sim-to-real-transfer-skill

# Sim-to-Real Transfer Skill ## Overview Expert skill for bridging the simulation-to-reality gap through domain randomization, system identification, and transfer validation techniques. ## Capabilities - Implement domain randomization (physics, appearance, dynamics) - Configure system identification for simulation parameters - Set up adaptive domain randomization - Implement domain adaptation techniques - Configure noise injection for robust policies - Set up reality gap metrics and monitoring - Implement progressive network transfer - Configure latency simulation - Set up sensor noise modeling - Implement hardware-in-the-loop validation ## Target Processes - sim-to-real-validation.js - digital-twin-development.js - rl-robot-control.js - field-testing-validation.js ## Dependencies - Simulation environments (Gazebo, Isaac Sim) - Physical robot access - System identification tools ## Usage Context This skill is invoked when processes require transferring simulation-trained models or behaviors to real robot hardware with minimal performance degradation. ## Output Artifacts - Domain randomization configurations - System identification results - Reality gap analysis reports - Transfer validation metrics - Sensor noise models - Calibrated simulation parameters