Introduction
As robots become increasingly sophisticated, they are being applied to a growing number of industrial processes. Six-DOF (6 Degrees of Freedom) robots are capable of performing complex movements, thus demonstrating excellent performance in many complex industrial tasks, such as palletizing, handling, gluing, and welding. Six-DOF robots can lift and manipulate heavy loads with dexterity, and precisely control loads through complex geometric movements.
Due to their excellent performance and flexibility, 6-DOF robots are used in conjunction with various tools for many different tasks. However, each time a new tool is introduced, precise recalibration is required before the robot can perform its functions effectively. Recalibration is often time-consuming and inaccurate, hindering the smooth operation of industrial processes and causing production delays.
Servotronix has developed an efficient position teaching method for rapid calibration of new tools used by 6-DOF robots, without relying on manufacturer measurements or external sensors. This method is simple, accurate, and highly effective for practical applications.
calibration
Six-DOF robots need to hold and move tools when performing tasks. To achieve satisfactory performance, the robot must know the precise position of the tool during operation. Each time a different tool is assembled, the robot must be precisely recalibrated.
There are different methods for calibrating a 6-DOF robot. These include contact with a reference component, using a distance sensor, and measuring with a laser interferometer. Alternatively, external sensors, such as a camera system, can be mounted at different locations on the robot to obtain the precise position of the reference object used for calibration.
These methods are both time-consuming and complex. Servotronix has developed a simpler method that has achieved excellent results. We will explain this method below.
Determine the center point of the tool
We use the kinematic criterion method to determine the tool center point (TCP), which is the reference point for defining all robot localization. The TCP is defined in the world coordinate system – a Cartesian coordinate system that can locate any point in the world. This coordinate system will always remain stationary relative to the robot.
Tool coordinate system
The tool coordinate system defines the tool's position and orientation, with its zero point set at the tool's center point (TCP). The robot's TCP moves to its programmed position as it performs Cartesian motion. Changing the tool will alter its coordinate system, thus requiring recalibration to ensure the new TCP accurately reaches the target position.
In many robotic applications, the TCP (Tool Trajectory) means a complex path within the robot's workspace, typically a straight path for a tool that changes orientation. This tool itself needs to be changed occasionally, or even frequently. Each time the tool is changed, a new set of geometric parameters must be determined and configured before the robot resumes operation.
In most industrial applications, position teaching is the most practical method for robot task programming. When using this method, high-precision tool parameters (usually from the manufacturer) are required, including the tool's angular offsets (yaw, pitch, and roll) and Cartesian offsets, in order to generate a straight path with controllable tool posture.
Unfortunately, operators often find that tool geometry identification is limited by factors such as: (1) lack of information from the manufacturer regarding tool dimensions; (2) lack of available hardware assistance; and (3) inability to determine how to mount the tool onto the robot flange. Faced with these limitations, operators must waste considerable time calibrating the tool each time it is changed.
Accurate evaluation and simplified calibration
Servotronix has developed a method for quickly and accurately estimating tool geometry parameters without the need for external sensors, vision, or other assistance, or tool disassembly. In this position teaching method, the operator simply positions the 6-DOF robot's TCP in several different poses and then automatically inputs the data into Servotronix's tool size evaluation algorithm. This algorithm rapidly determines the precise calibration parameters for the new tool, enabling it to be quickly put into use.
The accuracy of this calibration method improves with an increase in the number of tool pose samples. Our experiments show that using the inverse homogeneous transformation matrix may not produce the desired results, but using the least squares method will yield accurate calibration values.
Servotronix's method
We used a 6-DOF robot equipped with a tool, six Servotronix high-performance CDHD servo drives, and a Servotronix softMC controller for testing. Our approach involves only analytical calculations without disassembling the tool. We only evaluate the XYZ dimensions and assume the tool's endpoints are in constant Cartesian coordinates.
It is self-evident that all robot poses pointing to the same location must lie on a sphere, and the tool tip must be located at the center of the sphere:
TCP can be calculated by measuring points on the sphere.
Where t represents the center
R2 = (X - Xt)2 + (Y - Yt)2 + (Z - Zt)2
Equation (1) contains four unknown parameters (R, Xt, Yt, Zt). The values of X, Y, and Z are calculated using forward kinematics. To achieve acceptable accuracy, our method requires at least four points to define a sphere. Therefore, the four such settings would be:
R2 = (X1 - Xt)2 + (Y1 - Yt)2 + (Z1 - Zt)2
R2 = (X2 - Xt)2 + (Y2 - Yt)2 + (Z2 - Zt)2
R2 = (X3 - Xt)2 + (Y3 - Yt)2 + (Z3 - Zt)2
R2 = (X4 - Xt)2 + (Y4 - Yt)2 + (Z4 - Zt)2
By using equation subtraction, we can not only eliminate the unknown variable R, but also remove all nonlinear components from the equations. This will generate a set of first-order polynomial equations, which can be solved by least-squares fitting. Using more than four points will produce more equations and higher accuracy.
This step-by-step method takes only a few minutes to complete and requires at least four measurements, as shown below:
Summarize
Servotronix's method is fast, accurate, and economical, allowing for calibration without disassembling tools. This method requires no specialized hardware, saving the time and effort needed to install new tools. Machine manufacturers can easily use this method to quickly, accurately, and virtually cost-free recalibrate the tools of 6-DOF robots, thereby enhancing the robot's operational smoothness in a wide range of applications and accelerating production speed.
