ACAS X, the Algorithm Replacing TCAS, and the Collision Avoidance Architecture Being Built for an Airspace Full of Aircraft That Have No Pilot Aboard

ACAS X - the Airborne Collision Avoidance System next generation - is the FAA- and ICAO-backed successor to TCAS, built to handle an airspace increasingly shared between airliners, drones, and autonomous vehicles. Where TCAS matched situations against a pre-computed table, ACAS X runs a continuous optimization that adapts in real time, incorporates GPS-derived position data, and extends to aircraft with no pilot aboard.

The Night TCAS Worked and People Died Anyway

On July 1, 2002, just after midnight Central European time, a Bashkirian Airlines Tupolev 154 and a DHL Boeing 757 converged over southern Germany. The Tupolev carried 69 people, including 52 schoolchildren bound for Spain. The Boeing was climbing northbound through the same airspace.

TCAS activated on both aircraft simultaneously. The Tupolev received a climb advisory. The Boeing received a descend advisory. The commands were coordinated - designed to resolve the conflict by separating the aircraft vertically.

The DHL crew flew their advisory immediately. The Tupolev crew did not. Zurich Area Control had just issued a descent instruction, and the crew followed the controller’s voice instead of the system.

Both aircraft descended. They collided at approximately 34,900 feet. Seventy-one people died.

TCAS did not fail at Überlingen. It issued correct, coordinated, complementary advisories. The failure was procedural: a crew trusted a human voice over an algorithm that could compute the geometry with more clarity than any controller in the loop. That lesson reshaped commercial aviation training globally and remains the foundation of how crews are taught to respond to Resolution Advisories today.

How TCAS Actually Works - And Where Its Architecture Breaks Down

TCAS Two was formalized in the early 1980s, mandated for most U.S. commercial aircraft by 1993, and eventually required globally by ICAO. It is now installed on essentially every commercial airliner flying.

The system works by interrogating nearby transponders. Your TCAS unit broadcasts radio signals; nearby transponders reply; the system measures reply timing to calculate range and uses directional antenna arrays for bearing. From that, it builds a surveillance picture of surrounding traffic.

A potential conflict triggers a Traffic Advisory (TA) first - “Traffic. Traffic.” - giving the crew situational awareness. If closure rates continue and projected miss distance drops within roughly 40 seconds, TCAS escalates to a Resolution Advisory (RA): “Climb. Climb.” or “Descend. Descend. Don’t deviate.” When two TCAS-equipped aircraft are in conflict, their units exchange coordination messages via a dedicated datalink. If one gets a climb advisory, the other gets a descent advisory. The maneuvers are designed to work together.

The system’s structural constraint is how those advisories are generated: a pre-computed lookup table.

When MIT Lincoln Laboratory engineers developed TCAS, they simulated thousands of encounter scenarios and computed the best response to each. That work was encoded into fixed avionics tables. During a real encounter, the aircraft doesn’t calculate anything - it matches the current situation to the nearest table entry and issues the stored advisory. This was the right design for 1980s hardware. But the table is immutable.

It cannot adapt to encounter geometries the designers didn’t anticipate. It cannot incorporate GPS-derived position data - GPS wasn’t a civil aviation tool when TCAS was designed. It ignores the precise position each aircraft is already broadcasting via ADS-B and instead builds its own picture from transponder timing alone. And it generates nuisance alerts on paired approaches to closely-spaced parallel runways, where two aircraft on adjacent final approach courses look geometrically threatening even when both are exactly where they belong.

Alert fatigue is a documented human factors finding, not a theoretical concern. Any safety system that produces false positives erodes the reliability of the response when a real alert fires. Reducing nuisance advisories is a safety improvement, not a comfort improvement.

The table also has an architectural ceiling that matters more now than it did in 1993: it was built for crewed aircraft. It has no mechanism to operate on an unmanned vehicle where no pilot is present to hear the advisory and respond.

What Makes ACAS X Fundamentally Different

ACAS X is being developed by MIT Lincoln Laboratory in coordination with the FAA and EUROCONTROL (the European Organisation for the Safety of Air Navigation). ICAO has formally adopted it as the future global standard for collision avoidance. Technical standards are published; certification work is underway.

The core departure from TCAS is how the advisory is generated.

