Photoelectric encoder with spherical lens

Abstract

This disclosure relates to an optoelectronic encoder with a spherical lens. An image-based optoelectronic encoder for a joint is provided, the encoder having a readhead for a first portion of the joint; and an optical system including a spherical lens for a second portion of the joint. The spherical lens has a front surface that is transparent to measurement light, and a constructed rear surface that is at least partially reflective. The spherical lens has a refractive index of at least approximately 2 or an equal radial variation for the measurement light, such that a beam of measurement light reflected from a point on the rear surface of the spherical lens within the spherical lens is refracted at the front surface of the spherical lens to form at least approximately parallel beams. The readhead and the spherical lens are rotatable relative to each other in at least two degrees of freedom. Such rotational positions of the readhead and the spherical lens relative to each other are determined with respect to at least two degrees of freedom using an image of at least a portion of the rear surface, by image evaluation of the structure of the imaged rear surface.

Description

Technical Field

[0001] This invention relates to photoelectric encoders and methods for determining the rotational position of a joint. Background Technology

[0002] A series of joints equipped with encoders and links connecting the end effector to the base are central to many systems used in portable metrology today. The joints used today have a single degree of freedom (DOF), meaning they ideally restrict the motion between two links to translation or rotation, and are equipped with an encoder that measures this single DOF. It is assumed that the motion in the other DOFs is small enough to be negligible or at least highly reproducible so that it can be eliminated through calibration. This results in precision, expensive joints.

[0003] One approach to increasing the number of DOFs is to equip low-quality single-DOF joints with sensors that measure all associated parasitic DOFs in addition to the main DOF, thereby facilitating the correction of joint defects.

[0004] use Figure 1a , Figure 1b Examples of metering devices with encoders known in the art are given. However, these instruments have features such as... Figure 1a , Figure 1b Several drawbacks are described in the context of this. Summary of the Invention

[0005] This invention provides a novel optical encoder with several advantages compared to known optical encoders in the prior art. One object of this invention is to provide an encoder capable of measuring more than one rotational degree of freedom (DOF).

[0006] This invention relates to an optoelectronic encoder for joints, preferably for measuring the joints of an articulated arm or robot. The encoder includes a readhead having at least a first measuring light source (e.g., an LED) and an image sensor (particularly a CMOS sensor) for imaging the reflected measuring light. The encoder also includes an optical system having at least one spherical lens and an aperture. The readhead and the spherical lens are rotatable relative to each other in at least two degrees of freedom.

[0007] The spherical lens includes a front surface that is transparent to the measurement light and faces the reading head (and the light source), and a constructed rear surface that is at least partially reflective. Furthermore, the spherical lens has a (constant) refractive index for the measurement light that is at least approximately 2 (or close to 2), or a radially varying refractive index, particularly a stepped or gradient refractive index (the sum of which equals approximately 2). This refractive index, in any case, ensures that the beam of measurement light reflected from a point on the rear surface of the spherical lens within the lens is refracted at the front surface of the lens to form at least approximately parallel beams.

[0008] The rotational position of the readout head and the ball lens relative to each other can be determined with respect to the at least two degrees of freedom using an image of at least a portion of the rear surface, by evaluating the structure of the imaged rear surface, which is captured by an image sensor via an optical system.

[0009] The concept particularly offers the advantage of measurement independent of the axial distance between the read head and the ball lens when the measurement beam is collimated by the ball lens. Furthermore, lateral offset between the read head and the ball lens does not cause image shift / translation.

[0010] As a preferred option, the readhead includes at least one marker pattern that can be imaged on an image sensor using measurement light to compensate for changes in the translational position of the image sensor relative to the optical system by using the position in the image using the marker pattern. Using such a reference marker, for example, can compensate for positional changes in the image sensor caused by thermal effects. The marker is located, for example, at the center of the axis of the optical system. Alternatively, the marker (or multiple markers) can be placed around the perimeter of the optical axis, for example, at a point that has the advantage of fewer aberrations in the measured image.

