Fiducial marker for a robot

Abstract

A system is disclosed which comprises a robot 2, at least one fiducial marker 6 supported on the robot, wherein each fiducial marker 6 comprises a backing (11; Figs. 2A, 2B, 3A, 3B) having a front sur

Description

Field The present invention relates to a fiducial marker for a robot particularly, but not exclusively, for a modular robot. Background Pose estimation is a common problem in computer vision. Although there has been significant progress in recent year, position estimate can still be difficult to achieve. When estimating the pose of humans, motion capture suits are used with visually salient points that can be illuminated. When capturing the pose of a robot, image processing can be aided by passive patterns (or "fiducial marks" or simply "fiducials") or specifically-shaped active markers, such as LEDs or other lights), which are easier to pick up compared to a darker background. Electrically-powered light sources, however, need power provided by wires. This can present a problem when the wire passes through or over a joint of a robot or when the surface of a robot needs to be kept clean for ingress protection or hygiene reasons. Furthermore, the location of the fiducial does not necessarily provide orientation information to the observer. This complicates the reliability of that data at the sensor fusion stage. Summary According to a first aspect of the present invention there is provided a system comprising a robot and at least one fiducial marker supported on the robot, wherein each fiducial marker comprises a backing having a front surface, a phosphorescent layer disposed on the front surface, a convex transparent outer shell covering the phosphorescent layer and configured to form a generally flat enclosure between the shell and the backing, and a disc or bubble freely disposed in the enclosure. The system further comprises at least one camera system configured to view at least part of robot arm and a computer system configured to receive image(s) of the at least one fiducial marker from the at least one camera system and to process the image(s) so as to determine position of the robot. This arrangement can have one or more advantages. A phosphorescent marker does not require electricity to run. A light in a charging station or exposure to sunlight can charge a marker for many hours. The marker provides a gravity sensor and so gives orientation information. The marker is cheap and easily mounted to the robot. The amount of computer processing needed to determine position of the robot can be reduced. The marker (once exposed to illumination) can work in darkness or in environments in which passive markers might not work. A phosphorescent marker tends to be high contrast, even in high ambient light levels, allowing the system to work in all light levels. The enclosure may be filled with a liquid or gas. The robot may comprise at least two modules assembled to form the robot. The at least one camera system may be separate from (or "external to") the robot. The at least one camera system may be integrated into the robot. The camera system may comprise an omnidirectional camera. The camera system may comprise a camera having an optical axis, and a parabolic mirror disposed on the optical axis and arranged so as to capture an image in a direction generally perpendicular to the optical axis. According to a second aspect of the present invention there is provided a method comprising attaching at least one fiducial marker to a robot wherein each fiducial marker comprises a backing having a front surface, a phosphorescent layer disposed on the front surface, a convex transparent outer shell covering the phosphorescent layer and configured to form a generally flat enclosure between the shell and the backing, and an opaque disc freely disposed in the enclosure. The method may further comprise configuring at least one camera system to view at least part of robot arm and configuring a computer system to receive image(s) of the at least one fiducial marker from the at least one camera system and to process the image(s) so as to determine position of the robot. According to a third aspect of the present invention there is provided use of phosphorescent marker as a fiducial marker, wherein the fiducial marker comprises a backing having a front surface, a phosphorescent layer disposed on the front surface, a convex transparent outer shell covering the phosphorescent layer and configured to form a generally flat enclosure between the shell and the backing, and an opaque disc freely disposed in the enclosure. According to a fourth aspect of the present invention there is provided a fiducial marker comprising a backing having a front surface, a phosphorescent layer disposed on the front surface, a convex transparent outer shell covering the phosphorescent layer and configured to form a generally flat enclosure between the shell and the backing, and an opaque disc freely disposed in the enclosure. The disc may support a symbol for recognised by a vision processing system. The disc may have a weight or be more dense to help to retain orientation of the symbol. Brief Description of the Drawings Certain embodiments of the present invention will now be described, by way of example, with reference to the accompanying drawings, in which: Figure 1 is a perspective view of a modular robot in the form of an articulated arm and a plurality of fiducial markers disposed on joints of the articulated arm; Figures 2A and 2B are top plan views of a first and second types of fiducial markers; Figure 3A and 3B are side view of the first and second types of fiducial markers shown in Figures 2A and 3A respectively; Figure 4 is a plan top plan view of a fiducial marker that has been illuminated in one position and has been moved to have a different rotation and / or orientation; Figure 5 illustrates a system comprising the modular robot and fiducial markers shown in Figure 1, an external camera system and a computer system receiving an image from the external camera system and displaying the wireframe model of the modular robot based on the fiducial markers; Figure 6 is a perspective view of a modular robot in the form of an articulated arm having cameras integrated into the joints of the articulated arm and a plurality of fiducial markers disposed on joints of the articulated arm; Figure 7 is a detailed perspective view of a joint of the robot shown in Figure 6 which shows a fiducial marker disposed on the joint of the robot and a camera system integrated into the joint arranged to view another fiducial marker on an adjacent joint; and Figure 8 illustrates a system comprising the modular robot, fiducial markers and integrated camera systems shown in Figure 6, and a computer system receiving images from the camera systems and displaying the wireframe model of the modular robot