demonstrating the flow of the position prediction process which concerns on 2nd Embodiment. It is a schematic diagram for demonstrating the flow of the position prediction process which concerns on 2nd Embodiment. It is a schematic diagram for demonstrating the flow of the position prediction process which concerns on 2nd Embodiment. It is a schematic diagram for demonstrating the flow of the position prediction process which concerns on 2nd Embodiment. It is a schematic diagram for demonstrating the flow of the position prediction process which concerns on 2nd Embodiment. It is a schematic diagram for demonstrating the flow of the position prediction process which concerns on 2nd Embodiment. It is a schematic diagram for demonstrating the flow of the position prediction process which concerns on 2nd Embodiment. It is a schematic diagram for demonstrating the flow of the position prediction process which concerns on a modification.

  Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. The following embodiment is an example embodying the present invention, and does not limit the technical scope of the present invention.

<I. First Embodiment>
<1. Processing apparatus 100>
A configuration of the processing apparatus 100 on which the position prediction apparatus 1 according to the first embodiment is mounted will be described with reference to FIGS. 1 and 2. FIG. 1 is a diagram schematically illustrating a schematic configuration of a processing apparatus 100 in which the position prediction apparatus 1 is mounted. FIG. 2 is a block diagram illustrating a hardware configuration of the control unit 105.

  The processing apparatus 100 includes a stage 101 on which the object 9 is placed, a processing head 102 disposed above the stage 101 and fixed to the base unit 110 of the processing apparatus 100, and the stage 101 on the base unit 110. And a drive mechanism 103 that moves relative to the processing head 102.

  In the processing apparatus 100, the driving mechanism 103 moves the stage 101 on which the object 9 is placed (in this embodiment, for example, in the + X direction), while the processing head 102 moves the object 9. The process for is executed. That is, the processing head 102 sequentially performs processing on each position of the target object 9 while relatively moving in the −X direction with respect to the target object 9.

  The processing head 102 is, for example, a drawing head that exposes (draws) a pattern (for example, a circuit pattern) on a substrate (specifically, a substrate on which a layer of a photosensitive material such as a resist is formed). In this case, the processing head 102 specifically uses pattern data (specifically, for example, CAD data generated using CAD (computer aided design)) that describes data to be drawn on the substrate. The light is spatially modulated in accordance with the rasterized data, and the substrate that is the object 9 is irradiated with the spatially modulated light. However, the processing head 102 is configured to spatially modulate light using, for example, a GLV (Grating Light Valve) (“GLV” is a registered trademark) which is a diffraction grating type light modulation element. can do. When such a processing head 102 is provided, the processing apparatus 100 constitutes a substrate processing apparatus (specifically, a drawing apparatus). That is, in the processing apparatus 100 that is a drawing apparatus, the drawing head that is the processing head 102 irradiates light to each position of the substrate while moving relative to the substrate that is the object 9, thereby forming a pattern on the substrate. The drawing process is performed.

  The position prediction device 1 is mounted on the processing device 100. The position prediction apparatus 1 predicts the relative position of the object 9 with respect to the processing head 102 at the time (processing time) when the processing from the processing head 102 to the object 9 is executed. The configuration and operation of the position prediction apparatus 1 will be described later.

  In addition, the processing device 100 includes a correction unit 104 that corrects the processing position of the processing head 102 based on the predicted position of the object 9 output from the position prediction device 1. Specifically, for example, based on the predicted position of the object 9 at the processing time output from the position prediction device 1, the correction unit 104 calculates the amount of deviation from the ideal position of the object 9 at the processing time. This is calculated and acquired as the “positional deviation amount”. Then, the correction unit 104 corrects the processing position of the processing head 102 at the processing time (the irradiation position of light when the processing head 102 is a drawing head, for example) so as to be shifted by the positional deviation amount. As a result, the target position of the object 9 is accurately processed.

  When the processing head 102 is, for example, a drawing head, the correction unit 104 includes, for example, an optical component (for example, a glass plate) disposed on the path of light emitted from the drawing head, and the optical component. What is necessary is just to comprise including the drive mechanism which changes an attitude | position. In this case, the position where the light emitted from the drawing head enters the substrate can be shifted by a desired amount by changing the attitude of the optical component by the drive mechanism. That is, the light irradiation position can be corrected by a desired amount.

  In addition, the processing device 100 includes a control unit 105 that is electrically connected to each unit included in the processing device 100 and controls the operation of each unit of the processing device 100 while executing various arithmetic processes.

  As illustrated in FIG. 2, the control unit 105 is configured by a general computer in which, for example, a CPU 151, a ROM 152, a RAM 153, a storage device 154, and the like are interconnected via a bus line 155. The ROM 152 stores basic programs and the like, and the RAM 153 is used as a work area when the CPU 151 performs predetermined processing. The storage device 154 is configured by a nonvolatile storage device such as a flash memory or a hard disk device. The storage device 154 stores a processing program P0. The CPU 151 as the main control unit performs arithmetic processing according to the procedure described in the processing program P0, and is configured to realize various functions. Yes. The processing program P0 is normally stored and used in advance in a memory such as the storage device 154, but is recorded in a recording medium such as a CD-ROM or DVD-ROM or an external flash memory (program product). ) (Or provided by downloading from an external server via a network, etc.) and stored in a memory such as the storage device 154 in addition or exchange. Note that some or all of the functions realized in the control unit 105 may be realized in hardware by a dedicated logic circuit or the like. In the control unit 105, an input unit 156, a display unit 157, and a communication unit 158 are also connected to the bus line 155. The input unit 156 includes various switches, a touch panel, and the like, and receives various input setting instructions from an operator. The display unit 157 includes a liquid crystal display device, a lamp, and the like, and displays various types of information under the control of the CPU 151. The communication unit 158 has a data communication function via a LAN or the like.

<2. Position prediction apparatus 1>
As described above, the position predicting device 1 is mounted on the processing device 100 and predicts the relative position of the object 9 with respect to the processing head 102 at the processing time when the processing from the processing head 102 to the object 9 is executed. In the following description, it is assumed that the processing time to be predicted by the position prediction apparatus 1 is the time at which the processing head 102 starts processing the object 9 (“processing start time” described later).

  The configuration of the position prediction apparatus 1 will be described with reference to FIG. 3 in addition to FIGS. FIG. 3 is a block diagram illustrating a configuration included in the position prediction apparatus 1.

  The position prediction apparatus 1 includes a measurement unit 10 that measures the position of the object 9 and a prediction unit 20 that predicts the position of the object 9 based on the measurement position of the object 9 acquired by the measurement unit 10. The imaging control unit 12, the analysis processing unit 13, and the prediction unit 20 included in the measurement unit 10 are, for example, as a main control unit according to the procedure described in the position prediction program P1 stored in the storage device 154 of the control unit 105. This is realized by the CPU 151 performing arithmetic processing. The position prediction program P1 is normally stored and used in advance in a memory such as the storage device 154, but is recorded in a recording medium such as a CD-ROM or DVD-ROM or an external flash memory (program). Product (or provided by downloading from an external server via a network, etc.) and additionally or exchangeably stored in a memory such as the storage device 154. In addition, a part or all of these functional units 10 and 20 may be realized in hardware by a dedicated logic circuit or the like.

<I. Measuring unit 10>
The measuring unit 10 has a plurality of times within the time before the processing time (in this case, the processing start time) that is the target of prediction after the object 9 and the processing head 102 start moving relatively. Each of them is “measurement time”, and the position of the object 9 at each of the plurality of measurement times is measured. Specifically, the measurement unit 10 includes an imaging unit 11, an imaging control unit 12, and an analysis processing unit 13.

<Imaging unit 11>
The imaging unit 11 is fixed to the processing head 102 and images the object 9 on the stage 101. The imaging unit 11 includes n imaging units 111 (where n is an integer equal to or greater than 2, for example, n = 10).

  Each of the n imaging units 111 is a mechanism that images an in-plane region of the object 9 on the stage 101 and has the same configuration. That is, each imaging unit 111 includes, for example, a light source configured by LEDs, a lens barrel, an objective lens, and a CCD image sensor configured by an area image sensor (two-dimensional image sensor). In this configuration, each imaging unit 111 captures an image of the object 9 on the stage 101 in accordance with an instruction from the imaging control unit 12 described later, and an in-plane region of the object 9 (immediately below the imaging unit 111). Two-dimensional image data (imaging data) obtained by imaging (region) is acquired.

  The n imaging units 111 are fixed with respect to the processing head 102. Therefore, when the drive mechanism 103 moves the stage 101, the plurality of imaging units 111 are also moved relative to the object 9 together with the processing head 102.

  Further, each of the n imaging units 111 is arranged on the −X side of the processing head 102 in a state of being arranged in a line along the X axis. As described above, the processing head 102 performs processing on the target object 9 while relatively moving in the −X direction with respect to the target object 9. That is, the n imaging units 111 are arranged along the moving direction on the downstream side of the processing head 102 with respect to the moving direction in which the processing head 102 is moved relative to the object 9 when the processing is executed. . Note that the n imaging units 111 may be arranged at a certain interval, or may be arranged so that the interval decreases as the processing head 102 is approached, for example.

  Now, the object 9 placed on the stage 101 has a position detection mark (specifically, near the + X side end (that is, the processing start position) of the processing target area defined within the surface. For example, it is assumed that a cross mark (M) is attached. The measuring unit 10 detects the mark M as a characteristic part on the object 9 and detects the position of the object 9 by detecting the mark M that is the characteristic part.

  When the processing on the object 9 is executed by the processing apparatus 100, first, the mark M on the object 9 is arranged on the −X side from directly below the imaging unit 111 on the most −X side. In this state, the drive mechanism 103 starts moving the stage 101 in the + X direction (the state shown in FIG. 1 and also referred to as “movement start state”). When the stage 101 starts to move in the + X direction, the n imaging units 111 and the processing heads 102 start to move relative to the object 9 in the −X direction. First, the most imaging unit 111 on the −X side passes above the mark M (the state shown in FIG. 4), and then the second imaging unit 111 from the −X side passes above the mark M. (The state shown in FIG. 5), and so on, each imaging unit 111 passes over the mark M one after another. Then, after the most + X side imaging unit (n-th imaging unit) 111 passes above the mark M, the processing head 102 reaches directly above the processing start position (state shown in FIG. 10). When the processing head 102 reaches directly above the processing start position, the processing head 102 performs processing on the object 9 (in the case where the processing head 102 is a drawing head, for example, irradiation of light on the substrate that is the target 9). Will be started.

  Here, the time at which the movement of the stage 101 is started from the movement start state (that is, the time at which the relative movement between the processing head 102 and the object 9 is started) is also referred to as “movement start time”. The time at which the processing head 102 arrives immediately above the processing start position (that is, the time at which the processing head 102 starts processing the object 9) is also referred to as “processing start time”. That is, in the processing apparatus 100, each imaging unit 111 passes over the mark M one after another within the time after the movement start time and before the processing start time. Each time when the mark passes over the mark M is defined as a “measurement time”.

<Imaging control unit 12>
The imaging control unit 12 controls the imaging unit 11 to acquire imaging data of the object 9 at each of a plurality of measurement times. That is, as described above, when the stage 101 starts moving in the + X direction from the movement start state, each of the n imaging units 111 sequentially passes above the mark M, and the imaging control unit 12 At the time when each imaging unit 111 reaches above the mark M (that is, at the timing when the mark M is captured in the imaging region of the imaging unit 111), the imaging unit 111 above the mark M An in-plane region of the object 9 is imaged. That is, the imaging unit 111 is caused to image the mark M that is a characteristic part of the object 9. Thereby, imaging data of the object 9 at different measurement times (specifically, imaging data obtained by imaging the mark M on the object 9) is acquired from each of the n imaging units 111. .

<Analysis processing unit 13>
The analysis processing unit 13 analyzes the imaging data sequentially acquired by the imaging unit 11 and specifies the position of the object 9 at the measurement time at which the imaging data is acquired. Specifically, for example, the analysis processing unit 13 performs pattern matching between the imaging data acquired by the imaging unit 111 and the template image data stored in advance in the storage device 154, and sets the mark M from the imaging data. To detect. When the mark M is detected, the analysis processing unit 13 further specifies the position of the mark M, and acquires the specified position as the measurement position of the object 9 at the measurement time when the imaging data is acquired. To do. However, this measurement position (and a predicted position calculated based on the measurement position) is information represented by, for example, a two-dimensional coordinate position defined by an orthogonal coordinate system defined on the base 110. The relative position of the object 9 with respect to the processing head 102 is shown.

  As described above, the imaging unit 11 sequentially acquires imaging data of the object 9 captured at each of a plurality of measurement times. Therefore, the analysis processing unit 13 analyzes the captured image data acquired one after another, thereby acquiring the measurement position of the object 9 at each of a plurality of measurement times.

<Ii. Prediction unit 20>
The prediction unit 20 predicts the position of the object 9 at the processing start time based on the measurement position of the object 9 at each of a plurality of measurement times acquired by the measurement unit 10. Specifically, the prediction unit 20 includes a predicted position calculation unit 21, a prediction error acquisition unit 22, a parameter adjustment unit 23, and an output unit 24.

<Predicted position calculation unit 21>
The predicted position calculation unit 21 uses the prediction model including the processing parameter to measure the position of the object 9 at the target time at two or more (two in this embodiment) measurement times before the target time. Is predicted based on the measurement position of the object 9 in each of the above. However, here, all of the third and subsequent measurement times and the process start time are set as the target times.

  More specifically, when the measurement unit 10 acquires the measurement position of the object 9 at a certain measurement time (current measurement time), the predicted position calculation unit 21 determines the measurement position of the object 9 at the current measurement time and Based on the measurement position of the object 9 at the measurement time before the current measurement time, the object 9 at the next measurement time after the current measurement time (or the processing start time when the current measurement time is the last measurement time) Predict the position of. Note that the predicted position calculation unit 21 acquires various types of information (for example, information such as the moving speed of the stage 101 at each measurement time) from the drive mechanism 103 as necessary in predicting the position of the object 9. Then, the position of the object 9 may be predicted in consideration of the information.

  The prediction model used by the predicted position calculation unit 21 may be any type. For example, a prediction model that assumes that the object 9 vibrates in a direction orthogonal to the traveling direction in a horizontal plane can be employed. Further, for example, a prediction model that assumes that the object 9 vibrates in a direction parallel to the traveling direction in a horizontal plane can be employed. Further, for example, it is possible to adopt a prediction model that assumes that a vibration in which a plurality of vibration components having different directions and vibration frequencies are combined is generated. In each of these prediction models, for example, the frequency, amplitude, initial phase, and the like of the vibration component may be included as processing parameters. In addition, for example, a prediction model that assumes that the object 9 is traveling along a linear axis that is not parallel to the X axis in a horizontal plane can be employed. Further, for example, a prediction model that is assumed to vibrate while proceeding along a linear axis that is not parallel to the X axis can be employed. In these prediction models, for example, an angle formed by a linear axis that is not parallel to the X axis with respect to the X axis may be included as a processing parameter.

<Prediction error acquisition unit 22>
The prediction error acquisition unit 22 predicts the position of the object 9 for a certain measurement time (that is, the position of the object 9 predicted by the prediction position calculation unit 21 using the prediction model) and the object 9 for the measurement time. Is calculated as a prediction error of the prediction model at the measurement time, by calculating the difference from the measurement position (that is, the position of the object 9 actually measured and acquired by the measurement unit 10).

<Parameter adjustment unit 23>
The parameter adjustment unit 23 is included in the prediction model so that the prediction error approaches zero when the prediction error of the prediction model at a certain measurement time is other than zero (or exceeds the allowable value). Adjust processing parameters. That is, the processing parameters are optimized so that the prediction error approaches zero.

  However, when the parameter adjustment unit 23 changes the processing parameter, the predicted position calculation unit 21 thereafter predicts the position of the object 9 using a prediction model including the changed processing parameter. And according to the prediction error calculated about the obtained prediction position, the parameter adjustment part 23 adjusts a process parameter again. In this way, every time a prediction error for each measurement time is acquired, adjustment of the processing parameters included in the prediction model is repeated. As the adjustment of the processing parameters is repeated, the prediction model is adapted to the displacement state of the object 9 after the start of relative movement. That is, the prediction accuracy of the prediction model increases.

<Output unit 24>
The output unit 24 outputs the predicted position of the object 9 at the processing start time acquired by the predicted position calculation unit 21 using the prediction model as the final predicted position. However, as described above, before the predicted position calculation unit 21 predicts the position of the object 9 at the processing start time, adjustment of the processing parameters included in the prediction model used for this is repeated. That is, the position of the object 9 at the processing start time is predicted using a prediction model including the processing parameter adjusted by the parameter adjusting unit 23. Therefore, the obtained predicted position is a highly reliable value.

  The output unit 24 includes an evaluation unit 241 that evaluates the prediction model. The evaluation unit 241 accumulates prediction errors at each of a plurality of measurement times, and evaluates the prediction model based on the accumulated group of prediction errors. Specifically, for example, the evaluation unit 241 has a case where the prediction error is within a predetermined allowable range (for example, a range whose absolute value is equal to or less than a predetermined threshold) after a certain time (preferably, zero). Give a positive rating to the prediction model.

  When the evaluation unit 241 gives a negative evaluation to the prediction model, the output unit 24 performs predetermined error processing instead of (or in combination with) the output of the final prediction position. Specifically, the error process may be a process for causing the display unit 157 to display a message indicating that the position prediction process has failed.

<3. Flow of processing>
The flow of the process (position prediction process) in which the position prediction apparatus 1 predicts the position of the object 9 at the process start time will be specifically described with reference to FIGS. 4 to 10 are schematic diagrams for explaining the flow of the position prediction process, and each of the processes at different times is schematically shown. In FIGS. 4 to 10, the processing performed by each processing unit 21, 22, 23 at each time is schematically shown as one block.

  As described above, at the time of executing the process, first, the mark M on the object 9 is set to the movement start state arranged on the −X side from directly below the −X side imaging unit 111 (see FIG. 1). In this state, the drive mechanism 103 starts moving the stage 101 in the + X direction (movement start time). Then, after the n imaging units 111 pass over the mark M one after another (that is, after n measurement times have passed), the processing head 102 reaches directly above the processing start position, and at this time, Processing for the object 9 is started from the processing head 102 (processing start time). In the following, the time (measurement time) when the i-th (i = 1, 2,..., N) imaging unit 111 counting from the −X side passes above the mark M is referred to as “measurement time t (i)”. The time when the processing head 102 arrives just above the processing start position (processing start time) is indicated as “processing start time t (n + 1)”.

<Measurement time t (1)>
When a certain time elapses from the movement start time, as shown in FIG. 4, the first imaging unit 111 counting from the −X side passes above the mark M (measurement time t (1)). Then, the imaging control unit 12 causes the imaging unit 111 immediately above the mark M to image the in-plane region of the object 9. Thereby, imaging data of the object 9 at the measurement time t (1) is acquired. When the imaging data is acquired, the analysis processing unit 13 subsequently analyzes the imaging data and acquires the measurement position A (1) of the object 9 at the measurement time t (1). The acquired measurement position A (1) is input to the predicted position calculation unit 21.

<Measurement time t (2)>
When a certain time elapses from the measurement time t (1), the second imaging unit 111 counting from the −X side passes over the mark M as shown in FIG. 5 (measurement time t (2)). . Then, the imaging control unit 12 causes the imaging unit 111 immediately above the mark M to image the in-plane region of the object 9. Thereby, imaging data of the object 9 at the measurement time t (2) is acquired. When the imaging data is acquired, the analysis processing unit 13 subsequently analyzes the imaging data and acquires the measurement position A (2) of the object 9 at the measurement time t (2). The acquired measurement position A (2) is input to the predicted position calculation unit 21.

  When the predicted position calculation unit 21 acquires the measurement position A (2) at the measurement time t (2), the predicted position calculation unit 21 and the measurement position A (2) at the measurement time t (1) acquired previously. Based on 1), the position of the object 9 at the measurement time t (3) is predicted and acquired as the predicted position B (3). The acquired predicted position B (3) is input to the prediction error acquisition unit 22.

<Measurement time t (3)>
When a certain time elapses from the measurement time t (2), as shown in FIG. 6, the third imaging unit 111 counting from the −X side passes over the mark M (measurement time t (3)). . Then, the imaging control unit 12 causes the imaging unit 111 immediately above the mark M to image the in-plane region of the object 9. Thereby, imaging data of the object 9 at the measurement time t (3) is acquired. When the imaging data is acquired, the analysis processing unit 13 subsequently analyzes the imaging data and acquires the measurement position A (3) of the object 9 at the measurement time t (3). The acquired measurement position A (3) is input to the predicted position calculation unit 21 and the prediction error acquisition unit 22.

  When the predicted position calculation unit 21 acquires the measurement position A (3) at the measurement time t (3), the predicted position calculation unit 21 and the measurement position A (2) at the time t (2) acquired earlier. ) And the position of the object 9 at the measurement time t (4) is predicted and acquired as a predicted position B (4). The acquired predicted position B (4) is input to the prediction error acquisition unit 22.

  On the other hand, when the prediction error acquisition unit 22 acquires the measurement position A (3) at the measurement time t (3), the prediction error acquisition unit 22 and the prediction position at the measurement time t (3) acquired previously. A difference from B (3) is calculated and acquired as a prediction error C (3) at measurement time t (3). The obtained prediction error C (3) is input to the parameter adjustment unit 23 and is accumulated in the evaluation unit 241.

  When the parameter adjustment unit 23 acquires the prediction error C (3) at the measurement time t (3), the parameter adjustment unit 23 is included in the prediction model used by the prediction position calculation unit 21 for prediction so that the prediction error C (3) approaches zero. Adjust processing parameters. When the parameter adjustment unit 23 adjusts the processing parameter, the predicted position calculation unit 21 thereafter predicts the position of the object 9 using a prediction model including the adjusted processing parameter.