[0011] Alternatively, the back surface is constructed with respect to reflectivity, and the measured light is incoherent. In other words, the back surface is constructed with respect to specular reflection (i.e., the pattern of reflective and absorptive portions, or the distribution of regions, areas, or points with higher reflectivity in regions of lower reflectivity).

[0012] Alternatively, the structures of the light source and the back surface are adjusted relative to each other in such a way that when the back surface is illuminated with measurement light, a speckled image is formed on the image sensor, for which the correlation of the measurement light is adapted to the surface roughness. Therefore, at least a portion of the image of the back surface is a speckled image.

[0013] In a preferred embodiment, the effective aperture is a virtual aperture formed at the center of the spherical lens by imaging the real aperture at one focal point of the relay lens in the optical system. Therefore, independent of the rotational position of the readhead and the spherical lens relative to each other, the principal ray of light reflected from a point on the rear surface (which is imaged on the image sensor) passes through the center of the spherical lens.

[0014] As an option, the marking pattern is located in the image plane of an optical system (which may include the relay lens mentioned above), such that the marking is imaged onto the image sensor. Alternatively or additionally, the system includes an additional lens that is part of the optical system and images the marking pattern and / or the marking pattern located at the center of the relay lens onto the image sensor.

[0015] Alternatively, as another option, at least one translational degree of freedom of the spherical lens and the reading head relative to each other can be determined. This is achieved by evaluating images of the same portion of the rear surface generated using different wavelengths, utilizing the dispersive properties of the spherical lens. For this purpose, the reading head includes a second light source with a second wavelength different from the first wavelength of the first measurement light source. Preferably, the two light sources can be turned on alternately.

[0016] Optionally, the encoder includes a beam splitter, preferably a prism, configured such that the emitted measurement light can be split into two paths, thereby enabling imaging of the rear surface from different perspectives.

[0017] The present invention also relates to a method for determining, particularly for measuring, the rotational position of a joint of an articulated arm or robot with respect to at least two, and particularly all three, angular degrees of freedom, the joint having an optoelectronic encoder according to the invention. In the process of said method, measuring light is emitted at a spherical lens, measuring light reflected back from the rear surface of the spherical lens through an optical system is received, the reflected measuring light is captured as an image using an image sensor, and the rotational position relative to said at least two, and particularly all three, degrees of freedom is determined by image evaluation of the image.

[0018] Optionally, the encoder includes a marker pattern as described above, which is imaged together with the back side in an image, and the image of the back side is evaluated with respect to the imaged marker pattern.

[0019] As an alternative, only regions of interest (ROIs) distributed within the image of the rear surface are evaluated for location determination, as if they were portions of the imaged rear surface. This option allows for, for example, higher frame rates or measurement rates compared to evaluation of the entire image. ROIs may optionally be uniformly distributed around the image and / or have a rectangular shape. Alternatively, each ROI may include a marker pattern.

[0020] As another option, the emission of the measurement light source and the capture of the reflected measurement light are synchronized relative to each other.

[0021] The encoder of the present invention preferably enables self-calibration or provides a self-calibration function. As for calibration purposes (e.g., high frame rates do not have high priority), then even in the measurement mode that only evaluates the ROI as described above, the entire image is preferably evaluated for self-calibration, or the encoder is calibrated by evaluating the image of the back surface as a whole. Attached Figure Description

[0022] Hereinafter, the encoder and method according to the invention will be described in more detail by way of example only, with reference to exemplary embodiments schematically depicted in the accompanying drawings.

[0023] More specifically, in the diagram:

[0024] Figure 1a , Figure 1b A measuring instrument with an articulated arm, specifically implemented as a prior art, is schematically shown.

[0025] Figure 2 A joint with an optoelectronic encoder according to the present invention is schematically shown, the optoelectronic encoder including a spherical lens as a scale.

[0026] Figure 3 The spherical lens and the measuring beam are depicted;

[0027] Figure 4 A simplified optical system with a spherical lens, an image sensor, another lens, and an optical measurement path is shown.

