156B Week 4
Overview of Status of Analysis, Design, Fabrication, Tests, etc.
Continued working on dynamic analysis in Matlab as well as the frequency/motion analysis in SolidWorks.
Accomplishments from Previous Week
Created visualization of kinematics in Matlab.
Completed the simplified SolidWorks model.
Goals for Next Week (list names after each item). Use specific and measurable objectives.
Optimize simulation parameters and results.
For the kinematics MATLAB model:
We have to figure out a way to send the time history of the angle positions for each joint for a given trajectory to the robot.
The time step for each angle needs to be small enough so that the motors are not overloaded given the inertial properties of the robot arm
Include this in our MATLAB program to determine the duration of this time step after inputting the inertial parameters of the robot and the properties of the motors.
For the CAD model:
We need to empirically verify the simplified Solidworks model
Determine how we want to obtain empirical data
Use phone apps, spring scales, etc.
Determine design enhancements to improve robot performance (e.g. Natural Frequency)
Reduce the mass, moving center of mass, increase the stiffness in the joints, or tuning the motor's internal PID controller.
Research whether Solidworks has the capability to simulate motion from point A to point B and analyze the dynamics of this motion.
Sponsor Comments from Last Meeting and Actions Taken to Address these Comments (indicate date of comments and if via email or in person)
4/21 Sponsor Meeting
They like the two approaches, but are hoping that we can have more hands-on experiment.
they were confused about why we don't use a 3x4 jacobian matrix that includes the wrist joint.
we explained to them that we excluded the wrist joint for simplification. But we kept the wrist joint in the back of our minds on how to keep it level throughout the arm's motion.
We also ended up including the wrist joint and using a 3x4 matrix. We used pinv(A) in matlab to calculate the pseudo inverse of the 3x4 matrix.
they thought the MATLAB program was a very good step. Now they just want us to figure out a way to send the time history of the angle positions for each joint for a given trajectory from point A to point B to the robot's motor joints and execute this motion.
We're not sure how to send the time history of the angles to the motors.
the way the motor controls work is we send target positions (in angles) to the motors and the motor's internal PID position control determines the motion.
However, what we did, was calculate the joint angles for a given target xyz coordinate using our inverse kinematic program and send those joint angles to the robot.
Sponsors were impressed with the Solidworks model. They asked to propose design enhancements to improve the natural frequency.
Instructor Comments from Last Meeting and Actions Taken to Address these Comments (indicate date of comments and if via email or in person)
Prof. Delson also likes how we combine Matlab and SolidWorks.
Buy a spring scale to obtain stiffness of each motor.
We bought a spring scale.
Comments from Other Students in the Class (indicate date of comments and if via email or in person)
N/A
Risks and Areas of Concern
Not being able to conduct too much experiments and data acquisition in real life. Edward's room is very tiny and there isn't much space to conduct tests. However, having the robot at Edward's apartment was most convenient since Edward and William is part of the control interface team.
We don't know how to get instantaneous position values (and other parameters such as torque, speed, current, voltage) from the motors as it is in motion. We can only obtain it once the motion is complete
Still unsure whether our solidworks model is accurate.
Resources or Information Required but not Available
N/A
Schedule
Describe upcoming milestone
Update Gantt chart.
Budget (list amount spent and amount remaining)
N/A
Progress on Report and Webpage
N/A