知识驱动下数据融合的Petri网建模

Knowledge-Driven Data Fusion Modeling Using Petri Nets

针对数据融合中存在的数据表达能力不强、融合效率低的问题,提出采用知识驱动的Petri网进行数据融合建模.首先,对多传感器采集的数据进行信息处理,以此为基础,将知识通过产生式规则进行知识表示.然后,通过Petri网理论设计其映射模型和映射算法,建立Petri网的数据融合模型.最后,将该模型在搭建的自主追踪机器人实验平台上进行验证.结果实现了三种异类传感器的数据融合,使自主追踪机器人能较好的完成追踪过程,验证了该方法的正确性和适用性,可为多源信息的数据融合提供方法与模型支持.

Aiming at the issue that low data expression ability and low fusion efficiency in data fusion, a data fusion modeling methed based on Petri net by knowledge-driven is proposed. Firstly, the data of multi-sensors was processed. Based on this, the knowledge through the generation of rules to representation knowledge. Then the mapping model and the mapping algorithm are designed by using Petri net theory to establish the fusion model. Finally, the model is validated on the experimental platform of autonomous tracking robot. The experimental results show that three kinds of heterogeneous sensor data fusion can help autonomous tracking robot to better complete the tracking process. The correctness and applicability of this method are verified, which can be widely used in data fusion modeling of multi-source information.