ACAS X, the Next Generation Collision Avoidance System Replacing TCAS II, and the Markov Decision Process That Changes the Math Behind Every Resolution Advisory

ACAS X replaces TCAS II's fixed logic tables with a Markov Decision Process, cutting nuisance resolution advisories by 20–25% while improving safety performance.

Aviation Technology Analyst

ACAS X - the Airborne Collision Avoidance System X - is the FAA- and Eurocontrol-backed successor to TCAS II, the collision avoidance system that has protected transport category aircraft since the mid-1980s. Developed over more than a decade at MIT Lincoln Laboratory, ACAS X replaces the deterministic logic architecture at the heart of TCAS II with a mathematical framework called a Markov Decision Process. The result is a system that is fundamentally better at distinguishing real conflicts from encounters that are close but benign. ACAS Xa, the transport category variant, reached ICAO Standards and Recommended Practices status in 2022.

Why TCAS II Was Always Going to Have Limits

TCAS II was not invented in a vacuum. It was demanded by tragedy. The 1978 midair collision over San Diego - a Pacific Southwest Airlines Boeing 727 and a Cessna 172 - killed 144 people. The 1956 Grand Canyon collision killed 128 and triggered decades of airspace reform. By the time the FAA mandated TCAS II in transport category aircraft in the early 1990s, the need was undeniable.

The system is built on a logic table. Engineers mapped every possible encounter geometry between two aircraft, then hardcoded a response for each one. Put in the same inputs, get the same output - every time. That deterministic architecture was an elegant solution for the hardware capabilities of the 1980s and early 1990s.

The problem is that aviation kept getting more complicated. Busier airspace, different autopilot behaviors, reduced vertical separation minima, and eventually unmanned aircraft created a class of encounter the original logic table wasn’t designed to handle cleanly. The result: nuisance resolution advisories - alerts that fired on encounters that were close but not genuinely conflicting.

Studies of TCAS II performance, including analysis by Lincoln Lab itself, found crews sometimes not following resolution advisories. Not out of negligence, but because the alert didn’t match what the crew could see. A collision avoidance system only works when pilots trust it enough to comply without filtering the advisory through their own judgment. When that trust erodes, the system’s safety case erodes with it.

How the Markov Decision Process Changes the Architecture

The Lincoln Lab team didn’t set out to rewrite the logic table with better numbers. They replaced the architecture entirely.

A Markov Decision Process (MDP) is a tool for making optimal decisions under uncertainty over time, named for Russian mathematician Andrei Markov. Instead of matching a state to a fixed response, the MDP defines what outcome to minimize - in this case, collision probability - models the uncertainty in all inputs, and computes the best policy across all possible futures simultaneously.

Applied to collision avoidance, ACAS X doesn’t look up an answer. It asks: given everything known about where the other aircraft probably is, where it’s probably going, and how much uncertainty exists in those estimates, what action minimizes collision probability while also minimizing unnecessary alerts?

The critical architectural difference is that TCAS II operates on point estimates while ACAS X operates on probability distributions. TCAS II interrogates a transponder, gets a reply, computes a closure rate, and applies the rule. ACAS X treats the other aircraft’s position and trajectory as a probability cloud and reasons across the entire cloud when deciding whether to issue an advisory.

FAA modeling, based on simulation of large encounter databases, projects ACAS X could reduce nuisance resolution advisories by 20 to 25 percent compared to TCAS II, while maintaining or improving actual safety performance. That reduction isn’t just a performance metric - it’s a safety metric. Fewer false alerts means higher crew trust in the alerts that do fire.

The ACAS X Family: Three Variants for Three Operational Environments

ACAS X is not a single product. It is a family of standards.

ACAS Xa is the direct replacement for TCAS II in transport category aircraft. From the pilot’s perspective, the experience is structurally similar - same resolution advisory voice commands, same compliance expectation. Underneath, the mathematics are fundamentally different. ACAS Xa can receive ADS-B position reports and fold that data directly into its probability model. In environments without ADS-B, it falls back to transponder interrogation, the same mode TCAS II uses today.