[0028] Figure 5a , Figure 5b This illustrates another example of a simplified optical system;

[0029] Figure 6a , Figure 6b Two examples of images generated by the encoder's image sensor are shown;

[0030] Figure 7 Another example of a measurement image is shown; and

[0031] Figure 8 A simplified schematic diagram illustrates the further development of image-based multi-DOF encoders. Detailed Implementation

[0032] Figure 1a and Figure 1bA measuring instrument is shown, specifically implemented as a hinged arm 100 known in the prior art. The arm 100 includes a plurality of segments 102 connected to a base 101 and interconnected by corresponding joints 104. The joints 104 enable rotation about (only) one axis, such as... Figure 1a As indicated by the middle arrows Rz, Ry1, and Ry2. Using this measuring arm 100, the dimensions of workpieces, etc., can be measured by contacting the workpiece with the arm's probe 103 and reading the corresponding rotational positions Rz, Ry1, or Ry2 of each single-axis joint 104. Figure 1b The angle encoder 106 within the corresponding joint 104 depicted combines these positions with the known length of segment 102 to determine the position of probe 103. Figure 1b A cross-section of the joint 104 supported by the bearing 109 is shown, wherein the end of the segment 102 is connected to the base 101 (or another segment) within the frame structure 105.

[0033] In the arm 100 with improved measurement accuracy, in addition to the encoder 106 for determining rotation about the intended rotation axis Rz, Ry1, or Ry2, the joint 104 also includes additional position sensors, such as a distance sensor 107 or a tilt sensor 108, for determining unintended movements of the joint 104. That is, such a single-degree-of-freedom (DOF) joint 104 can be equipped with sensors 107, 108 that measure all relevant parasitic DOFs, in addition to the angle encoder 106 that measures the main DOF, thereby facilitating the correction of joint defects.

[0034] Therefore, such measuring devices (or similar robotic arms) as known in the art have the disadvantage of using joints 104 such that each joint 104 provides only one DOF, thus requiring several joints 104 (and segments 102) to enable probe 102 to achieve more than one DOF. Furthermore, additional effort is required to reduce or determine measurement errors caused by structural defects in the joints 104, for example, by implementing additional sensors 107, 108 as shown.

[0035] Figure 2 A joint 1 with a photoelectric encoder 2 according to the present invention is shown. (As shown in the image) Figure 1a , Figure 1b In contrast to the prior art's 1-DOF or single-axis joint shown, joint 1 is a multi-axis joint, enabling at least 2-DOF or all three rotational DOFs depicted: rotation Rz about the z-axis, rotation Rey about the y-axis, and rotation Rx about the x-axis. That is, the two connecting portions 1a and 1b of joint 1 (or the two arm segments 102, see...) Figure 1a It can rotate around all axes.

[0036] An encoder 2 is provided to measure the corresponding rotational positions Rz, Ry, and Rx. The encoder 2 includes a measuring head or reading head 5 located in the first part 1a of the joint 1, and a 3D measuring instrument in the form of a ball lens 3 located in the mating part 1b of the joint. Therefore, the reading head 5 and the ball lens 3 are capable of rotating relative to each other, and rotating relative to each other when the joint 1 rotates. The rotational position or rotational angle is measured using an image of the rear surface 4 of the lens (i.e., the surface not directly opposite the reading head / the surface not facing the reading head, or in other words, the more distant surface / side / hemispherical surface of the ball 3 as seen from the reading head).

[0037] The image is generated using a measurement beam 7 from a light source 6 of the readhead 5, directed towards the front surface (the surface facing the readhead 5) of the spherical lens 3. The front surface and the spherical lens 3 are transparent to the measurement light 7, while the rear surface 4 is at least partially reflective. Therefore, at least a portion of the measurement light 7 is reflected back, re-enters the readhead 5 through the aperture 8, and is guided by the beam splitter 10 to the image sensor 11 (e.g., a CMOS or CCD sensor). Preferably, image acquisition by the image sensor 11 and light emission from the light source 6 are synchronized.

[0038] Therefore, at least a portion of the rear surface 4 of the sphere 3 is imaged onto the image sensor 11 through an optical device comprising at least an aperture 8 and a spherical lens 3. In other words, one hemisphere of the sphere 3 is transparent, and the other hemisphere is reflective and, for example, sandblasted and / or coated. The transparent side acts as a lens for the imaging system.

