Intelligent Fault-Tolerant Control of Delta Robots: A Hybrid Optimization Approach for Enhanced Trajectory Tracking

The kinematic complexity and multi-actuator dependence of Delta-type manipulators render them vulnerable to performance degradation from faults. This study presents a novel approach to Active Fault-Tolerant Control (AFTC) for Delta-type parallel robots, integrating an advanced fault diagnosis system with a robust control strategy. In the first stage, a fault diagnosis system is developed, leveraging a hybrid feature extraction algorithm that combines Wavelet Scattering Networks (WSNs), Principal Component Analysis (PCA), Linear Discriminant Analysis (LDA), and Meta-Learning (ML). This system effectively identifies and classifies faults affecting single actuators, sensors, and multiple components under real-time conditions. The proposed AFTC approach employs a hybrid optimization framework that integrates Genetic Algorithms and Gradient Descent to reconfigure a Type-2 fuzzy controller. Results show that the methodology achieves perfect fault diagnosis accuracy across four classifiers and enhances robot performance by reducing critical degradation to moderate levels under multiple faults. These findings validate the robustness and efficiency of the proposed fault-tolerant control strategy, highlighting its potential for enhancing trajectory tracking accuracy in complex robotic systems under adverse conditions.