The integration of ADS-B data is significant. Because ACAS Xa can receive ADS-B broadcasts passively rather than relying solely on active interrogation, it can build a more accurate geometry picture earlier in an encounter. Earlier detection enables gentler maneuvers - a nudge at 30 seconds to conflict is far less disruptive than a hard turn at 15.

ACAS Xu is designed for unmanned aircraft systems. Because there’s no pilot to respond to a verbal advisory, ACAS Xu interfaces directly with the autopilot or flight management system and commands the avoidance maneuver autonomously. The system is necessarily more conservative in certain respects - there’s no human with situational awareness who could recognize a bad call and override it.

ACAS Xo addresses specific operations like closely-spaced parallel approaches and formation flying, where standard collision avoidance logic would fire constantly due to intentional proximity. ACAS Xo can suppress alerts within defined operational parameters while still protecting against genuine departures from the planned geometry.

The existence of Xu and Xo reflects where aviation is heading. Collision avoidance is no longer about protecting two transport jets from each other at flight level 350. It’s about protecting an increasingly diverse airspace population where unmanned and manned aircraft share altitude bands.

Where the Transition Actually Stands

ACAS Xa reached ICAO SARPS status in 2022. That means the technical standard exists and has international consensus. It does not mean hardware is in production or that an operator mandate is imminent.

Avionics transitions are slow. The original TCAS mandate took years to implement. The upgrade to TCAS II took years after that. Airlines replacing working, certified avionics require a compelling regulatory or operational reason to spend the money.

New aircraft deliveries will likely incorporate ACAS Xa first. Retrofit mandates, if they come, will be phased over years to give operators time to plan and budget. TCAS II will continue flying in operational aircraft for years to come.

What This Means for Pilots

For pilots flying transport category equipment, ACAS Xa is coming. The pilot-facing experience will be largely transparent - the transition should require no fundamental change in how crews respond to resolution advisories. The nuisance rate will be lower. The system will handle complex traffic environments with greater precision. The fundamental rule remains unchanged: when the system issues a resolution advisory, comply immediately. That imperative is as important under ACAS Xa as it is under TCAS II.

For Part 91 general aviation pilots, the more immediately relevant point is understanding what current technology does and doesn’t do. The ADS-B traffic picture on a panel is situational awareness. It is not a collision avoidance system and was never designed to tell a pilot when and how to maneuver. The majority of mid-air collisions in the light aircraft fleet occur in visual conditions, below 10,000 feet, frequently near airports in the traffic pattern - an environment where tight geometry and variable closure rates limit what any electronic system can accomplish. See-and-avoid, attention management, and traffic pattern discipline remain the primary safety tools in that environment.

For anyone watching the unmanned aircraft integration story, ACAS Xu is infrastructure still being tested against real-world mixed-airspace operations. How it performs at real density with real edge cases is something the entire industry will be watching.


Key Takeaways

  • ACAS X replaces TCAS II’s fixed logic table with a Markov Decision Process that reasons across probability distributions rather than point estimates, making it fundamentally better at distinguishing genuine conflicts from close-but-benign encounters.
  • FAA modeling projects a 20–25% reduction in nuisance resolution advisories under ACAS X, with maintained or improved safety performance.
  • ACAS Xa (transport category), ACAS Xu (unmanned systems), and ACAS Xo (formation/parallel approach operations) each address a distinct segment of the evolving airspace population.
  • ACAS Xa achieved ICAO SARPS status in 2022; full fleet transition will be gradual, with TCAS II remaining operational for years.
  • The core pilot action doesn’t change: when a resolution advisory fires, comply immediately - regardless of what you think you see outside.

Radio Hangar. Aviation talk, built by pilots. Listen live | More articles