What is a difference between degrees of freedom (DOF) and degrees of motion (DOM)? I know that DOF is the number of independent movements a manipulation arm can make and robot system can have max 6 independent DOF and unlimited number of DOM but I do not distinguish them from each other.
$\begingroup$
$\endgroup$
1
-
1$\begingroup$ DOF is never limited to 6 in robotics. Actually, we tend to use redundant manipulator, for which it holds DOF > 6 (when we span the whole Cartesian space). Therefore, you should better clarify what is the context in which you raised your question... $\endgroup$– Ugo PattaciniApr 29, 2016 at 17:13
Add a comment
|
2 Answers
$\begingroup$
$\endgroup$
16
The way you are describing it, DOM is the number of independent dimensions in $\vec q$. DOF is the number of independent dimensions in $\vec x$. In practice, a robotics engineer will use DOF to represent the number of independent actuators of the robot, which you are calling DOM. Better notation would be to call DOM the mobility of the system, as described in the Mobility formula section here.
Using your notation:
Think of a telescoping pole made out of 5 sections that can collapse inside each other. The degree of freedom is the length of the pole. That value can be anything between its fully-collapsed length, up to its fully-extended length. DOF = 1.
If there is a sliding joint between each section of the pole, there will be four independent motions you can use to set the total length. DOM = 4.
In kinematics, the length would typically be denoted by $x$, and the individual joint lengths would be assigned values of $q_1$ through $q_4$.
$\vec x = f(\vec q)$
-
$\begingroup$ The DOF are actually the number of independent movements that can be achieved by the structure and usually map 1 to 1 into the number of independent joints. In your example DOF=4 (given that the first section is fixed). $\endgroup$ Apr 29, 2016 at 16:28
-
$\begingroup$ Also, you make an improper use of the word rank. See en.wikipedia.org/wiki/Rank_(linear_algebra). $\endgroup$ Apr 29, 2016 at 16:31
-
$\begingroup$ @Ugo, your comment about DOF is usually true for robotics, where we refer to the number of actuators as DOF. But when discussing DOM vs DOF as in the OP's question, he is referring to output DOF, and DOM refers to the number of independently controlled actuations that accomplish the given number of output DOFs. So for his question, my example has DOF = 1 (length of pole) and DOM = 4 (number of independent motions). If any of the joints are mechanically coupled to another, such as $q_1 = q_2$, the DOM would reduce to 3. Hence the use of the word rank - independent terms of $\vec q$. $\endgroup$– SteveOApr 29, 2016 at 16:58
-
1$\begingroup$ Yes, relative to your original question. When you get into multi-axis robot control you will start to call the number of actuators DOF. But for where your learning is now, on the mechanism side, that is correct. $\endgroup$– SteveOApr 29, 2016 at 18:31
-
1$\begingroup$ Sorry, but I still believe that both the question and the answer remain quite misleading for future readers of this post. $\endgroup$ Apr 30, 2016 at 11:13
$\begingroup$
$\endgroup$
It appears the industry need to consider some standardization of the terms. In my search I have found differing definitions for the term DOF.
From Yaskawa/Motoman: "Degrees Of Freedom : The number of independent directions or joints of the robot (R15.07), which would allow the robot to move its end effector through the required sequence of motions. For arbitrary positioning, 6 degrees of freedom are needed: 3 for position (left-right, forward-backward, and up- down) and 3 for orientation (yaw, pitch and roll)." Note: R15.07 would appear to reference the ANSI/RIA/ISO materials but at RIA, I only found reference to R15.06.
From Robot Worx: Degrees of Freedom - The amount of values in a system possible of variation. A robotic joint is equal to one degree of freedom.
From ATI automation another RIA member: "Six Degrees of Freedom - A fancy way of saying Fx, Fy, Fz, Tx, Ty and Tz."
I would have to disagree with the Robot Worx, because at a minimum both directions of rotation or travel must be considered for each plane. So Yaskawa's explanation of 3 planar linear directions and three planar rotations makes more sense as well as the definition from ATI. If this is accepted, then what of DOM? This term must inevitably be more complex and account for many things other than direction. I see no terminology so far that accounts for actual degrees of motion of servos/actuators for example servos designed for 180°/360° motions. There should be some way to relate this functionality in the glossary of terms whether it be DMO or yet another term. Some method of using all the available functions of a robotic device should be somehow described.
In my opinion there should also be a standard definition for a description that includes, such other functions as: time, acceleration, speed, grip/release, etc.
