Materials and CAD Files

Structural Components

Electronics and Motors

Hardware

Custom Part Files

Tube Materials

Assembly Guides

Custom Part Manufacturing Guide

Full System Assembly

Cart Assembly

Tube Fabrication

Electronics Setup

Electronics Board Setup

Firmware Installation

Wire Management

Full System Test

Kinematics

Calculating Kinematics of a CTR

Implementing Kinematics

Troubleshooting Kinematics

Experiments

Installing Tubes

Homing the Robot

Graph Paper In-Plane Bending Experiment

Out-of-Plane Rotation Experiment

For this experiment, you will need a piece of graph paper, tape, a ruler, the graph paper mounting bracket, and your set of concentric tubes. We will demonstrate using two tubes, but you can apply the same procedure to three tubes.

This experiment will only test in-plane bending. We will not be rotating the tubes, but this will still give you an idea of how the tubes change shape as they interact with each other.

1. Home the robot as described in Homing the Robot.
2. Cut out a rectangular piece of graph paper that will fit on the graph paper mounting bracket.

1. Tape the graph paper onto the mounting bracket, and attach the mounting bracket to the front of the robot using M4 button head screws and M4 nuts. The graph paper should sit just barely behind the tubes, as shown in the image below.

   

2. Translate both tubes about 25mm out from the home position. This will give you more room to measure them

3. Mark the location where the tube curvatures begin, and mark the end tip of the tubes as shown below.

1. Translate only the inner tube by 35mm forward. Mark the new end tip of the robot in a new color.

1. Remove the board from the robot. You should see the following markings:

1. Measure the distance shown between the points as shown below. Note the coordinate frame (shown below in blue) where Z points to the right, and X points up. This is in line with our kinematics frame.

1. Using the perform_kinematics.m script to calculate the expected transformation matrix for q_var = [0 0 0 0]. This should match below. Here we have highlighted the values for $x$ and $z$ position in meters.

1. Convert the $x$ and $z$ translation values to millimeters and compare to your actual results. Here, we measured that the robot was at 19mm in $x$ and 47 mm in $z$. We predicted that the robot should be at 21.7mm in $x$ and 43.1mm in $z$. We have an error of ~3mm in $x$ and ~4mm in $z$.
2. Repeat the error calculation for q_var = [0 35 0 0]. Here, our measured values (shown in red above) are 76 mm in $z$ and 36 mm in $x$. Our expected values are 44.3mm in $x$ and 63.6 mm in $z$. We have an error of ~8 mm in $x$ and ~13 mm in $z$.

1. Once you have calculated your error, consider the following questions:
   1. Do your error values make sense?
   2. Were you able to measure and predict values that were within 10 mm?
   3. If you weren’t, why do you think that might be?
   4. What sources of error can you identify in this experiment? For example, the using a wide tip pen could result in marking the exact end of the tubes. What else can you think of?

Congratulations! You have completed your first kinematic validation experiment!

While this is exciting, you most likely noticed a few drawbacks to this approach. First, this method takes time. We need to manually mark positions, measure locations, and calculate kinematics separately. Secondly, we introduce human error with all of the manual positioning and measuring. Thirdly, we can only measure positions in the x-z plane using this approach. We cannot test our out-of-plane rotation functions.

In the Experiments section of this page, we provide instructions for how to use our actuation system to conduct more controlled kinematic experiments.