A Temperature Compensation Method for aSix-Axis Force/Torque Sensor Utilizing Ensemble hWOA-LSSVM Based on Improved...

Area of Science:

• Robotics and Control Systems
• Sensor Technology
• Machine Learning Applications

Background:

• Six-axis force/torque (F/T) sensors exhibit performance degradation in extreme environments due to temperature sensitivity.
• Accurate F/T sensing is critical for applications requiring precise force and torque measurements.

Purpose of the Study:

• To develop an advanced ensemble temperature compensation method for F/T sensors.
• To improve the precision and stability of F/T sensors operating under varying ambient temperatures.

Main Methods:

• An ensemble method combining whale optimization algorithm (WOA) tuned least-square support vector machine (LSSVM) with trimmed bagging was proposed.
• The WOA was hybridized with simulated annealing (SA) to mitigate local entrapment issues.
• An adaptive trimming strategy and inverse quote error (invQE) with cross-validation were used for enhanced training.

Main Results:

• The maximum absolute temperature-induced measurement error was reduced from 3.34% to 3.9×10-3% of full scale.
• The proposed hWOA-LSSVM ensemble method demonstrated significant improvements in F/T sensor precision and stability.
• Experimental analyses confirmed the method's effectiveness and adaptability.

Conclusions:

• The developed ensemble temperature compensation method effectively addresses the performance limitations of F/T sensors in extreme environments.
• The integration of WOA, LSSVM, and trimmed bagging offers a robust solution for enhancing sensor accuracy and reliability.
• The method showcases strong local search capabilities and adaptability, making it suitable for demanding applications.