Human-Cyber-Physical Systems Lab （HCPS)

Our lab pioneers an integrative Human-Cyber-Physical Systems (HCPS) framework to address safety-critical decision-making challenges in dynamic, partially observable environments like healthcare and autonomous systems. Unlike traditional Cyber-Physical Systems (CPS) that prioritize full autonomy, we focus on human-machine collaboration under strict ethical and regulatory requirements, where human accountability remains central. By co-evolving an environmental digital twin (personalizing context-aware models with causal reasoning) and a cognitive digital twin (adapting to human operators’ latent states), our framework optimizes decisions through bidirectional alignment—respecting physical causality while adapting to human cognitive dynamics. This approach advances safety, transparency, and reliability in high-stakes domains where human judgment and machine intelligence must synergize under time pressure and uncertainty.

Our research spans two primary directions:

1. Personalized Medical Diagnosis and Therapy

Leveraging the dual-twin paradigm to model patient-specific physiological and environmental variables, enabling adaptive treatment plans under uncertainty.
Integrating multimodal data (e.g., biomarkers, clinician expertise) for real-time therapy adjustments while preserving interpretability and ethical alignment with medical workflows.

2. Safety-Centric Collaborative Driving Decision Intelligence

Designing human-cyber-physical interfaces for autonomous vehicles that balance driver intent, traffic dynamics, and regulatory constraints.
Enhancing collision avoidance and cooperative driving through causal risk prediction and dynamic control authority allocation between humans and machines.

By bridging causal AI, cognitive modeling, and domain-specific constraints, our work redefines safety-critical HCPS architectures where human and machine intelligence coexist as interdependent partners rather than isolated agents.