New MPPI-PID control method optimizes PID gains for smoother path following

Researchers have developed a novel control method called Model Predictive Path Integral PID (MPPI-PID) control, which optimizes PID gains online rather than directly sampling control input sequences. This approach aims to improve sampling efficiency and produce smoother control inputs by optimizing low-dimensional gain-space instead of high-dimensional input sequences. The method was evaluated on a mini forklift performing path following using a hybrid physical and neural network model, demonstrating improved tracking performance and smaller input increments compared to traditional PID and conventional MPPI. AI

IMPACT This new control method could lead to more efficient and smoother robotic path following in industrial applications.

RANK_REASON The cluster contains a research paper detailing a new control method. [lever_c_demoted from research: ic=1 ai=0.7]

Read on arXiv cs.LG →

• artificial neural network
• arXiv
• mini forklift
• model predictive control
• model predictive path integral
• MPPI-PID control
• Teruki Kato

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →