一个用于机器人自然语言理解的英语句法分析系统

在低挡微机中速度较慢的串行处理硬设备条件下，利用本文提出的启发式概念，分层搜索和匹配策略以及设置最大搜索长度等方法，可使推理速度提高一个数量级以上．此外，通过引入语义信息，分阶段消除歧义，自顶向下与自底向上相结合，以及把一般疑问句一律变成相应陈述句的方法，解决了自动英语句法分析中的一系列难题，缩小了知识库的规模。

Under the condition of serial processing hardware in low  end  microcomputers the speed of which is lower, the inference process can be sped up more than 10 times by means of the concept of “heuristic processing”, hierarchical search strategy and matching strategy, and setting the maximun search length presented in this paper. Furthermore, a serial problems of automatic english parsing have been solved and the size of  knowledgebase has been reduced after intraducing semantic information, staged resolving syntax ambiguities, integrating Top  down and Bottom  up analysis, and transforming all Yes/No questions into corresponding declarative sentences.

中国科学院机器人学开放研究实验室基金