Autonomous planning of service robot based on natural language tasks in intelligent space

Abstract

It is promising for service robots to be able to accomplish natural language tasks given by human automatically. In this paper, we propose a flexible framework, which is aimed at resolving Chinese service tasks into executable sequences of primitive actions. Firstly, we collect a number of service tasks described in free-form Chinese natural language. We then build the intelligent space corpus by using the method combined semantic parsing with the conditional random field (CRF) to extract key information, which is used to convert Chinese tasks into robot control instructions. Based on this, a semantic knowledge ontology model, referred to as action-environment ontology model, is constructed by STRIPS standard structure, allowing a feasible solution (an action sequence) for a particular task to be created. A recursive forward-back search algorithm is proposed to generate executable actions sequence for task planning. At last, the nature language processing experiment is performed with the complex index (F) value of 95.7%, and the primitive actions sequence is executed in robot experimental platform under intelligent space environment.