Vehicle Decision System Based on Domain Knowledge Graph

The decision-making system constitutes a crucial module of the unmanned ground vehicle. Considering the issue that unmanned vehicles deal with a vast amount of information and the traditional knowledge representation fails to meet the requirements, a decision-making approach based on the domain knowledge graph is herein proposed. The knowledge graph offers distinct advantages. The integration of this technology can enhance the understanding of the environment, thereby improving the level of autonomy and intelligence. Nevertheless, existing models typically represent knowledge in a static manner, overlooking dynamic information. Consequently, an unmanned vehicle knowledge graph model layer is established, vehicle dynamic information modeling is designed, a temporal model layer is proposed, and an intelligent knowledge-driven decision-making system is constructed. This method integrates the dynamic and static aspects of environmental information. Specifically, the construction of a speed series temporal layer is proposed to provide more information for the decision-making module and enhance the credibility of decision results. The proposed method was tested on a real vehicle platform. The experimental results demonstrate that this system can effectively recognize the environment, make rational decisions by integrating dynamic and static information, and enhance safety and reliability.