A neural blueprint for human-like intelligence in soft robots

Source Domain: news.mit.edu

Adaptive AI Control System: The new AI system enables soft robotic arms to learn and adapt to diverse motions and tasks without retraining.
Real-world Applications: This technology enhances soft robotics for practical uses such as assistive, rehabilitation, and medical devices.
Structural and Plastic Synapses: The AI combines structural synapses (offline learning) with plastic synapses (real-time adaptation) to improve adaptability and stability.
Interdisciplinary Collaboration: Researchers from MIT, the National University of Singapore, and Nanyang Technological University co-developed this system.
Versatility and Stability: The system maintains performance across different tasks and disturbances, ensuring reliability in unpredictable environments.
Generalizability: The method works across various soft-arm platforms and supports multiple task types, like trajectory tracking and object placement.
High Performance: Demonstrated significant improvements, achieving reductions in tracking error and high shape accuracy under disturbances.
Future Applications: Potential to advance manufacturing, logistics, medical, and industrial soft robots, improving efficiency in various fields.

A neural blueprint for human-like intelligence in soft robots | MIT News