Deep Reinforcement Learning-Based Motion Planning and PDE Control for Flexible Manipulators
Amir Hossein Barjini, Seyed Adel Alizadeh Kolagar, Sadeq Yaqubi, Jouni Mattila
AI summary
Problem
Flexible robotic manipulators suffer from significant endpoint vibrations during motion, which traditional control methods struggle to suppress while maintaining precise trajectory tracking using only a single base torque.
Approach
A high-level Soft Actor-Critic DRL agent generates optimized trajectories that inherently minimize vibrations, which are then tracked by a low-level nonlinear PDE controller proven to guarantee closed-loop stability via Lyapunov analysis.
Key results
- Developed a nonlinear PDE controller ensuring exponential stability with a single base torque input
- Trained a SAC-based DRL planner to generate trajectories that actively suppress endpoint vibrations
- Validated the framework through simulations and real-world experiments on a hydraulically actuated manipulator
- Achieved superior vibration suppression and tracking accuracy compared to conventional PID and cubic polynomial methods
Why it matters
This synergistic approach provides a practical, high-precision control strategy for lightweight flexible robots, advancing their reliability in industrial automation and delicate manipulation tasks.
Abstract
This article presents a motion planning and control framework for flexible robotic manipulators, integrating deep re- inforcement learning (DRL) with a nonlinear partial differential equation (PDE) controller. Unlike conventional approaches that focus solely on control, we demonstrate that the desired trajectory significantly influences endpoint vibrations. To address this, a DRL motion planner, trained using the soft actor-critic (SAC) algorithm, generates optimized trajectories that inherently minimize vibra- tions. The PDE nonlinear controller then computes the required torques to track the planned trajectory while ensuring closed-loop stability using Lyapunov analysis. The proposed methodology is validated through both simulations and real-world experiments, demonstrating superior vibration suppression and tracking ac- curacy compared to traditional methods. The results underscore the potential of combining learning-based motion planning with model-based control for enhancing the precision and stability of flexible robotic manipulators. IndexTerms—Flexiblerobotics,integratedplanningandcontrol, motion control, reinforcement learning.