<Measurement time t (i) (i = 4, 5,..., N−1)>
The processing performed at each of measurement time t (4) (state shown in FIG. 7), measurement time t (5),..., Measurement time t (n−1) (state shown in FIG. 8) This is the same as the processing performed at the measurement time t (3) described above.

  That is, when a certain amount of time elapses from the measurement time t (i−1), the i-th imaging unit 111 counting from the −X side passes over the mark M (measurement time t (i)). Then, the imaging control unit 12 causes the imaging unit 111 immediately above the mark M to image the in-plane region of the object 9. Thereby, imaging data of the object 9 at the measurement time t (i) is acquired. When the imaging data is acquired, the analysis processing unit 13 subsequently analyzes the imaging data and acquires the measurement position A (i) of the object 9 at the measurement time t (i). The acquired measurement position A (i) is input to the predicted position calculation unit 21 and the prediction error acquisition unit 22.

  When the predicted position calculation unit 21 acquires the measurement position A (i) at the measurement time t (i), the predicted position calculation unit 21 and the measurement position A at the time t (i−1) acquired previously. Based on (i-1), the position of the object 9 at the measurement time t (i + 1) is predicted and acquired as a predicted position B (i + 1). The acquired predicted position B (i + 1) is input to the prediction error acquisition unit 22.

  On the other hand, when the prediction error acquisition unit 22 acquires the measurement position A (i) at the measurement time t (i), the prediction position at the measurement time t (i) that has been acquired previously is obtained. A difference from B (i) is calculated and obtained as a prediction error C (i) at the measurement time t (i). The acquired prediction error C (i) is input to the parameter adjustment unit 23 and is accumulated in the evaluation unit 241.

  When the parameter adjustment unit 23 acquires the prediction error C (i) at the measurement time t (i), the parameter adjustment unit 23 is included in the prediction model used by the prediction position calculation unit 21 for prediction so that the prediction error C (i) approaches zero. Adjust processing parameters.

<Measurement time t (n)>
When a certain time elapses from the measurement time t (n−1), as shown in FIG. 9, the + X-side imaging unit 111 passes above the mark M (measurement time t (n)). Then, the imaging control unit 12 causes the imaging unit 111 immediately above the mark M to image the in-plane region of the object 9. Thereby, imaging data of the object 9 at the measurement time t (n) is acquired. When the imaging data is acquired, the analysis processing unit 13 subsequently analyzes the imaging data and acquires the measurement position A (n) of the object 9 at the measurement time t (n). The acquired measurement position A (n) is input to the predicted position calculation unit 21 and the prediction error acquisition unit 22.

  When the predicted position calculation unit 21 acquires the measurement position A (n) at the measurement time t (n), the prediction position calculation unit 21 and the measurement position A at the time t (n−1) acquired previously. Based on (n−1), the position of the object 9 at the processing start time t (n + 1) is predicted and acquired as a predicted position B (n + 1). The acquired predicted position B (n + 1) is input to the output unit 24.

  On the other hand, when the prediction error acquisition unit 22 acquires the measurement position A (n) at the measurement time t (n), the prediction position at the measurement time t (n) that has been acquired previously is obtained. The difference from B (n) is calculated and acquired as the prediction error C (n) at the measurement time t (n). The acquired prediction error C (n) is input to the output unit 24.

  When the output unit 24 acquires the predicted position B (n + 1) of the object 9 at the processing start time t (n + 1), the output unit 24 outputs the value of the acquired predicted position B (n + 1) as the final predicted position. However, prior to outputting the final predicted position, the evaluation unit 241 determines that the group of accumulated prediction errors C (i) (i = 3, 4,..., N) (that is, measurement times t (3), t The prediction model is evaluated based on the prediction errors C (3), C (4),..., C (n)) of (4),. When the evaluation unit 241 gives a negative evaluation to the prediction model, the output unit 24 performs predetermined error processing instead of (or in combination with) the output of the final prediction position.

<Process start time t (n + 1)>
When a certain time elapses from the measurement time t (n), as shown in FIG. 10, the processing head 102 arrives immediately above the processing start position and starts processing the object 9 (processing start time t ( n + 1)). However, the control unit 105 causes the correction unit 104 to output the target object at the final predicted position output from the position prediction device 1 (that is, the process start time t (n + 1)) until the process start time t (n + 1) arrives. 9 predicted position), an instruction is given to correct the processing position by the processing head 102 as necessary. In response to the instruction, the correction unit 104 corrects the processing position of the processing head 102 at the processing start time t (n + 1), thereby accurately processing the target position of the object 9 at the processing start time t (n + 1). Will be given.

<4. Effect>
According to the first embodiment, the target of the processing start time is based on the measurement position of the target 9 at each of the plurality of measurement times after the target 9 and the processing head 102 start moving relatively. Predict the position of the object 9. According to this configuration, the position of the object 9 can be predicted in consideration of the displacement state of the object 9 after the start of relative movement. Therefore, the position of the object 9 when the process is executed can be predicted with sufficient accuracy. As a result, the target position of the object 9 can be accurately processed. For example, when the processing head 102 is a drawing head, the target position on the substrate can be accurately irradiated with light.

  In particular, according to the first embodiment, the position of the object 9 at the measurement time is predicted using the prediction model. Then, the processing parameter included in the prediction model is adjusted so that the difference between the obtained predicted position and the actual measurement position approaches zero. According to this configuration, the processing parameters included in the prediction model are adjusted according to the displacement state of the object 9 after the start of relative movement, so that the prediction accuracy can be further improved.

  In particular, according to the first embodiment, the imaging unit 11 images the object 9 at each of a plurality of measurement times, and the measurement position of the object 9 at each measurement time is based on the acquired imaging data. Identified. According to this configuration, the measurement position of the object 9 at each of a plurality of measurement times can be acquired with a simple configuration.

<II. Second Embodiment>
<1. Position Prediction Device 2>
A position prediction apparatus 2 according to the second embodiment will be described. In the following description, descriptions of points that are not different from the position prediction apparatus 1 according to the first embodiment are omitted, and the same elements are denoted by the same reference numerals.

  Similar to the position prediction apparatus 1 according to the first embodiment, the position prediction apparatus 2 is mounted on the processing apparatus 100 (see FIG. 1), and processing time at which processing from the processing head 102 to the object 9 is executed. The relative position of the object 9 with respect to the processing head 102 is predicted. Here again, it is assumed that the processing time to be predicted by the position prediction device 2 is the time at which the processing head 102 starts processing the object 9 (processing start time).

  The configuration of the position prediction apparatus 2 will be described with reference to FIG. 11 in addition to FIGS. FIG. 11 is a block diagram illustrating a configuration included in the position prediction device 2.

  The position prediction device 2 includes a measurement unit 10 that measures the position of the object 9 and a prediction unit 50 that predicts the position of the object 9 based on the measurement position of the object 9 acquired by the measurement unit 10. The configuration of the measurement unit 10 is the same as that of the measurement unit 10 included in the position prediction apparatus 1 according to the first embodiment.

  The prediction unit 50 predicts the position of the object 9 at the processing start time based on the measurement position of the object 9 at each of a plurality of measurement times acquired by the measurement unit 10. Specifically, the prediction unit 50 includes a plurality of (for example, three in this embodiment) prediction position calculation units 51x, 51y, and 51z, a prediction error acquisition unit 52, a parameter adjustment unit 53, and an output. Part 54.

