RobotControl-MATLAB-Project

Kwangwoon University, Seoul, South Korea
Term project for Robot Control(Second semester, 2021.)

PDF, lecture material is own Kwangwoon Univ.

---

1. 3-DOF links free fall simulation

Code is here and dynamics modeling using Lagrangian function is here.

---

2. 2-DOF Dynamics parameter estimation

Code is here and two link dynamics equation is here.

To estimation, two algorithms are available.

2.1 Kalman Filter based parameter estimation algorithm

You can see that every parameters are converge quickly without large overshoot. But also little errors are exist.

2.2 Error minimization algorithm

Every parameters are converge very quickly than Kalman algorithm and error is almost zero. But in the beginning, very big overshoot exists.

---

3.1 1-DOF Joint Space PID CTM Controller

• Target position: 0 deg -> 90 deg
• Target velocity: 30 deg/s
• Use PID controller & Try several gain value

For position controller at critically damped system, P gain K_p is equal to square of system frequency ω_n² and D gain K_d is equal to 2ω_n(∵ ζ=1 at critically damped system). If want to decrease steady-state error, we can add I controller.

3.1.1 ω_n=5, K_p=ω_n², K_d=2ω_n, K_i=0

3.1.2 ω_n=20, K_p=ω_n², K_d=2ω_n, K_i=0

3.1.3 ω_n=20, K_p=ω_n², K_d=2ω_n, K_i=250

3.2 2-DOF Cartesian Space PID CTM Controller

• Target trajectory: radius-0.1m, period-1s, drawing cricle
• Use PID controller & Try several gain value

Control gain is equal to 3.1.

3.2.1 ω_n=5, K_p=ω_n², K_d=2ω_n, K_i=0

3.2.2 ω_n=20, K_p=ω_n², K_d=2ω_n, K_i=0

3.2.3 ω_n=20, K_p=ω_n², K_d=2ω_n, K_i=250

4.1 3-DOF Joint Space PID CTM Controller

Final goal is extension of 3.1 & 3.2 to 3-DOF

ω_n=20, K_p=ω_n², K_d=2ω_n, K_i=250

4.2 3-DOF Cartesian Space PID CTM Controller

ω_n=20, K_p=ω_n², K_d=2ω_n, K_i=250