[0039] To compensate for drift of the image sensor 11 caused, for example, by thermal effects (heat), the optical system includes markers or marker patterns 9 in the optical path. In other words, the (e.g., thermally induced) translation, rotation, and / or deformation of the image sensor 11 is corrected by the markers 9 imaged onto the sensor 11 defining a zero point. The purpose of the markers 9 is to have an image with a pattern for compensating for mechanical instability and thermal expansion of the readhead 5. For example, the markers 9 are plates with reflective lithographic patterns and an absorption layer on the rear side. The optical system may include an additional lens for imaging the markers 9, for example, the additional lens having approximately the same size as the markers 9 and located directly in front of the markers 9.

[0040] The rear surface 4 of the spherical lens 3 is constructed such that the image depends on the orientation of the spherical lens 3 relative to the reading head 5 or the rotational position Rz, Rx, and Ry of the joint 1. Therefore, the orientation or rotational position is determined through image evaluation. Preferably, the spherical lens 3 has a refractive index n (for the wavelength of the measurement light 7) very close to 2 (±1%) (e.g., 2.002 or 1.985), as referenced. Figure 3 A more detailed explanation follows.

[0041] Figure 3 A spherical lens 3 with a measuring beam 7 is schematically shown. The spherical lens 3 is made of a material with a refractive index n≈2 for the selected wavelength of the beam 7. The beam 7 is reflected or scattered back at the constructed rear side 4 of the spherical lens 3. According to geometric principles, Snell's law, and small-angle approximations sin(α)≈α and similarly sin(2α)≈2α, it can be seen that the light rays emitted from point A on the rear side 4 of the spherical lens 3 will be refracted at the front side 4f of the spherical lens to form a parallel beam 7b.

[0042] Because the refractive index of the spherical lens material is close to n=2, the light rays diffused backward from a scattering point A on the rear side 4 are almost perfectly collimated at the output of the sphere 3. This allows for a substantial relaxation of the positional tolerance requirements for the spherical lens relative to the reading head 5.

[0043] In the paraxial approximation, and for n=2, the system is only sensitive to rotation / orientation and not to translation.

[0044] An alternative to the uniform spherical lens 3 shown is a spherical lens with a refractive index that is practically (almost) equal to 2 but varies (rotatably) in the direction of the center of the sphere, for example, a stepped or gradient radially varying refractive index, which has the same refractive effect as the spherical lens 3 with a constant refractive index of approximately 2. An example is the so-called Luneburg lens.

[0045] Figure 4 A simplified optical system with a spherical lens 3 and an image sensor 11 is shown, the image sensor being used to image measurement light 7 reflected from the rear surface 4 of the lens 3. The system includes: a (thin) lens 12 with a focal length f equal to the distance from the image sensor 11; and a spherical lens 3 with a refractive index n = 2. Some light rays 7 are emitted from point A on the rear side 4 of the spherical lens 3, and after being refracted by the front side of the spherical lens 3, the light rays 7 become parallel. The thin lens 12 then focuses the parallel beam onto point A' on the image sensor 11. Thus, point A is mapped to point A'. This can be done for any other point on the rear side 4 of the spherical lens 3, i.e., to obtain an imaging map from the rear surface 4 of the spherical lens 3 to the image sensor 11.

[0046] The encoder may include stored decoding information, which may be in the form of an image or feature descriptor of the rear surface 4 to the image sensor 11. The decoding information or image may be an intensity image of a pattern / spot, a stereo projection, or a list of coordinates of identifiable features (such as the center of a spot).

[0047] The magnification is given by the ratio of the focal length f of the thin lens 12 to the radius of the spherical lens 3, for example, 30mm:5mm = 6. This is passive except for the image sensor 11.