<Predicted position calculation units 51x, 51y, 51z>
The process performed by each of the three predicted position calculation units 51x, 51y, 51z is the same as the process performed by the predicted position calculation unit 21 described above. That is, each of the three predicted position calculation units 51x, 51y, and 51z uses the prediction model including the processing parameters to set the position of the object 9 at the target time to two or more positions before the target time (this In the embodiment, the prediction is performed based on the measurement position of the object 9 at each of the two measurement times. However, as in the first embodiment, in this embodiment, all of the third and subsequent measurement times and the processing start time are set as the target times.

  However, the three predicted position calculation units 51x, 51y, and 51z predict the position of the object 9 using different prediction models. In the following, the prediction model used by the first predicted position calculation unit 51x is also referred to as “first prediction model”, the prediction model used by the second predicted position calculation unit 51y is also referred to as “second prediction model”, and the third The prediction model used by the predicted position calculation unit 51z is also referred to as a “third prediction model”.

  Note that the prediction model used by each of the predicted position calculation units 51x, 51y, and 51z may be any type as in the first embodiment. For example, the first prediction model employs a prediction model that assumes that the object 9 vibrates at a relatively long period in a direction orthogonal to the traveling direction in the horizontal plane, and the second prediction model uses the object 9 Adopts a prediction model that is assumed to vibrate at a relatively short period in a direction orthogonal to the traveling direction in the horizontal plane, and the third prediction model is a direction in which the object 9 is parallel to the traveling direction in the horizontal plane. It is possible to adopt a prediction model that is assumed to vibrate.

<Prediction error acquisition unit 52>
The process performed by the prediction error acquisition unit 52 is substantially the same as the process performed by the prediction error acquisition unit 22 described above. However, the prediction error acquisition unit 52 has three prediction positions (that is, a prediction position predicted by the first prediction position calculation unit 51x using the first prediction model and a second prediction position calculation) for each measurement time. The prediction position predicted by the unit 51y using the second prediction model and the prediction position predicted by the third prediction position calculation unit 51z using the third prediction model) are input. Therefore, the prediction error acquisition unit 52 calculates the difference between the measurement position of the object 9 at a certain measurement time and each of the three prediction positions of the object 9 at the measurement time, thereby calculating the three predictions. An error (a prediction error of the first prediction model at the measurement time, a prediction error of the second prediction model at the measurement time, and a prediction error of the third prediction model at the measurement time) is acquired.

<Parameter adjustment unit 53>
The process performed by the parameter adjustment unit 53 is almost the same as the process performed by the parameter adjustment unit 23 described above. However, when three prediction errors are input to the parameter adjustment unit 53 for each measurement time, the parameter adjustment unit 53 causes the first prediction model to approach the prediction error of the first prediction model close to zero. The processing parameters included in the are adjusted. Furthermore, the parameter adjustment unit 53 adjusts the processing parameter included in the second prediction model so that the prediction error of the second prediction model approaches zero. Furthermore, the parameter adjustment unit 53 adjusts the processing parameter included in the third prediction model so that the prediction error of the third prediction model approaches zero.

  In this case as well, when the parameter adjustment unit 53 changes the processing parameter, each predicted position calculation unit 51x, 51y, 51z thereafter uses the prediction model including the changed processing parameter to determine the target 9 Predict location. And according to the prediction error calculated about each obtained prediction position, the parameter adjustment part 53 again adjusts the process parameter contained in each prediction model. In this way, each time a prediction error for each measurement time is acquired, the adjustment of the processing parameters included in each prediction model is repeated, and each prediction model is used for the object 9 after the start of relative movement. It will be according to the displacement situation.

<Output 54>
The output unit 54 outputs the final predicted position based on the predicted position of the object 9 at the processing start time acquired by the predicted position calculation units 51x, 51y, and 51z using different prediction models.

  The output unit 54 includes an evaluation unit 541 that evaluates each of the three prediction models used by the prediction position calculation units 51x, 51y, and 51z for prediction. The evaluation unit 541 accumulates prediction errors at each of a plurality of measurement times for each prediction model, and evaluates each prediction model based on the accumulated group of prediction error values. Specifically, the evaluation unit 541 evaluates the first prediction model based on the prediction error of the first prediction model at each of a plurality of measurement times. Furthermore, the evaluation unit 541 evaluates the second prediction model based on the prediction error value of the second prediction model at each of the plurality of measurement times. Furthermore, the evaluation unit 541 evaluates the third prediction model based on the prediction error value of the third prediction model at each of the plurality of measurement times. For example, the evaluation unit 541 has converged to zero when the prediction error is within a predetermined allowable range (for example, a range whose absolute value is equal to or less than a predetermined threshold) after a certain time, for example. A positive evaluation of the prediction model.

  The output unit 54 selects the prediction model given a positive evaluation by the evaluation unit 541 as the optimal prediction model, and obtains the predicted position of the object 9 at the processing start time acquired using the optimal prediction model as the final prediction model. Output as predicted position. In addition, when there are a plurality of prediction models given a positive evaluation, the output unit 54, for example, among the plurality of prediction models given a positive evaluation, the prediction having the smaller prediction error acquired last. The model may be selected as the optimal prediction model. Alternatively, the output unit 54 selects all of the plurality of prediction models given a positive evaluation as the optimal prediction model, and averages the predicted positions at the processing start time acquired using each of the plurality of optimal prediction models. The value may be output as the final predicted position.

  When the evaluation unit 541 gives a negative evaluation to all prediction models, the output unit 54 performs predetermined error processing instead of (or in combination with) the output of the final prediction position. Do.

<2. Flow of processing>
The flow of the process (position prediction process) in which the position prediction device 2 predicts the position of the object 9 at the process start time will be specifically described with reference to FIGS. 12 to 18 are schematic diagrams for explaining the flow of the position prediction process, and each of the processes at different times is schematically shown. In FIGS. 12 to 18, the processing performed by each processing unit 51, 52, 53 at each time is schematically shown as one block.

<Measurement time t (1)>
At a measurement time t (1) when a certain time has elapsed from the movement start time, as shown in FIG. 12, the imaging control unit 12 applies an in-plane region of the object 9 to the imaging unit 111 directly above the mark M. To image. And the analysis process part 13 analyzes the acquired imaging data, and acquires the measurement position A (1) of the target object 9 of measurement time t (1). The acquired measurement position A (1) is input to the three predicted position calculation units 51x, 51y, and 51z.

<Measurement time t (2)>
At the measurement time t (2), as illustrated in FIG. 13, the imaging control unit 12 causes the imaging unit 111 immediately above the mark M to image the in-plane region of the object 9. And the analysis process part 13 analyzes the acquired imaging data, and acquires the measurement position A (2) of the target object 9 of measurement time t (2). The acquired measurement position A (2) is input to the three predicted position calculation units 51x, 51y, 51z.

  When each of the three predicted position calculation units 51x, 51y, 51z acquires the measurement position A (2) at the measurement time t (2), the measurement position A (2) and the previously acquired measurement time are obtained. Based on the measurement position A (1) at t (1), the position of the object 9 at the measurement time t (3) is predicted. In other words, the first predicted position calculation unit 51x uses the first prediction model, and based on the two measurement positions A (2) and A (1), the position of the object 9 at the measurement time t (3). Is obtained as a predicted position Bx (3). In addition, the second predicted position calculation unit 51y uses the second prediction model to determine the position of the object 9 at the measurement time t (3) based on the two measurement positions A (2) and A (1). Is obtained as a predicted position By (3). In addition, the third predicted position calculation unit 51z uses the third predicted model to determine the position of the object 9 at the measurement time t (3) based on the two measurement positions A (2) and A (1). Is obtained as a predicted position Bz (3). The three predicted positions Bx (3), By (3), Bz (3) acquired by the three predicted position calculators 51x, 51y, 51z are input to the prediction error acquiring unit 52.

