Air Traffic Control and Pilot Voice Communications Analysis with 12,200 Elements

12,200 communication elements from 4,500 transmissions between pilots and controllers at 3 terminal facilities (TRACONs) were analyzed. Each element was categorized by speech act (e.g., instruction, request) and aviation topic (e.g., heading, altitude) and evaluated for communication errors. The analysis was conducted by O. Veronika Prinzo using the aviation topic-speech act taxonomy (ATSAT).

Use Cases

• Modeling communication error patterns based on speech act categories and aviation topics.
• Training or evaluating speech recognition systems on structured ATC/pilot dialogue.
• Analyzing verbosity and conversational strategies in advisory or request speech acts.
• Studying error mitigation strategies like number grouping or omission in time-critical communications.

Strengths

• Analysis is based on a substantial corpus of 12,200 categorized communication elements.
• Provides specific error rates: 40% of 2,500 controller elements and 59% of 5,900 pilot elements contained at least one error.
• Error analysis is detailed by speaker role (pilot/controller), facility, speech act, and topic.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Last update date is unknown; freshness unverified.

Provenance

Source

O. Veronika Prinzo

Collection Method

Analysis of voice communications parsed into elements and categorized using the ATSAT taxonomy.

Time Range

The time range of the recorded communications is not specified.

Freshness

Last update date is unknown; freshness unverified.

Geography

Data was collected from 3 terminal air traffic control facilities (TRACONs) in an unspecified location.