Adaptive Tracking Control of Anchoring Unit for Pipeline Intelligent Plugging Robot Based on Improved Deep Deterministic Policy Gradient

A pipeline intelligent plugging robot (PIPR) is an important tool in subsea pipeline maintenance and emergency repair. Precise position-tracking control is crucial for the in-pipe plugging operation of a PIPR. The anchoring module is the key component responsible for fixed-point braking, which faces the challenges of insufficient structural adaptability within a narrow space. Additionally, traditional PID control may lead to poor robustness under fluctuating working conditions and load disturbances. To address these issues, this study designs a novel anchoring module combining screw transmission, an eccentric crank–slider mechanism, and a parallelogram linkage. To achieve adaptive tracking control, the improved deep deterministic policy gradient (DDPG) algorithm is introduced to optimize the parameters of the PID controller. A reward function with mechanical constraint penalties and a dual-phase strategy is proposed for dynamic parameter optimization. All control performances are analyzed and verified through simulations. The results indicate that the proposed method outperforms traditional PID control as regards response speed, overshoot, and robustness, which can achieve precise anchoring. This study provides a theoretical foundation for ensuring the precision of the plugging process.