<Measurement time t (3)>
At the measurement time t (3), as illustrated in FIG. 14, the imaging control unit 12 causes the imaging unit 111 immediately above the mark M to image the in-plane region of the object 9. And the analysis process part 13 analyzes the acquired imaging data, and acquires the measurement position A (3) of the target object 9 of measurement time t (3). The acquired measurement position A (3) is input to each of the three predicted position calculation units 51x, 51y, 51z and the prediction error acquisition unit 52.

  When each of the three predicted position calculation units 51x, 51y, 51z acquires the measurement position A (3) at the measurement time t (3), the measurement position A (3) and the previously acquired measurement time are obtained. Based on the measurement position A (2) at t (2), the position of the object 9 at the measurement time t (4) is predicted. The three predicted positions Bx (4), By (4), Bz (4) acquired by the three predicted position calculators 51x, 51y, 51z are input to the prediction error acquiring unit 52.

  On the other hand, when the prediction error acquisition unit 52 acquires the measurement position A (3) at the measurement time t (3), the prediction error acquisition unit 52 includes the measurement position A (3) and the previously acquired measurement time t (3). Differences between the predicted positions Bx (3), By (3), and Bz (3) are calculated. That is, the prediction error acquisition unit 52 calculates the difference between the prediction position Bx (3) acquired using the first prediction model and the measurement position A (3), and the prediction error Cx (3 of the first prediction model ) Get as. Further, the difference between the predicted position By (3) acquired using the second prediction model and the measurement position A (3) is calculated and acquired as the prediction error Cy (3) of the second prediction model. Further, the difference between the predicted position Bz (3) acquired using the third prediction model and the measurement position A (3) is calculated and acquired as the prediction error Cz (3) of the third prediction model. The three obtained prediction errors Cx (3), Cy (3), Cz (3) are input to the parameter adjustment unit 53 and accumulated in the evaluation unit 541.

  When the parameter adjustment unit 53 acquires the three prediction errors Cx (3), Cy (3), and Cz (3), the parameter adjustment unit 53 brings the prediction errors Cx (3), Cy (3), and Cz (3) close to zero. As described above, the processing parameters included in each prediction model are adjusted. That is, the parameter adjustment unit 53 adjusts the processing parameter included in the first prediction model so that the prediction error Cx (3) of the first prediction model approaches zero. In addition, the processing parameters included in the second prediction model are adjusted so that the prediction error Cy (3) of the second prediction model approaches zero. In addition, the processing parameters included in the third prediction model are adjusted so that the prediction error Cz (3) of the third prediction model approaches zero. When the parameter adjustment unit 53 adjusts the processing parameter, each predicted position calculation unit 51x, 51y, 51z thereafter predicts the position of the target 9 using the adjusted processing parameter.

<Measurement time t (i) (i = 4, 5,..., N−1)>
The processing performed at each of measurement time t (4) (state shown in FIG. 15), measurement time t (5),..., Measurement time t (n−1) (state shown in FIG. 16) This is the same as the processing performed at the measurement time t (3) described above.

  That is, at the measurement time t (i), the imaging control unit 12 causes the imaging unit 111 immediately above the mark M to image the in-plane region of the object 9. And the analysis process part 13 analyzes the acquired imaging data, and acquires the measurement position A (i) of the target object 9 of measurement time t (i). The acquired measurement position A (i) is input to each of the three predicted position calculation units 51x, 51y, 51z and the prediction error acquisition unit 52.

  When each of the three predicted position calculation units 51x, 51y, 51z acquires the measurement position A (i) at the measurement time t (i), the measurement position A (i) and the previously acquired measurement time are obtained. Based on the measurement position A (i-1) at t (i-1), the position of the object 9 at the measurement time t (i + 1) is predicted. The three predicted positions Bx (i + 1), By (i + 1), Bz (i + 1) acquired by the three predicted position calculators 51x, 51y, 51z are input to the prediction error acquiring unit 52.

  On the other hand, when the prediction error acquisition unit 52 acquires the measurement position A (i) at the measurement time t (i), the prediction error acquisition unit 52 includes the measurement position A (i) and the previously acquired measurement time t (i). The difference between each of the predicted positions Bx (i), By (i), and Bz (i) is calculated, and the prediction errors Cx (i), Cy (i), Cz (i) at the measurement time t (i) are calculated. Get as. The three obtained prediction errors Cx (i), Cy (i), and Cz (i) are input to the parameter adjustment unit 53 and accumulated in the evaluation unit 541.

  When the parameter adjustment unit 53 acquires the three prediction errors Cx (i), Cy (i), and Cz (i), the parameter adjustment unit 53 brings the prediction errors Cx (i), Cy (i), and Cz (i) close to zero. As described above, the processing parameters included in each prediction model are adjusted.

<Measurement time t (n)>
At the measurement time t (n), as illustrated in FIG. 17, the imaging control unit 12 causes the imaging unit 111 immediately above the mark M to image the in-plane region of the object 9. And the analysis process part 13 analyzes the acquired imaging data, and acquires the measurement position A (n) of the target object 9 of measurement time t (n). The acquired measurement position A (n) is input to each of the three predicted position calculation units 51x, 51y, 51z and the prediction error acquisition unit 52.

  When each of the three predicted position calculation units 51x, 51y, 51z acquires the measurement position A (n) at the measurement time t (n), the measurement position A (n) and the previously acquired measurement time are obtained. Based on the measurement position A (n−1) of t (n−1), the position of the object 9 at the processing start time t (n + 1) is predicted. The three predicted positions Bx (n + 1), By (n + 1), Bz (n + 1) acquired by the three predicted position calculators 51x, 51y, 51z are input to the output unit 54.

  On the other hand, when the prediction error acquisition unit 52 acquires the measurement position A (n) at the measurement time t (n), the prediction error acquisition unit 52 includes the measurement position A (n) and the previously acquired measurement time t (n). The difference between each of the predicted positions Bx (n), By (n), and Bz (n) is calculated, and the prediction errors Cx (n), Cy (n), Cz (n) at the measurement time t (n) are calculated. Get as. The three obtained prediction errors Cx (n), Cy (n), and Cz (n) are input to the output unit 54.

  When the output unit 54 acquires the three predicted positions Bx (n + 1), By (n + 1), Bz (n + 1) of the object 9 at the processing start time t (n + 1), the output unit 54 acquires the three predicted positions Bx ( n + 1), By (n + 1), Bz (n + 1) are output based on the final predicted position. Specifically, first, the evaluation unit 541 applies a group of prediction errors Cx (i), Cy (i), Cz (i) (i = 3, 4,..., N) accumulated for each prediction model. Based on this, each prediction model is evaluated. The output unit 54 selects, for example, the prediction model given a positive evaluation by the evaluation unit 541 as the optimal prediction model, and the object at the processing start time t (n + 1) acquired using the optimal prediction model. The predicted position of 9 is output as the final predicted position. FIG. 17 illustrates a state in which the first prediction model is selected as the optimal prediction model, and the prediction position Bx (n + 1) of the target 9 obtained using the first prediction model is output as the final prediction position.

