Exploring the application of artificial intelligence speech to text technology in air traffic control emergency response——Targeting automatic filling and contingency plan matching scenarios
探索人工智能语音转文字技术在空管应急处置中的应用——面向自动填报与预案匹配场景

摘要

空管应急处置具有突发性强、信息交互密集、处置节奏紧凑的显著特点，同时，信息通报是空管突发事件处置中的核心工作，需要管制员持续完成信息报送任务。传统处置模式下，管制员在开展指挥处置的同时，还要查阅应急处置检查单，持续梳理填报应急信息，多重任务并行会加重席位管制员工作负荷，间接增大运行风险。本文依托已实践的管制语音转文字试验平台，以平台输出文本作为数据基础，重点探索两类业务应用场景：一是基于转写文本关键词，智能匹配相关应急处置场景的应急处置卡；二是基于自动提取“时间戳、地点、事件类型”等关键字段，完成《突发事件信息报告表》等应急管理系统表单的智能预填报工作。

Abstract

The emergency response of air traffic control has significant characteristics of strong suddenness, intensive information exchange, and tight response rhythm. At the same time, information notification is the core work of air traffic control emergency response, which requires controllers to continuously complete information reporting tasks. Under the traditional disposal mode, while conducting command and disposal, the controller also needs to consult the emergency disposal checklist, continuously sort out and fill in emergency information, and multiple tasks in parallel will increase the workload of the seat controller, indirectly increasing operational risks. This article relies on a practical controlled speech to text experimental platform, using the platform's output text as the data foundation, and focuses on exploring two types of business application scenarios: one is an emergency response card based on transcribed text keywords, which intelligently matches relevant emergency response scenarios; The second is to automatically extract key fields such as “timestamp, location, and event type” to complete the intelligent pre filling work of emergency management system forms such as the “Emergency Event Information Report Form”.