Robot and sensor error determination system

• US 4,967,370 A
• Filed: 10/21/1988
• Issued: 10/30/1990
• Est. Priority Date: 10/21/1988
• Status: Expired due to Term

• Current Assignee: Rudolph Technologies Incorporated (Onto Innovation Inc.)
• Original Assignee: Robotic Vision Systems, Inc. (Acuity Brands, Inc.)
• Inventors: Stern, Howard K., Schmuter, Samson L.
• Primary Examiner(s): MacDonald, Allen
• Application Number:
  US07/260,862
• Time in Patent Office: 739 Days
• Field of Search: 364/513, 364/474.34, 364/474.35, 364/474.37, 364/192, 364/571.05, 364/571.02, 364/571.01, 901/46, 901/47, 318/568.22, 318/632
• US Class Current: 700/254
• CPC Class Codes: B25J 9/1692 Calibration of manipulator

Abstract

A method for measuring the position and orientation changes of a 3-D sensor relative to a robot mount and for measuring the positioning and orienting changes of the robot mount. All measurements reveal changes relative to a baseline measurement. A novel reference provides a low-cost measurement aid. A reference object is placed within the work volume of the robot carrying the sensor which is capable of making three-dimensional measurements. The reference object has a shape that enables unambiguous determination of its location and orientation. The reference object is measured from predetermined orientations, and the measurements are recorded in a memory. These measurements form a baseline measurement set. The measurement step is repeated at a later time, and another measurement set is produced. The baseline measurement set and the other measurement set are processed to determine robot positioning errors and sensor location errors.

Claims

1. A method for determining robot positioning errors and sensor location errors, comprising the steps of:

placing a reference object within a work volume of a robot carrying a sensor capable of making three dimensional measurements, said reference object having a symmetrical shape about an axis, said shape unambiguously defining location and orientation of said object for its small deviations relative to the robot position;

measuring said reference object from a single predetermined robot position at fixed sensor orientations rotationally incremented about a robot axis that is substantially aligned to said axis of symmetry;

recording said measurements in a memory, said measurements forming a baseline measurement set;

repeating said measurement step at a later time and producing another measurement set; and

processing said baseline set and said another measurement set to determine separately robot positioning errors and sensor location errors.

2. A method as defined in claim 1, wherein said reference object is a square block mounted on a second square block, said square block having its edges respectively parallel to the diagonals of said second square block.

3. A method as defined in claim 2, wherein the measurements are made in four 90 degree increments of sensor rotation.

4. The method as defined in claim 1 wherein said reference object is a square cavity located in a square block, said cavity having its edges respectively parallel to the diagonals of said square block.

5. A method as defined in claim 1, wherein said reference object is a four-sided pyramid mounted on a square block, said pyramid having its base edges respectively parallel to the diagonals of said square block.

6. A method as defined in claim 1, wherein said reference object is a three-sided pyramid mounted on a symmetric base.

7. A method as defined in claim 6, wherein the measurements are made in three 120 degree increments of sensor rotation.

8. A method as defined in claim 1, including the step of displaying separated robot and sensor errors that are determined for an operator to make corrections.

9. A method as defined in claim 1, wherein said robot is positioned with respect to said reference object such that motions of robot joints individually transform to sensor motions predominantly along one principal axis of the said reference object with minimum cross coupling between robot axes.

10. A method as defined in claim 1, wherein the robot is positioned with respect to said reference object such that one robot joint causing sensor motions in similar directions with another robot joint is located at the distance to said reference object that is substantially greater than the distance from said another robot joint to said reference object.