<Process start time t (n + 1)>
At the processing start time t (n + 1), as shown in FIG. 18, the processing head 102 arrives immediately above the processing start position and starts processing the object 9. However, the control unit 105 sets the processing position by the processing head 102 to the correction unit 104 based on the final predicted position output from the position prediction device 2 as necessary until the processing start time t (n + 1) arrives. To give corrections. In response to the instruction, the correction unit 104 corrects the processing position of the processing head 102 at the processing start time t (n + 1), thereby accurately processing the target position of the object 9 at the processing start time t (n + 1). Will be given.

<3. Effect>
According to the second embodiment, the same effect as in the first embodiment can be obtained. Furthermore, according to the second embodiment, the position of the object 9 is predicted using each of a plurality of different prediction models, and a prediction model given a positive evaluation is used among the plurality of prediction models. The predicted position acquired in this way is output as the final predicted position. According to this configuration, it is possible to widely cope with various displacement situations of the object 9, and high versatility is realized. In addition, since the predicted position acquired using the prediction model that best matches the displacement state of the object 9 after the start of relative movement is output as the final predicted position, high prediction accuracy can be stably achieved. Can be realized.

<III. Modification>
In the position prediction devices 1 and 2 according to each of the above embodiments, the processing time to be predicted by the position prediction devices 1 and 2 is the time when the processing head 102 starts processing the object 9 (processing start time). However, in addition to (or instead of) the processing start time, the position prediction devices 1 and 2 may set an arbitrary time after the processing start time as a prediction target.

  Further, in the position prediction devices 1 and 2 according to each of the above-described embodiments, a plurality of position prediction processes may be performed during the period from the start of movement to the stop of movement of the stage 101 from the movement start state. Good. For example, a plurality of processing regions are formed in the surface of the object 9, and the first mark M1 is attached near the + X side end of the first processing region, so that the second processing is performed. It is assumed that the second mark M2 is attached near the + X side end of the region. In this case, as shown in FIG. 19, first, the mark M1 is detected, the position of the object 9 at the processing start time with respect to the first processing region is predicted, the mark M2 is detected, and the second The position of the object 9 at the processing start time with respect to the processing area may be predicted.

  In the second embodiment, the analysis processing unit 13 may be included in each of the predicted position calculation units 51x, 51y, and 51z included in the prediction unit 50. In this case, the imaging data acquired by the imaging unit 111 is directly input to the prediction unit 50, and the imaging data is analyzed by the predicted position calculation units 51x, 51y, and 51z. Here, each predicted position calculation unit 51x, 51y, 51z may execute analysis processing of different algorithms to specify the position of the mark M from the imaging data. In this case, the parameter adjustment unit 53 includes a process included in the prediction model used by the prediction position calculation unit so that the prediction error calculated for the prediction position acquired by each prediction position calculation unit 51x, 51y, 51z approaches zero. While adjusting a parameter, it is good also as a structure which also adjusts the processing parameter contained in the analysis algorithm of the imaging data which the said predicted position calculation part uses.

  Further, in the processing apparatus 100 on which the position prediction apparatuses 1 and 2 according to each of the above embodiments are mounted, the processing head 102 and the object 9 are relatively moved when the stage 101 is driven by the drive mechanism 103. Although it is configured to be moved, when the processing head 102 is moved with respect to the fixed stage 101 (or when the stage 101 and the processing head 102 are moved together), the processing head 102 and the object 9 are moved. And may be moved relatively.

  In the above embodiment, the measurement unit 10 measures the position of the object 9 by detecting the alignment mark M attached to the object 9, but the detection object is Such a mark M is not necessarily required. That is, the detection target may be any part as long as it can be a characteristic part on the object 9. For example, the detection target may be an end face edge of the target 9, An existing pattern may be used.

  In the above embodiment, the measurement unit 10 measures the position of the object 9 by imaging the object 9 with the imaging unit 11 and detecting the position of the mark M or the like to be detected. However, the mode of measuring the position of the object 9 is not necessarily limited to this. For example, various displacement meters and various sensors (for example, the object 9 or the stage 101 are set as the measurement target, the laser beam is emitted toward the measurement target, the reflected light is received, and the reflected light is output. You may measure the position of the target object 9 using the interference type laser length measuring device etc. which measure the position of a measuring object from interference with incident light.

  In the above-described embodiment, the imaging unit 11 uses the plurality of imaging units 111 to sequentially acquire imaging data of the objects 9 at a plurality of measurement times. However, the imaging unit 11 does not necessarily include a plurality of imaging units 11. The imaging unit 111 may not be provided. For example, an optical system capable of selecting a plurality of optical paths as an optical path incident on the area image sensor from the light source through the object 9 is formed by a mirror, a lens, etc., and one light source and one area image sensor. Thus, a function equivalent to that of the plurality of imaging units 111 may be realized.

  In the above-described embodiment, a mechanism is provided for moving each imaging unit 111 along the axis (that is, the axis fixed to the processing head 102 and extending along the X axis). The separation interval of each imaging unit 111 may be adjustable.

  Further, in the above embodiment, the correction unit 104 is based on the amount of displacement of the object 9 at the processing time (specifically, based on the position of the object 9 at the processing time predicted by the position prediction device 1). Although the processing position of the processing head 102 at the processing time is corrected according to the acquired positional deviation amount of the target object 9, the correction unit 104 corrects the positional deviation amount of the target object 9 at the processing time. Accordingly, the drive mechanism 103 may be made to correct the position of the stage 101 at the processing time.

  In addition, the processing apparatus 100 on which the position prediction apparatuses 1 and 2 according to the above-described embodiments are mounted, for example, scans the photosensitive material on the substrate with spatially modulated light, thereby directly forming a pattern on the photosensitive material. Although the substrate processing apparatus to be exposed (specifically, the drawing apparatus) is described, as described above, the processing apparatus 100 is not necessarily limited to the drawing apparatus.

  For example, the processing apparatus 100 may be a substrate processing apparatus (specifically, an exposure apparatus) that exposes a photosensitive material formed on a substrate in a planar shape. In this case, the object 9 is a substrate on which a layer of a photosensitive material such as a resist is formed, and the processing head 102 is an exposure head that irradiates the substrate with light using a light source and a photomask. In the processing apparatus 100 that is an exposure apparatus, the exposure head that is the processing head 102 moves relative to the substrate that is the object 9 while moving light (specifically, for example, with a photomask). By irradiating the shaped light), a pattern (circuit pattern) is exposed to the substrate.

  Further, for example, the processing apparatus 100 may be a substrate processing apparatus (specifically, an inspection apparatus) that inspects a pattern or the like formed on the substrate. In this case, the object 9 is a substrate to be inspected, and the processing head 102 is, for example, an inspection head for imaging a partial region of the substrate and inspecting a pattern shape or the like formed on the surface thereof. In the processing apparatus 100 that is an inspection apparatus, the inspection head that is the processing head 102 images each position of the substrate while moving relative to the substrate that is the object 9, and analyzes the obtained imaging data. As a result, the pattern formed on the substrate is inspected.

