Scribd
Upload
12 views

Lab 4 - Trajectory Interpolation Reference Guide

The document discusses two methods for smoothly interpolating between joint positions of a robot arm: 1) using a quintic polynomial and 2) using a trapezoidal velocity profile. The quintic polynomial method defines a 5th order polynomial equation to calculate intermediate joint positions based on a scalar value between 0 and 1. This method peaks briefly but is inefficient. The trapezoidal velocity profile method explicitly defines the joint velocity to follow a trapezoidal shape, allowing for more efficient use of motor capabilities. Both methods aim to define smooth transitions between joint positions to reduce motor workload.

Uploaded by

Ng Chris
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
12 views

Lab 4 - Trajectory Interpolation Reference Guide

The document discusses two methods for smoothly interpolating between joint positions of a robot arm: 1) using a quintic polynomial and 2) using a trapezoidal velocity profile. The quintic polynomial method defines a 5th order polynomial equation to calculate intermediate joint positions based on a scalar value between 0 and 1. This method peaks briefly but is inefficient. The trapezoidal velocity profile method explicitly defines the joint velocity to follow a trapezoidal shape, allowing for more efficient use of motor capabilities. Both methods aim to define smooth transitions between joint positions to reduce motor workload.

Uploaded by

Ng Chris
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

Creating a trajectory: Interpolation from A to B

By Jonathan Woolfrey

There are two ways to approach moving a robot arm: joint space, or end-effector/task space. In specifying
end-effector motion, we still have to resolve the necessary joint velocities to achieve the task. This can get
complicated. For now, we’ll look at how to smoothly transition a joint from one position to another.

We don’t want our joint motors to be working too hard, so if the joint positions transition smoothly, then the
joint velocities will also be smooth.

We can define an intermediate joint position between two joint angles and as a function of a
smooth scalar .

= 1− + = 0,1 ❶

It can be seen then that:

when s = 0
=
when s = 1

There are two main approaches for achieving this:

1. Quintic Polynomial: an easy, but inefficient method.


2. Trapezoidal Velocity Profile: a more complex, but much more efficient method.
1. Quintic Polynomial

Let be a 5th order polynomial function of time:

= + + + + +! ❷
"=5 +4 +3 +2 +
' = 20 + 12 +6 +2

When = 0

) =!
") =
') = 2

And at some final time *

+ = * + * + * + * + *+!
" =5 * +4 * +3 * +2 *+
+

+' = 20 * + 12 * + 6 * + 2

) 0 0 0 0 0 1
. ") 1 . 0 0 0 0 1 01 . 1
- ') 0 - 0- 0
- 0 - 0 0 0 2 0 00 - 0
We pack these in a matrix for easier handling: + =
- 0 - * * * * * 10 - 0
- +" 0 - 5* 4* 3* 2* 1 00 - 0
, + ' / ,20* 12* 6* 2 0 0/ , ! /

0 0 0 0 0 1 23 )
. 0 . ") 1
. 1 0 0 0 1 01
- 0 - 0 - ') 0
Then solve for the polynomial coefficients: - 0= - 0 0 0 2 0 00 - 0 ❸
- 0 - * * * * * 10 - +
0
- 0 - 5* 4* 3* 2* 1 00 - +" 0
,! / ,20* 12* 6* 2 0 0/ , + ' /

Then plug ❸ back in to ❷, and ❷ back in to ❶.

Fortunately, Robot Toolbox


can solve this for us using,
jtraj(q1,q2,stepCount).

The joint velocity profiles will


look something like this:
2. Trapezoidal Velocity Profile

The velocity profile using Quintic Polynomials peaks for a brief instance in time which leads to two problems:

1. Not making full use of the motor’s speed capabilities


2. The motor has to use a lot of power accelerating and decelerating. This is not energy efficient.

Instead, if we explicitly define the velocity " to


follow a trapezoidal shape, as shown:

The area under the trapezoid defines


the total distance we have
travelled in that time period.

Using some geometry and algebra


we can define:

0 for = )
7
5 ;< − ) + ½> − )
23
+ for ) < < 3
@;< 3 − ) + ½> 3 − ) + ;A − 3 B 23
+ for 3 ≤ <
6
5 ;< − + ½> − + ;A − + ;A − − ½> − for ≤ <
23
3 ) 3 ) 3 + +
4 1 for t ≥ +

We also have to be careful when selecting the


parameters of the velocity profile, or we could end up in
a situation like this:

Robot Toolbox can also help us generate trapezoidal


profiles, albeit with a little extra work.

s = lspb(0,1,steps);
qMatrix = nan(steps,6);
for i = 1:steps
qMatrix(i,:) = (1-s(i))*q1 + s(i)*q2;
end

You might also like