[0048] The optical system may also include a prism (not shown), preferably serving as the emitting optical element of the readout head. Using a beam splitter configured to split the emitted measurement light 7 into two optical paths, the rear surface 4 is imaged from different viewing angles. That is, the measurement light 7 can be split into two paths to generate multiple fields of view on the image sensor (see also...). Figure 6b For example, the viewing angle is (relative to the optical axis) + / - 30°.

[0049] Figure 5a , Figure 5b Another example of a simplified optical system is illustrated, in which, Figure 5a A system with the depicted master ray is described. Figure 5b A system with the depicted wavefront is described. The system includes a spherical lens 3 and an image sensor 11, by which the image sensor images the measuring light 7 reflected by the rear surface 4 of the lens 3 and passing through the aperture 8 and the lens 12 having a focal length f. The optical system in this example also includes a relay optics or lens 13 (preferably as part of the front optics of the readhead), which is specifically implemented as two lenses having a focal length f1, and the distance between the two lenses is twice the focal length f1. Also as shown, the marking pattern 9, as described above, is located in the image plane formed inside the relay lens 13.

[0050] The relay lens 13 is located between the aperture 8 and the spherical lens 3. The distance between the aperture 8 and the lens 13 is equal to the focal length f1 of the lens 13. Thus, a virtual aperture 8' (or entrance pupil) is formed by imaging the aperture 8 into the spherical lens 3. Therefore, the relay lens 13 maps the aperture 8 onto the sphere 3. Since the image of the real aperture 8 is "seen" from the object space, the "virtual" aperture 8' or entrance pupil is located at the center of the spherical lens 3. Therefore, this design is super-centered / near-centered at the spherical lens 3 (the lens system where the entrance pupil is in front of the lens in the space where the object is located or may be located; the optical system is far-centered at the image sensor 11).

[0051] Due to this arrangement, a virtual aperture 8' exists, centered at the center of the spherical lens 3. The purpose of the relay lens 13 is to form an image of aperture 8 at the center of the spherical lens 3, such that the principal rays at each field point (by definition, rays passing through the center of the aperture) are almost perpendicular to the surface of the sphere, as... Figure 5a As shown.

[0052] In other words, at a given point on sphere 3, a set of identical rays 7, independent of the sphere's orientation, are used. This is achieved by forming a synthetic aperture 8' at the center of sphere 3 through imaging the physical aperture 8 via relay lens 13. Or, as... Figure 5b As shown, the wavefront 18 (the set of points seeing the same optical path) is depicted from the surface of the spherical object to the image plane 11. Because the curvature of the wavefront 18 matches the surface of the sphere in the object plane, all points follow the same optical path, since the planar wavefront 18 is parallel to the image surface in the image plane. This allows for high accuracy in the encoder.

[0053] Figure 6a It shows examples of Figure 2 An example of an image 14 generated by the image sensor 11 of the encoder 2 is depicted. The image has a marked pattern 9 at its center. It is surrounded by the image of the rear surface 4i of the spherical lens.

[0054] In the example, the back surface image is a speckle image 4i. To generate such a speckle image 4i, the back surface 4 is a rough surface (preferably with a reflective coating), and the measurement light 7 is coherent (preferably with low coherence (minimum temporal coherence)), reducing interference effects (especially avoiding parasitic effects), but still in speckle form. The rough surface can be formed, for example, by sandblasting, silver coating, and / or ALD coating. The aperture 8 defines the speckle size or resolution of the imaging optics. In practice, the correlation of the measurement light 7 is adapted to the roughness of the surface 4.

[0055] As an alternative to or addition to the speckle image, image formation is achieved through specular reflection, i.e., a pattern of the reflective and absorptive portions of the rear surface 4 of the sphere. In this case, the encoder 2 includes an incoherent measurement light source. The rear surface 4 may, for example, be a reflective coating defined by photolithography, such as being covered with black paint or chromium, and / or display a ring or dot pattern, for example, with a clearly defined specific distribution.

