Robot Task Planning Based on State Semantic Network

Abstract

The generation of robot task planning is a challenging issue in uncertain and dynamic real world. Semantic knowledge can be used to support the task planning of robots as a source of implied knowledge. At present, many literatures elaborate semantic networks, which integrate spatial information and semantic information, and use semantic knowledge to carry out robot task planning. Aiming at the demand of automatic robot task planning, this paper improved the semantic network with the trend of the development of the current semantic network. This paper proposed a state semantic network (SSN) based on state machine and semantic network. The state semantic network is a semantic network composed of semantic objects with current state information, and the current state information of the objects in the SSN is determined by the state machine associated with the semantic network objects. In the case of reasoning based on the state semantic network objects, only when the object being reasoned and its current state satisfy specific conditions, can the reasoned object be iterated as a semantic object associated with the current object that conforms to the particular condition, and achieves further reasoning.