How far can i reach? A case study in robot knowledge acquisition

This paper mainly describes robot knowledge acquisition and reasoning underlying actual applications, starting with introducing some basic problems with constructing an intelligent system. On one hand, growing neural gas (GNG) and Natural Actor-Critic (NAC) are used for acquiring source knowledge; Afterward on the other hand, active logic and metacognitive loop (MCL) are incorporated as reasoning engines/mechanisms and robot knowledge base (KB) is established for knowledge representation and reasoning. Several cases including “How far can I reach” are raised to support our arguments. Across these experiments and potential applications, we show that with suitable symbolic representation of knowledge and a well-designed systematic structure, our autonomous robot agent can basically perceive, learn from and interact with the environment, and reason about and handle unexpected problems using its knowledge base. Some analysis and future research directions are presented in the end.