DC12 - Measure and Compensation of the Sim2Real Gap using Hybrid Simulation

Objectives

1. To develop a tool-supported testing methodology that bridges the gap in performance and accuracy between simulation and reality, enabling faster integration of systems into the real-world.
2. To develop a validation process of robots safety functions based on simulation testing of their actual perception systems so that simulation closely resembles and matches the real system.

Expected Results

1. Tool-supported methodology to reduce the Sim2Real gap of simulation and immersion infrastructures dedicated to AI systems validation.
2. Methods underpinned by parameter identification and data processing that reduce the Sim2Real gap for robots hybrid testing.

Planned Secondment(s)

1. S23 UBremen, Prof. Hochgeschwender, 2 months in M14-15
2. S24 PAL, Dr. Lemaignan, 2 months in M38-39

Required Skills

Essential

• Robotics, Computer Science or Applied Mathematics degree
• Experience in physical or robotic simulation
• Rigor and methodical approach

Desirable

• Knowledge in ROS and robot simulators (Gazebo, IsaacSim, 4DV, etc.)
• Basics of statistical analysis
• Experience in physical handling of robots, testing or benchmark
• Basic understanding of electronics and embedded systems