Joint Reconfiguration after Failure for Performing Emblematic Gestures in Humanoid Receptionist Robot.

Area of Science:

• Robotics
• Artificial Intelligence
• Control Systems

Background:

• Humanoid robots with anthropomorphic arms face challenges with actuator failures during tasks.
• Failures in robotic actuators can impede the execution of pre-programmed motions and gestures.

Purpose of the Study:

• To develop a joint reconfiguration algorithm for the Namo receptionist robot to ensure gesture performance post-actuator failure.
• To enable quick fault recovery response in humanoid robots with 7-DOF arms.

Main Methods:

• Utilized bio-inspired artificial intelligence: Genetic Algorithm (GA), Bacteria Foraging Optimization Algorithm (BFOA), and Artificial Bee Colony (ABC).
• Proposed a gesture similarity measurement as an objective function for joint reconfiguration.
• Employed grid search for parameter tuning and compared the performance of AI methods.

Main Results:

• Bio-inspired AI methods successfully suggested reconfigured gestures within 1 second after simulated joint motor failure.
• Both BFOA and ABC provided optimal reconfigured gestures, with ABC showing higher reliability and less result variability.
• The joint reconfiguration method demonstrated effectiveness across all possible joint failure scenarios.

Conclusions:

• The proposed method enables effective joint failure recovery in humanoid robots under strict time constraints.
• Bio-inspired AI algorithms, particularly ABC, offer a robust solution for real-time robotic fault recovery.
• This strategy enhances the operational resilience of humanoid robots in dynamic environments.