Behavior Selection Method of Humanoid Robots to Perform Complex Tasks

Abstract

This paper proposes a behavior selection method of humanoid robots to perform complex tasks using the degree of consideration-based mechanism of thought (DoC-MoT). The four input (context) symbols and seven target (atom behavior) symbols are defined to perform five complex tasks. The degree of consideration (DoC) for each input symbol is represented by the λ-fuzzy measure and the knowledge link strengths between input and target symbols are represented by the partial evaluation values. Each target symbol is globally evaluated by the fuzzy integral of the partial evaluation values with respect to the fuzzy measure values. Then, one target symbol with the highest evaluation value is selected and activated. To make corrections to the robot’s wrong behaviors, a learning process from a human’s behaviors is employed to update the DoCs and the knowledge link strengths. To show the effectiveness of the proposed method, simulations are performed in a text-based simulator developed in Visual Studio 2012. The results show that the proposed method can generate human-like behaviors.