AUTOBIO-KI
Scalable Evaluation and Robotics System for Automated High-Performance Biomass Production through AI Integration
AUTOBIO-KI develops a fully AI-controlled biological production system — combining Physics-Informed Neural Networks (PINNs), real-time Computer Vision, and autonomous Edge-AI robotics into a closed-loop "Sim-to-Real" architecture for industrial biomass cultivation.
As consortium coordinator, MindWaves AI is responsible for overall project management and leads the development of the Edge-AI vision pipeline (YOLOv8 / DINOv2) and the autonomous control logic deployed directly on edge hardware (Raspberry Pi 5 + Hailo-8L NPU).
MindWaves Contribution
- Consortium coordination & project management
- Edge-AI Computer Vision pipeline (biomass classification, anomaly detection)
- Autonomous control logic for robotic sensor systems
- Real-time inference at sub-100 ms latency on edge devices
Consortium Partners
- Energaia Labs GmbH — Industrial bioconversion & system validation
- Gemino GmbH — Cloud infrastructure & data architecture
- Epic Technologies UG — Mechatronic gantry systems & sensor integration
- Robert Boyle Institut e.V. — PINNs, Digital Twin & scientific leadership
The "Sim-to-Real Loop" — a closed feedback cycle where Physics-Informed Neural Networks continuously calibrate against real sensor data, enabling predictive resource optimisation that purely data-driven models cannot achieve.