  Further, for example, the processing apparatus 100 may be a CTP (Computer To Plate) apparatus that outputs a printing plate from data. In this case, the object 9 is a plate material such as an aluminum plate to be a printing plate. Further, the processing head 102 is, for example, a drawing head that irradiates a plate material with light and forms an image on the plate material. In the processing apparatus 100 which is a CTP apparatus, the drawing head which is the processing head 102 irradiates light to each position of the plate material while moving relative to the plate material which is the object 9, thereby An image is formed. As a result, a printing plate is obtained.

  For example, the processing apparatus 100 may be an image forming apparatus that forms an image or the like on a medium such as printing paper. In this case, the object 9 is, for example, printing paper. The processing head 102 is, for example, an inkjet head that forms an image by ejecting ink droplets onto a printing paper. In the processing apparatus 100 that is an image forming apparatus, an ink jet head that is a processing head 102 ejects ink droplets to each position of the printing paper while moving relative to the printing paper that is the object 9. An image is formed on the printing paper.

  In the above embodiment and each of the modifications described above, when the object 9 is a substrate, the substrate is included in, for example, a semiconductor substrate, a printed substrate, a color filter substrate, a liquid crystal display device, or a plasma display device. It may be various substrates such as a flat panel display glass substrate, an optical disk substrate, and a solar cell panel.

DESCRIPTION OF SYMBOLS 1, 2 Position prediction apparatus 10 Measurement part 11 Imaging unit 111 Imaging part 12 Imaging control part 13 Analysis processing part 20, 50 Prediction part 21, 51x, 51y, 51z Prediction position calculation part 22, 52 Prediction error acquisition part 23, 53 Parameter Adjustment unit 24, 54 Output unit 100 Processing device 101 Stage 102 Processing head 103 Drive mechanism 9 Object

Claims (11)

In a processing apparatus comprising a processing head for processing a target while moving relative to the target, the processing head is at a processing time at which the processing head executes the processing on the target. A position prediction apparatus for predicting the position of the object,
Each of a plurality of times within a time before the processing time after the object and the processing head start to move relative to each other is a measurement time, and the object at each of the plurality of measurement times A measuring unit for measuring the position of an object;
A prediction unit that predicts the position of the object at the processing time based on the measurement position of the object at each of the plurality of measurement times;
A position prediction apparatus comprising: The position prediction apparatus according to claim 1,
The prediction unit is
Using a prediction model including one or more measurement times of the plurality of measurement times and the processing time as a target time, and including a processing parameter, the position of the target at the target time is set before the target time. A predicted position calculation unit that predicts based on the measurement position of the object at each of two or more measurement times;
When the predicted position calculation unit predicts the position of the target object using any one of the plurality of measurement times as a target time, the predicted position obtained and the target object at the measurement time set as the target time. A prediction error acquisition unit that calculates a difference from the measurement position and acquires the difference as a prediction error;
A parameter adjustment unit for adjusting the processing parameter so that the prediction error approaches zero;
The predicted position obtained when the predicted position calculation unit predicts the position of the object using the prediction model including the processing parameter adjusted by the parameter adjustment unit with the processing time as the target time. Output as the final predicted position,
A position prediction apparatus comprising: The position prediction apparatus according to claim 2,
The prediction unit is
A plurality of the predicted position calculation unit,
The plurality of predicted position calculation units predict the position of the object using different prediction models,
The output unit is
An evaluation unit that evaluates each of the plurality of prediction models based on the prediction error calculated for the prediction position acquired using each of the plurality of prediction models;
Further comprising
Position that outputs the predicted position of the object at the processing time obtained as a final predicted position, obtained using a prediction model given a positive evaluation by the evaluation unit among the plurality of predicted models Prediction device. The position prediction apparatus according to any one of claims 1 to 3,
The measurement unit is
An imaging unit fixed to the processing head and imaging the object;
An imaging control unit that causes the imaging unit to image the object at each of the plurality of measurement times;
Analyzing the imaging data acquired by the imaging unit, identifying the position of the object at the measurement time at which the imaging data was acquired, and obtaining the measurement position of the object;
A position prediction apparatus comprising: The position prediction apparatus according to claim 4,
The imaging unit is
A plurality of imaging units,
The plurality of imaging units are
With respect to the movement direction in which the processing head is moved relative to the object, the processing head is arranged along the movement direction on the downstream side of the processing head,
The imaging control unit
Each of the plurality of imaging units causes the imaging unit to image the characteristic portion at a time when the imaging unit reaches above the characteristic portion on the object.
Position prediction device. In a processing apparatus comprising a processing head for processing a target while moving relative to the target, the processing head is at a processing time at which the processing head executes the processing on the target. A position prediction method for predicting the position of the object,
a) After the target object and the processing head are relatively started to move, each of a plurality of times within a time before the processing time is set as a measurement time, and at each of the plurality of measurement times, Measuring the position of the object;
b) predicting the position of the object at the processing time based on the measurement position of the object at each of the plurality of measurement times acquired in the step a);
A position prediction method comprising: The position prediction method according to claim 6,
Step b)
b1) Using one or more measurement times of the plurality of measurement times and the processing time as a target time, and using a prediction model including processing parameters, the position of the target at the target time is determined from the target time. Predicting based on the measurement position of the object at each of the previous two or more measurement times;
b2) In the step b1), when the position of the object is predicted using any one of the plurality of measurement times as the target time, the obtained predicted position and the target at the measurement time set as the target time Calculating a difference from the measurement position of the object and obtaining it as a prediction error;
b3) adjusting the processing parameters so that the prediction error approaches zero;
b4) Obtained when the position of the object was predicted using the prediction model including the processing parameter adjusted in step b3) with the processing time as the target time in step b1). Outputting a predicted position as a final predicted position;
A position prediction method comprising: The position prediction method according to claim 7,
In the step b1), the position of the object at the target time is predicted using each of a plurality of different prediction models,
Step b)
b5) evaluating each of the plurality of prediction models based on the prediction error calculated for the prediction position acquired using each of the plurality of prediction models;
Further comprising
In the step b4), the predicted position of the object at the processing time obtained using the prediction model given a positive evaluation in the step b5) is output as the final predicted position.
Location prediction method. The position prediction method according to any one of claims 6 to 8,
Step a)
a1) causing the imaging unit fixed to the processing head to image the object at each of the plurality of measurement times;
a2) analyzing the imaging data acquired by the imaging unit, identifying the position of the object at the measurement time at which the imaging data was acquired, and acquiring it as the measurement position of the object;
A position prediction method comprising: The position prediction method according to claim 9,
The imaging unit is
A plurality of imaging units,
The plurality of imaging units are
With respect to the movement direction in which the processing head is moved relative to the object, the processing head is arranged along the movement direction on the downstream side of the processing head,
In the step a1),
Each of the plurality of imaging units causes the imaging unit to image the characteristic portion at a time when the imaging unit reaches above the characteristic portion on the object.
Location prediction method. A processing head that irradiates the substrate with light to expose the pattern;
A drive mechanism for relatively moving the substrate and the processing head;
A position prediction unit that predicts the position of the substrate relative to the processing head at a processing time when the processing head irradiates the substrate with light;
A correction unit that corrects an irradiation position of light from the processing head based on the position of the substrate at the processing time predicted by the position prediction unit;
With
The position prediction unit is
After the substrate and the processing head start to move relatively, each of a plurality of times within a time before the processing time is a measurement time, and the object at each of the plurality of measurement times A measuring unit for measuring the position of
A prediction unit that predicts the position of the object at the processing time based on the measurement position of the object at each of the plurality of measurement times;
A substrate processing apparatus comprising:

Images