based on the fiducial markers. Detailed Description of Certain Embodiments Referring to Figures 1 to 5, a system 1 for determining the position of a robot 2 is shown. In this example, the robot 2 takes the form of a robot arm. The robot 2 is modular and comprises a series of modules 4 which can include a base 4b, linkages 4l and a gripper 4g. The robot 2 is articulated and has a series of joints 5. The system 1 includes a plurality of fiducial markers 6 (which are herein also referred to as "markers" or "fiducials") supported on the robot 2 and which are phosphorescent. In particular, the markers 6 are mounted on the joints 5. Preferably, each joint 5 supports a pair of marker 6 on opposite sides. Referring in particular to Figures 2A and 3A, a first type of fiducial marker 6, 6i is shown. Each fiducial marker 6, 6i comprises a backing 11 which is generally disc-shaped having a front surface 12 and a back surface 13, a thin phosphorescent layer 14 (or "coating") disposed on the front surface 12, a convex transparent outer shell 15 (or "cover") covering the phosphorescent layer 14 which is configured to form a generally flat enclosure 16 (or "space") between the shell and the backing 11, and an opaque disc 17 (or "circular token") freely disposed in the enclosure 16. The disc 17 may be formed of dense plastic or metal. The enclosure 11 is filled with a fluid (not shown) which may be a liquid, such as water or oil, or a gas, such as air. The back surface 13 of the backing 11 is coated with an adhesive (not shown). The marker 6 serves as a passive gravity token. The disc 17 may be significantly more dense than the fluid that fills the enclosure 16, for example, by being a metal disc in oil or water. Alternatively, the disc 17 may be significantly less dense than the fluid that fills the enclosure 16, for example, by being an air bubble in water or oil. This can help to give a reliable directional reading (for example metal in air or an air bubble in water). Referring in particular to Figures 2B and 3B, a second type of fiducial marker 6, 62 is shown. The second fiducial marker 6, 62 is the same as the first fiducial marker 6, 62 (Figure 2A) except that the disc 17' is painted, cut or otherwise marked with symbol 18 that can be easily recognised by a vision processing system. Additionally, the disc 17' may have a weight 19 or be more dense to help to retain orientation of the symbol 18. Referring to Figure 4, the fiducial marker 6 may be used to capture an initial reference position. If the marker 6 is exposed to light with the disc in a first position and then its orientation is changed so that the disc 17 is in a second position, then not only can the new position be detected, but also the initial position can determined due to a dark spot 20 in the phosphorescent layer 14 is left where the disc 17 covered the spot during illumination. Referring in particular to Figure 5, the system 1 further comprises at least one camera system 21 configured to view at least part of robot 2 and a computer system 31 configured to receive image(s) of the at least one fiducial marker 6 from the at least one camera system and to process the image(s) so as to determine position of the robot. The markers 6 are attached on both sides of each joint 5. In this case, the camera system 21 takes a form of a camera which is separate (or "remote") to the robot 2 so that it has a global view, in other words, its field of view contains the whole of the robot 2 and the visible markers 6. The camera 21 can detect the phosphorescent markers 6 even at low light levels. The image processing application software running on the computer system 31 (which may be embedded in the robot 2, in the camera 21 or a separate computer system) recognises the markers 6 and processes them. For example, image processing software can generate a wireframe model 40. Following this step, pose and / or position estimates can be used in a control application software and / or path planning application software, or displayed on the monitor 32. Referring to Figure 6 to 8, a second system 1, I2 for determining the position of a robot 2 is shown. The second system 1, I2 is the same as the first system 1, li (Figure 5) except that instead of having a camera system 21 which has a global view of the robot 2 and visible markers 6, cameras 22 are incorporated into each module 4, in this case in the joint 5. A camera system 21 takes the form of a multi-directional camera comprising a camera 22 and a parabolic mirror 24. This allows each camera 22 to detect fiducials that are within the visual field and to reconstruct the position of the robot. The distortion of the image caused by the camera can be accommodated for by the vision processing application software. Referring in particular to Figure 8, the screen 32 displays the outputs 41, including visible fiducials, of four cameras 22 and a reconstructed pose estimate 42 of the robot 2. Advantages The arrangement can have one or more advantages. Phosphorescent markers do not require electricity to run, a light in the charging station or exposure to sunlight can charge them for many hours. It can serve as a gravity sensor, giving orientation information, which is particularly important to mobile robotic platforms. The markers are cheap, easy to source and easy to mount to the robot. The amount of processing needed is decreased since a simpler algorithm or smaller AI model will be needed to recognise the markers. Phosphorescent markers can work in complete darkness where passive markers would not work. Phosphorescent markers are high contrast, even in high ambient light levels, allowing the system to work in all light levels. Modifications It will be appreciated that various modifications may be made to the embodiments hereinbefore described. Such modifications may involve equivalent and other features which are already known in the design, manufacture and use of robotic systems and component parts thereof and which may be used instead of or in addition to features already described herein. Features of one embodiment may be replaced or supplemented by features of another embodiment. Although claims have been formulated in this application to particular combinations of features, it should be understood that the scope of the disclosure of the present invention also includes any novel features or any novel combination of features disclosed herein either explicitly or implicitly or any generalization thereof, whether or not it relates to the same invention as presently claimed in any claim and whether or not it mitigates any or all of the same technical problems as does the present invention. The applicants hereby give notice that new claims may be formulated to such features and / or combinations of such features during the prosecution of the present application or of any further application derived therefrom.

Claims

1. A system (1) comprising:■ a robot (2);■ at least one fiducial marker supported on the robot, wherein each fiducial marker (6) comprises:- a backing (11) having a front surface (12),- a phosphorescent layer (14) disposed on the front surface,- a convex transparent outer shell (14) covering the phosphorescent layer and configured to form a generally flat enclosure (16) between the shell and the backing, and- a disc or bubble (17) freely disposed in the enclosure;■ at least one camera system (21) configured to view at least part of robot arm; and■ a computer system (31) configured to receive image(s) of the at least one fiducial marker from the at least one camera system and to process the image(s) so as to determine position of the robot.

2. The system of claim 1, wherein the robot comprises at least two modules (4) assembled to form the robot.

3. The system of claim 1 or 2, wherein the at least one camera system (21) is separate from the robot.

4. The system of claim 1 or 2, wherein the at least one camera system (21') is integrated into the robot.

5. The system of claim 4, wherein the camera system comprises:a camera (22) having an optical axis (23); anda parabolic mirror (24) disposed on the optical axis and arranged so as to capture an image in a direction (25) generally perpendicular to the optical axis.

6. A method comprising:attaching at least one fiducial marker to a robot wherein each fiducial marker (6) comprises:- a backing (11) having a front surface (12),- a phosphorescent layer (14) disposed on the front surface,- a convex transparent outer shell (14) covering the phosphorescent layer and configured to form a generally flat enclosure (16) between the shell and the backing, anda disc or bubble (17) freely disposed in the enclosure.

7. The method of claim 6, further comprising:■ configuring at least one camera system (21) to view at least part of robot arm; and■ configuring a computer system (31) to receive image(s) of the at least one fiducial marker from the at least one camera system and to process the image(s) so as to determine position of the robot.

8. The use of phosphorescent marker as a fiducial, wherein the fiducial marker (6) comprises:- a backing (11) having a front surface (12),- a phosphorescent layer (14) disposed on the front surface,- a convex transparent outer shell (14) covering the phosphorescent layer and configured to form a generally flat enclosure (16) between the shell and the backing, anda disc or bubble (17) freely disposed in the enclosure.

Images