HOME
Autonomous Ultrasound Probe in Unity
Project Timeline
Sep 2025 - Current
OVERVIEW
This project focuses on developing an autonomous ultrasound scanning system using Unity, ML-Agents, and Soft Actor-Critic (SAC) reinforcement learning. I built a high-fidelity simulation environment with deformable tissue models (CRESSim-MPM) and synthetic ultrasound rendering, enabling agents to learn realistic probe–tissue interactions. A major step forward was transitioning from location-based to image-based rewards, allowing the agent to learn directly from ultrasound imagery rather than positional cues. Through curriculum learning, domain randomization, and integrated soft-body physics, I created a robust and clinically relevant simulation pipeline for AI-guided medical imaging.
HighlightS
- ✨ Developed a first-of-its-kind autonomous ultrasound simulation combining SAC reinforcement learning with deformable tissue physics.
- 🧠 Transitioned from location-based to image-based rewards, enabling agents to learn tissue targeting directly from ultrasound images.
- 🎯 Integrated synthetic ultrasound rendering and tissue deformation models, creating a realistic environment for training medical AI.
- 🚀 Built a unified simulation pipeline connecting soft-body physics, ultrasound imaging, and RL policies for reproducible experiments.
- 🔄 Applied curriculum learning and domain randomization, dramatically improving agent robustness and generalization.
- ⚙️ Optimized SAC agent performance through reward restructuring, privileged state supervision, and environment tuning.
SKILLS
UnityCRESSimUltrasoundReinforcementSACSimulationPhysicsRenderingAutomationModeling
Additional Details
lowinertia
Portfolio Builder for Engineers