改进的分拣作业机械臂基于范例推理-信念期望意图推理机制

Improved CBR-BDI reasoning mechanism for sorting operation mechanical arm

针对以案例推理机制为推理核心的分拣作业机械臂系统不能用于物体信息较多的复杂场景的问题,提出一种改进的基于范例推理-信念期望意图(CBR-BDI)推理机制。首先,将输入的信息作为信念(Belief),通过分词与检索得到案例属性,将其作为期望(Desire);然后,加入地图匹配、期望分析和引导三个部分以完善期望;最后,完整的期望生成解决方案作为意图(Intention)。在多物体多信息的场景中,用户可通过对话指挥系统进行分拣作业。实验结果表明,与传统的案例推理(CBR)机制相比,改进的CBR-BDI推理机制具有分析和引导能力,能用于多物体复杂场景。

Focused on the issue that the sorting operation mechanical arm which used Case-Based Reasoning（ CBR）mechanism could not be used for the complex scenario with a lot of object information,an improved Case-Based ReasoningBelief,Desire,Intention（ CBR-BDI） reasoning mechanism was proposed. Firstly,the input information was regarded as belief,and the case properties obtained through sentence segmentation and retrieval was regarded as desire. Secondly,map matching,desire analysis and guidance were added to perfect desire. Finally,complete desire generated solution which was regarded as intention. In the scenario of multiple object and information,users could command system for sorting operation through dialogue. The experimental results show that compared with the traditional CBR mechanism,the improved CBR-BDI reasoning mechanism possess the ability of analysis and guide,and can be used for the scenario of multiple object and information.