[0056] The speckle image 4i is measured relative to the marked image 9i, which is independent of the spherical lens 3. This makes the measurement, for example, independent of the image sensor translation. For example, if the image sensor is unstable, then not only is there a translation of the speckle pattern 4i, but also a translation of the marked image 9i. Without the marked image, the translation of the speckle pattern would be incorrectly interpreted as a rotation of the spherical lens 3. Therefore, the marked image 9, for example, allows for the correction of, for example, thermal deformation (change in shape) or lateral translation and rotation about the optical axis / rotation in the lateral plane of the image sensor by means of the marked image 9 imaged onto the sensor 11 which defines the zero point and orientation.

[0057] Figure 6bAnother example of a measurement image 14 is shown. Such an image (in this example, a speckle image) is obtained by an encoder that includes a prism as part of the imaging optics, as shown above. Figure 4 As described in the context above. Therefore, there are multiple optical paths for measuring light, and thus multiple images 4i of the back surface as explained above. For embodiments including prisms, for example, two images with high angular accuracy about the optical axis and a wide baseline b can be produced.

[0058] Image 14 includes an image of a marking pattern in the center, as referenced as described above. Thus, the marking pattern can be integrated into the prism itself, or it can be a separate element in the beam path, for example, implemented as a marking plate with adjacent lenses between the prism and the image sensor.

[0059] Figure 7 Another example of measuring image 14 is shown. In this example, four regions of interest (ROIs) are depicted. Instead of reading out the entire image 14, only these ROIs are evaluated to increase the measurement rate (e.g., in the range of hundreds of Hz or up to kHz). Additionally, data compression can be achieved in the sensor via feature detection.

[0060] However, in a preferred embodiment of the encoder 2 with self-calibration capability, the entire image 14 is evaluated when self-calibration is performed. In other words, when measurement and high frame rates are required, only the ROI is evaluated; when calibration and minimizing errors are required, the entire image 14 or at least a larger frame is evaluated.

[0061] This self-calibration can include using at least one reference angle encoder (single DoF). Utilizing this reference value from at least one DoF allows for better calibration accuracy.

[0062] In any case, the marker pattern 9 and the rear side 4i are imaged in one image 14, and the image of the rear side is evaluated with respect to the imaged marker pattern 9i. In the example, each region of interest includes the marker pattern 9i. That is, image 14 includes multiple small markers 9i, rather than... Figure 6a and Figure 6b The single large marker is shown. Utilizing this alternative concept, using four markers in the corners of an allowable field of view can provide, for example, rotational accuracy of <1 arcsec, while reducing the total area of ​​marker 9. Since marker 9 doesn't move much at all, larger aberrations can be tolerated. The markers can be made from a diagonal stripe pattern. As shown, for example, four ROIs can be used, where each ROI contains a rectangular area of ​​one of markers 9i and spot 4i.

[0063] Figure 8Further developments in the image-based multi-DOF encoder are depicted. For clarity, only the readhead 5 and the spherical lens 3 are shown. In this embodiment, two different measurement wavelengths are used. For example, for illumination, light emitted by two different LEDs is used: the first LED emits light at 595 nm (orange light), and the second LED emits light at 730 nm (infrared light). Because the refractive index of the spherical lens 3 is not exactly 2 for the wavelengths used, the system is sensitive to the translation of the spherical lens 3 relative to the readhead 5 or the translation of the two articulated elements (if n=2 is perfectly satisfied, then the system is only sensitive to orientation and not to translation).

[0064] For each wavelength, the virtual image 15 or 16 on the back side of the spherical lens 3 is generated by the optical system. When n = 2, these virtual objects will have an infinite size; that is, viewing the spherical lens 3 with respect to image rotation and translation will be like viewing stars in the sky. For n < 2, virtual object 16 is located on one side of the spherical lens 3, and for n > 2, virtual object 15 is located on the other side. For the two selected wavelengths, the two virtual objects 15 and 16 for 595 nm and 730 nm respectively are located on opposite sides of the spherical lens 3.

[0065] Therefore, it can be said that two large virtual spherical surfaces 15 and 16 are imaged on opposite sides of the spherical lens 3 (virtual images exist for n < 2; real images exist for n > 2). The readhead 5 measures the smaller sub-parts on both sides. The optical axis 17 is the connection to the physical position corresponding to the center of the two images. Therefore, in addition to translation along the optical axis 17, the relative orientation (rotation and translation) of the readhead 5 and the spherical lens 3 can be determined for 5 DOFs. This final DOF can be determined based on the magnification of the two wavelengths.

