Robot Control, Where to Start?

A robot controller is the 'brain' of the robot. It sends commands to motors, reads sensor data, and ensures the robot moves with precise position and velocity.

• Sends position, velocity, and torque commands to motor drivers
• Reads sensor feedback in real-time to correct errors
• Handles safety functions like emergency stop and safety interlocks

Traditionally, PLCs (Programmable Logic Controllers) were the standard for industrial robots. But in the AI era, SoC (System on Chip) controllers with built-in GPUs are essential.

• PLC: Optimized for simple repetitive tasks, no AI capability
• SoC: AI compute + real-time control on a single chip
• WIM Robot Controller uses NVIDIA Jetson SoC to solve both simultaneously

For robots to move precisely, control commands must be delivered at exact time intervals. This is called 'Real-time Control.'

• Commands sent to motors every 1ms (1kHz)
• Lower jitter means higher precision
• Not possible with standard OS — requires a dedicated real-time OS

ROS 2 (Robot Operating System 2) is an open-source framework for robot software development. It's the de facto standard used by robot developers worldwide.

• Modular architecture — assemble only the features you need
• Proven library ecosystem including MoveIt2 and Nav2
• DDS-based communication for distributed systems

Applying AI to robots enables autonomous machines that can 'see, decide, and act.' This combines vision, motion planning, and VLA technologies.

• AI Vision: Recognize objects and estimate pose via cameras
• Motion Planning: Calculate optimal collision-free paths in real-time
• VLA (Vision-Language-Action): Instruct robots with natural language

EtherCAT is an industrial real-time communication protocol. It's the industry standard for connecting motor drivers, sensors, and I/O devices at high speed.

• Ultra-low latency communication in μs range
• Daisy-chain topology simplifies wiring
• Supports up to 65,535 nodes