About this video
A full run of LEG2: changing the quadruped's foot target coordinates and computing joint angles with inverse kinematics (IK) in RoboCo on Steam.
When you set a foot target as x·y·z coordinates, the author's Python library computes the IK and sends the resulting joint angles to RoboCo's ServoMotors.
The selected leg's foot moves forward/back, left/right, and up/down, then follows a circular path, showing how foot position relates to joint angles.
Keys in this exercise
| Key | Action |
|---|---|
| W / S | Foot x up / down |
| A / D | Foot y up / down |
| E / Q | Foot z up / down |
| U / O / J / L | Select leg FL / FR / HL / HR |
| I | Automatic foot-circle test |
| K | Selected leg to the stand pose |
| F1 | Key help |
| F2 | All four legs to the stand pose |
| F3 | Stop the circle |
The FK error on screen is the difference between the target and the position computed back from the joint angles. It isn't a separate measurement of where the foot actually is.
The calculation chain on screen
- Foot target
- IK
- HX · HY · KN angles
- Servo targets
- Compare with FK
What to watch
- All three joints moving together when the target changes
- The IK angles and the angles sent to the servos being different numbers
- The path the foot traces in the circle test
Exercise details
| Book | Learning Physical AI with RoboCo: Robot Structure and Gait Control (Korean edition) |
|---|---|
| Series | Physical AI with RoboCo · QUAD-01 (WALK) |
| Chapter | PART 3 · CONTROL · CHAPTER 11 |
| Related chapters | CH10·11·12·13 |
| Practice model | Book-QUAD-01-LEG2 |
| Python start file | q01_leg2_ik_start.py |
| Practice files version | v3.5.0 — download |
| Video ID | Q01-11-00 |
| YouTube | youtu.be/GdoPi8bZAd0 (opens in a new tab) |