[0066] Therefore, by utilizing the dispersive properties of the spherical lens 3, additional DOF can be measured to determine one or more translational positions in addition to the rotational position, by evaluating images of the same portion of the rear surface generated at different wavelengths. That is, a measurement using two wavelengths is provided, wherein at least one wavelength is selected such that n deviates slightly from 2. The translation can be determined based on subtle differences between the image transformations at the two wavelengths. In the illustrative embodiment, the readhead 5 uses at least two different LEDs to generate two images of the spherical lens 3. Using these two images, along with, for example, a pre-stored reference image and calibration parameters, the orientation (6DOF) of the spherical lens 3 relative to the readhead 5 can be determined.

[0067] Preferably, two wavelengths of light are combined and image acquisition is performed alternately using these two wavelengths. Different light sources can be turned on alternately to separate the images in time.

[0068] Preferably, the encoder measures two or three angles with a general dynamic range (e.g., ±45°) and displacement vectors (only two lateral components or all three components, including coaxial components) with a small actuation range (the joint allows relatively large rotations but only small translations in comparison). Of course, the encoder may include additional conventional position sensors for measuring one or more DOFs; for example, the encoder may be complementary to an additional 1-DOF sensor that measures rotation about an optical axis.

[0069] The ability to determine one or more translational positions is particularly advantageous for identifying small, unintended translational changes in joints, such as those caused by stress or heavy loads. Using two colors of measuring light allows for the differentiation between translation and rotation, enabling relaxation of mechanical tolerance requirements, for example, achieving mechanical clearance of approximately 50 micrometers.

[0070] When the spherical lens 3 is translated, the translation of the image is not the only visible effect. The local radius of curvature, the type of sphere (diffuse versus reflection), and the aperture size affect the translation sensitivity. This variation is the same for both wavelengths.

[0071] It is clear that the figures shown are merely schematic illustrations of possible exemplary implementations. Unless otherwise stated, the various methods described can also be combined with each other and with corresponding prior art devices and methods according to the invention.

Claims

1. A photoelectric encoder (2) for a joint (1), the photoelectric encoder (2) having: Read head (5), the read head includes: At least the first measuring light source (6), and Image sensor (11), the image sensor being used to image the reflected measurement light (7), An optical system comprising at least a spherical lens (3) and an aperture (8, 8'). in, The spherical lens (3) includes: The front surface (4f), which is transparent to the measuring light (7) and faces the reading head (5), and The constructed rear surface (4) is at least partially reflective. The spherical lens (3) has a refractive index (n) of at least approximately 2 for the measurement light (7) or a radially varying refractive index (n), such that the beam of the measurement light (7) reflected from point (A) on the rear surface (4) of the spherical lens (3) inside the spherical lens (3) is refracted at the front surface of the spherical lens (3) to form a beam that is at least approximately parallel. The reading head (5) and the spherical lens (3) are capable of rotating relative to each other in at least two degrees of freedom. The rotational position of the reading head (5) and the ball lens (3) relative to each other can be determined using an image (14) of at least a portion of the rear surface (4) by image evaluation of the structure of the imaged rear surface (4) with respect to the at least two degrees of freedom (Rx, Ry, Rz), the image (14) being captured by the optical system using the image sensor (11).

2. The photoelectric encoder (2) according to claim 1. Its features are, The photoelectric encoder (2) is a photoelectric encoder used to measure the joints (1) of an articulated arm or robot.

3. The photoelectric encoder (2) according to claim 1. Its features are, The image sensor (11) is a CMOS sensor.

4. The photoelectric encoder (2) according to claim 1. Its features are, The radially varying refractive index (n) is either a step refractive index or a gradient refractive index (n).

5. The photoelectric encoder (2) according to claim 1. Its features are, The reading head (5) includes at least one marker pattern (9) that can be imaged on the image sensor (11) using measurement light (7) to compensate for changes in the shape or orientation of the image sensor (11) relative to the optical system using the position of the at least one marker pattern (9) in the image (14).

