Skip to main content

Lesson

PD control

How robots move smoothly and how Cyberwave makes controller tuning safe.

PD (Proportional-Derivative) control is the fundamental technique that makes robot joints move smoothly to target positions. Getting the settings right is critical—too aggressive and the robot vibrates dangerously; too gentle and it's slow and weak.

What Is PD Control?

A PD controller tells a motor how hard to push based on two things:

  1. Proportional (P): How far am I from the target? Push harder when far away, lighter when close.
  2. Derivative (D): How fast am I moving? Slow down as you approach to avoid overshooting.

Think of it like driving a car: you accelerate more when you're far from your destination (P), and you brake as you approach to stop smoothly (D).

The balance between P and D—called "gains"—determines how the robot behaves:

  • P too high: The robot oscillates back and forth around the target
  • P too low: The robot is weak and may not overcome friction or gravity
  • D too high: Movement becomes sluggish and unresponsive
  • D too low: The robot overshoots and rings like a bell

Why Tuning Is Hard

Finding the right gains requires experimentation, but experimenting on real hardware is risky:

  • Oscillating robots can overheat motors in seconds
  • Violent vibrations damage gearboxes and bearings
  • Each robot may need different settings due to manufacturing variation
  • Changes that work in the lab may fail under production loads

Traditionally, tuning means carefully testing on hardware while watching for signs of trouble—slow, risky, and hard to scale.

How Cyberwave Helps

Cyberwave lets you tune controllers safely in simulation, then deploy to hardware with confidence.

Simulation-First Tuning

Adjust gains in Cyberwave's simulation environment. Test with different payloads, speeds, and scenarios. Find settings that work before touching real hardware.

Fleet-Wide Deployment

Push new controller settings to hundreds of robots at once. Track which configuration is running where. Roll back instantly if something goes wrong.

Automatic Monitoring

Cyberwave watches for signs of trouble in real-time:

  • Oscillation detection: Alert when vibration indicates unstable gains
  • Error tracking: Flag when the robot consistently can't reach targets
  • Safety stops: Halt automatically before motors overheat

Controller Policies

Define different configurations for different situations:

  • Precise mode: High stiffness for accurate positioning
  • Compliant mode: Soft settings for safe human interaction
  • Transport mode: Optimized for carrying payloads

Switch between policies per-task without re-tuning.

From Dangerous to Safe

Controller tuning used to mean risking expensive hardware. With Cyberwave, you iterate safely in simulation, deploy confidently to production, and monitor automatically for problems.

For implementation details, see the Cyberwave documentation.

More tracks

Keep exploring Cyberwave Learn

Every track is defined in markdown so adding new lessons stays fast and versioned.