Neural Networks Predicting Microbial Fuel Cells Output for Soft Robotics Applications.

Area of Science:

• Robotics
• Biotechnology
• Artificial Intelligence

Background:

• Biodegradable soft robotics require eco-friendly energy solutions.
• Microbial Fuel Cells (MFCs) offer a soft, environmentally benign power source.
• The unpredictable nature of MFCs necessitates advanced control strategies.

Purpose of the Study:

• To explore the use of MFCs as a power source for soft robots.
• To develop an AI-based control method for MFCs.
• To enhance the functionality and predictability of MFCs for robotic applications.

Main Methods:

• Development of soft, biodegradable Microbial Fuel Cells (MFCs).
• Application of artificial intelligence, specifically nonlinear autoregressive neural networks with exogenous inputs (NARX).
• Time-series prediction of MFC electrical output based on historical data and feeding volumes.

Main Results:

• Successfully predicted the electrical output of MFCs using a NARX neural network.
• Demonstrated the feasibility of using AI for real-time MFC output forecasting.
• Established a method for determining optimal feeding intervals and quantities for MFCs.

Conclusions:

• MFCs are a viable eco-friendly energy source for soft robotics.
• AI-driven prediction enhances MFC control and reliability.
• This approach enables the integration of predictable power management into soft robot behavior.