6. The photoelectric encoder (2) according to claim 5. Its features are, The change in the shape or orientation of the image sensor (11) relative to the optical system is a thermally induced change.

7. The photoelectric encoder (2) according to any one of claims 1 to 6. Its features are, The back surface (4) is constructed with respect to reflectivity, and the measurement light (7) is incoherent.

8. The photoelectric encoder (2) according to any one of claims 1 to 6. Its features are, The structures of the first measuring light source (6) and the rear surface (4) are adjusted to each other in such a way that when the rear surface (4) is illuminated by the measuring light (7), a spot image (14) is formed on the image sensor (11), and for this purpose the coherence of the measuring light (7) is adapted to the roughness of the rear surface (4).

9. The photoelectric encoder (2) according to any one of claims 1 to 6. Its features are, The aperture (8) is located in the reading head (5).

10. The photoelectric encoder (2) according to any one of claims 1 to 6. Its features are, The aperture (8, 8') is a virtual aperture (8') formed at the center of the spherical lens (3) by imaging the real aperture (8) in one focal point of the relay lens (13) of the optical system.

11. The photoelectric encoder (2) according to claim 5 or 6. Its features are, The at least one marker pattern is located in the image plane of the optical system, such that the at least one marker pattern is imaged onto the image sensor.

12. The photoelectric encoder (2) according to any one of claims 1 to 6. Its features are, Furthermore, the at least one translational position of the spherical lens (3) and the reading head (5) relative to each other with respect to at least one degree of freedom can be determined in the following manner: The dispersive properties of the glass using the spherical lens (3) are evaluated by evaluating images (14) of the same portion of the rear surface (4) generated using different wavelengths. For this purpose, the reading head (5) includes a second light source having a second wavelength that is different from the first wavelength of the first measuring light source (6). and / or The intensity and / or decorrelation of the image are evaluated.

13. The photoelectric encoder (2) according to claim 12. Its features are, For individual image generation, the first measurement light source and the second light source are used alternately.

14. The photoelectric encoder (2) according to any one of claims 1 to 6. Its features are, The photoelectric encoder (2) includes a beam splitter configured such that the emitted measurement light (7) can be split into two paths, thereby enabling imaging of the rear surface (4) from different perspectives.

15. The photoelectric encoder (2) according to claim 14. Its features are, The beam splitter is a prism.

16. The photoelectric encoder (2) according to any one of claims 1 to 6. Its features are, The photoelectric encoder (2) has a self-calibration function.

17. A method for determining the rotational position of a joint (1) about at least two rotational degrees of freedom, said joint (1) having a photoelectric encoder (2) according to claim 1, said method comprising the steps of: Measurement light (7) is emitted at the spherical lens (3); The optical system receives the measurement light (7) reflected back from the rear surface (4) of the spherical lens (3). The reflected measurement light (7) is captured as an image (14) using the image sensor (11); The rotational position with respect to the at least two degrees of freedom (Rx, Ry, Rz) is determined by image evaluation of the image (14).

18. The method according to claim 17, Its features are, The joint (1) is a joint of a metering articulated arm or robot.

19. The method according to claim 17, Its features are, The at least two rotational degrees of freedom (Rx, Ry, Rz) are all three rotational degrees of freedom (Rx, Ry, Rz).

20. The method according to any one of claims 17 to 19, wherein the method is used in the photoelectric encoder (2) according to claim 2. Its features are, The at least one marker pattern (9) and the rear surface (4) are imaged in an image (14), and the image (4i) of the rear surface (4) is evaluated with respect to the imaged marker pattern (9i).

21. The method according to any one of claims 17 to 19, Its features are, Only one or more regions of interest (ROIs) distributed in the image (14) of the rear surface (4) are evaluated as part of the imaged rear surface (4) for location determination.

22. The method according to any one of claims 17 to 19, Its features are, The emission of the measurement light (7) is synchronized with the capture of the reflected measurement light (7) by the image sensor (11).

Images