Sliding mode control gain optimization for a robot arm manipulator using an improved stochastic framework

Area of Science:

• Robotics
• Control Systems
• Optimization Algorithms

Background:

• Robotic manipulators face challenges like nonlinearities, couplings, and external disturbances.
• Achieving high trajectory tracking accuracy with reduced control effort and chattering is crucial for complex robotic applications.

Purpose of the Study:

• To develop an optimization-control framework for trajectory tracking of a 3-DoF manipulator.
• To automatically tune sliding mode controller (SMC) gains using an enhanced Stochastic Paint Optimizer (SPO-CL1).
• To enhance robustness, accuracy, convergence speed, and reduce chattering and actuation effort in robotic control.

Main Methods:

• An improved Stochastic Paint Optimizer (SPO-CL1) was developed with chaotic initialization, Opposition-Based Learning, and Lévy flight perturbations.
• SPO-CL1 automatically tuned the gains of a sliding mode controller by minimizing the Integrated Squared Error (ISE) cost function.
• Validation involved benchmarking against eleven algorithms on the CEC-2022 suite and a path planning experiment on a Lemniscate of Bernoulli trajectory.

Main Results:

• SPO-CL1 achieved the best Friedman rank (1.83) on the CEC-2022 benchmark suite, demonstrating statistically significant superiority over eleven other algorithms.
• In the path planning experiment, SPO-CL1 yielded the lowest ISE (1.51 × 10⁻⁴) with minimal inter-run variance.
• The approach demonstrated the fastest tracking error convergence and the tightest end-effector trajectory compared to all twelve algorithms.

Conclusions:

• SPO-CL1 is a competitive and reliable approach for automatic SMC gain tuning in complex robotic applications.
• The proposed framework effectively addresses nonlinearities, couplings, and disturbances, leading to improved robotic control performance.
• The integration of structure-aware enhancements significantly boosts the optimizer's performance in tuning control system parameters.