서비스 로봇을 위한 시-공간 상황 질의 프레임워크

Abstract

In this paper, we propose a context query language, called the spatio-temporal robotic context query language (ST-RCQL), and an efficient query processing system, ST-RCQP, for service robots operating in indoor environments. In order to accomplish their tasks successfully, indoor service robots need to not only recognize the current context that changes dynamically but also remember past contexts. To meet these requirements, the proposed context query language ST-RCQL is designed to efficiently retrieve 3D spatial relations between indoor objects that continuously change with the passage of time. Based on Allen’s interval algebra, ST-RCQL includes convenient temporal operators to find and compare different spatial contexts at different times. ST-RCQL has high expressive power to represent spatio-temporal context queries and has a precise grammar structure. The proposed ST-RCQP is a query processing system that finds answers for ST-RCQL context queries efficiently. In order to infer high-level spatial relationships between objects from real-time sensory data, ST-RCQP contains a backward spatial inference engine. Moreover, it has the facility to improve the query processing speed by maintaining both the temporal index and the spatial index for a large context knowledge base. Through various qualitative and quantitative experiments, we demonstrated the high efficiency and performance of both the proposed query language SP-RCQL and the query processing system ST-RCQP.