DC10 - Integrating Behavioral Specification and Automated Validation for Explainable Autonomy in Collaborative Robotics

Objectives

1. To develop a novel methodology grounded in Behavior-Driven Development (BDD) and tailored specifically for accommodating complexity and variability inherent in human-robot collaboration. By developing domain-specific concepts for acceptance criteria, the methodology will support the precise specification and automated validation of tasks where robots and humans work together.
2. To create a digital twin-based validation system, e.g., within NVIDIA Omniverse, equipped with measurement models that monitor robot behaviors in real-time to offer precise, automated validation of acceptance criteria for collaborative tasks.

Expected Results

1. A methodology and framework based on domain-specific models encoded as knowledge graphs as an efficient tooling to specify and validate collaborative robot tasks.
2. A digital twin-based validation system, e.g., within NVIDIA Omniverse, equipped with measurement models that monitor robot behaviors in real-time to offer precise, automated validation of acceptance criteria for collaborative tasks.

Planned Secondment(s)

1. S19 LNE, Dr. Mahtani, 2 months in M14-15
2. S20 PAL, Dr. Lemaignan, 2 months in M38-39

Required Skills

Essential

• Degree in Computer Science, Robotics, Artificial Intelligence, or a related field
• Experience in robotic simulation
• Excellent programming skills (e.g., Python)

Desirable

• Experience with formal or scenario-based specification techniques
• Familiarity with simulation environments such as Gazebo, Isaac Sim, orlema Work Designer
• Basic knowledge of collaborative robotics, human-robot interaction, or explainable AI methods