Key Takeaways
- 0.5% of all reported aviation accidents in the U.S. (1982–2016) involved collisions between aircraft, indicating the relative rarity of in-flight collision events compared with other accident types
- 3.9% of U.S. fatal general aviation accidents in 2016 involved midair collisions (including collisions with ground/aircraft in flight categories used by NTSB summaries)
- 58.0% of U.S. midair collision accidents between 1992–2011 occurred in visual meteorological conditions (VMC)
- The global TCAS market has expanded due to equipage and safety mandates; a report projects TCAS/ACAS integrated safety systems market reaching $3.1B by 2030 (forecast)
- The global ADS-B market is projected to reach $2.7B by 2031 (forecast), reflecting broad deployment relevant to collision avoidance via surveillance
- The global air traffic management (ATM) market is projected to reach $38.3B by 2030, with surveillance and safety systems contributing to collision-avoidance capabilities
- IATA’s 2023 report cited that 88% of airlines plan to invest in digital transformation initiatives affecting safety operations support systems
- MITRE’s 2022 evaluation framework for detect-and-avoid systems emphasizes that sensor fusion improves probability of timely conflict detection relative to single-sensor detection
- A 2018 peer-reviewed study in Aerospace Medicine and Human Performance found significant gaze/search limitations in see-and-avoid tasks, supporting the need for alerting systems to prevent midair collision
- A 2017 study in Ergonomics reported that human ability to detect small aerial targets degrades sharply with closure rate and target angular size, increasing midair collision risk in visual acquisition tasks
- A 2020 review article in Safety Science highlighted that see-and-avoid remains unreliable in high workload/complex airspace, motivating surveillance-based collision avoidance
- TCAS resolution advisory effectiveness has been evaluated in multiple studies; one peer-reviewed analysis reported that when resolution advisories are followed promptly, collision probability is substantially reduced
- A 2009 journal paper in Reliability Engineering & System Safety quantified that layered safety nets (surveillance + alerting) reduce top-event probability compared with single-layer systems
- A 2015 NASA report on detect-and-avoid demonstrated that sensor fusion and alerting thresholds improve probability of detection relative to single sensors in simulated encounters
- The global air traffic control systems market is projected to reach $24.7B by 2028 (forecast), reflecting investment in safety and conflict-management capabilities
Midair collisions are rare, but TCAS and surveillance improve detection and dramatically cut risk.
Related reading
01 · Category
Safety Incidents6 stats
Safety Incidents Interpretation
02 · Category
Technology Adoption3 stats
Technology Adoption Interpretation
03 · Category
Industry Trends2 stats
Industry Trends Interpretation
More related reading
04 · Category
Human Factors6 stats
Human Factors Interpretation
05 · Category
Risk Reduction7 stats
Risk Reduction Interpretation
06 · Category
Cost Analysis3 stats
Cost Analysis Interpretation
Midair collision context: where it happens and what helps
Midair collisions are relatively rare overall, but when they occur they’re concentrated in certain operational conditions (e.g., VMC and uncontrolled airspace). Surveillance/alerting and TCAS/ACAS approaches are reported to substantially reduce risk.
Cite This Report
This report is designed to be cited. We maintain stable URLs and versioned verification dates. Copy the format appropriate for your publication below.
Christopher Morgan. (2026, February 13). Mid Air Collision Statistics. Gitnux. https://gitnux.org/mid-air-collision-statistics
Christopher Morgan. "Mid Air Collision Statistics." Gitnux, 13 Feb 2026, https://gitnux.org/mid-air-collision-statistics.
Christopher Morgan. 2026. "Mid Air Collision Statistics." Gitnux. https://gitnux.org/mid-air-collision-statistics.
Sources & references
27 datasets cited across this report · attribution is report-level
+11 additional datasets cited (not shown individually)