ACAS X uses dynamic programming to solve what mathematicians call a Markov decision process - a framework for making optimal decisions under uncertainty when outcomes depend on both your action and on probabilistic future events. In the collision avoidance context, the system knows approximately where the conflicting aircraft is right now (with measurement uncertainty), approximately where each aircraft is heading, and not with certainty what the other aircraft will do next. It needs to compute the best response in the current moment while accounting for that full uncertainty picture simultaneously.

TCAS pre-computed that response decades ago and stored it. ACAS X runs the optimization continuously in real time, updating every fraction of a second as new surveillance data arrives.

When ADS-B data is available, ACAS X can use the other aircraft’s GPS-derived position directly rather than relying solely on transponder interrogation timing. The geometric picture is sharper, the uncertainty is smaller, and the advisory is more precisely calibrated.

MIT Lincoln Laboratory validates the algorithm by running it across tens of millions of encounter scenarios - including geometries no fixed table ever anticipated - looking for failure modes and edge cases. The published simulation results show ACAS X consistently producing fewer nuisance alerts than TCAS while maintaining equivalent or better collision avoidance performance. That combination is the engineering objective: fewer false alarms means crews trust the real ones, and trust is load-bearing in any safety system.

The ACAS X Family: Xa, Xo, and Xu

ACAS X is a family of variants, each designed for a specific operational context.

ACAS Xa is the direct successor to TCAS Two for large commercial transport aircraft. It is designed for backward compatibility: the verbal callouts are familiar, the display symbology fits existing crew mental models, and the transition from TCAS to ACAS Xa will not require crews to learn an alien system. What changes is the algorithm underneath and the performance that flows from it.

ACAS Xo addresses closely-spaced parallel runway operations specifically. It is built to understand the difference between a genuine conflict and two aircraft intentionally flying a managed paired approach procedure in close proximity - eliminating the nuisance alert scenario that TCAS handles poorly today.

ACAS Xu is designed for unmanned aircraft - and it may be the most architecturally significant variant for where aviation is heading.

ACAS Xu: Collision Avoidance Without a Pilot in the Loop

There is no pilot in an unmanned aircraft to hear an advisory and respond. ACAS Xu doesn’t issue verbal callouts. It communicates avoidance maneuvers directly to the flight management system, enabling autonomous response - or relays the advisory to a ground-based remote operator who may be managing multiple vehicles simultaneously.

That creates a layer of complexity TCAS never had to address. A remote operator receiving an ACAS Xu advisory for one of four vehicles they’re monitoring needs a response with enough precision that the operator’s reaction time doesn’t introduce a new hazard. The maneuver has to fit within the specific performance envelope of that particular unmanned platform. And the coordination has to work seamlessly alongside crewed aircraft carrying TCAS Two or ACAS Xa in the same airspace.

An unmanned cargo drone has a fundamentally different risk profile than a crewed airliner. Different performance envelope, different failure modes, different operational altitudes. ACAS Xu handles those differences while still coordinating correctly with the crewed aircraft sharing that airspace.

The existence of ACAS Xu is itself a design statement: the collision avoidance architecture now being built assumes not every aircraft in the sky has a human aboard.

Why the Timing Matters: What the Future Airspace Looks Like

The low-altitude airspace picture in ten to fifteen years is qualitatively different from anything the history of managed airspace has seen. Urban air mobility corridors in major cities. Commercial delivery drones at altitudes from a few hundred to several thousand feet. Electric air taxis sharing terminal airspace with helicopters and general aviation. Vehicle types, performance envelopes, and operational profiles with no real precedent.

A fixed lookup table designed around 1980s airliner encounters cannot be the collision avoidance backbone of that environment. The dynamic programming architecture of ACAS X - running continuously, adapting to actual encounter geometry, incorporating the best available position data, tunable across vehicle classes - matches the complexity that environment will generate.

When Will ACAS X Replace TCAS?

The honest timeline: ACAS Xa on new type certifications is probably the second half of the 2030s. Avionics certification at the highest design assurance levels takes years for legitimate reasons. The algorithm requires validation across millions of encounters. Failure modes must be fully characterized before the system enters service on commercial aircraft. That pace reflects the process working correctly, not an institutional bottleneck.

For most general aviation pilots, the near-term practical picture is limited. TCAS Two isn’t required in most GA aircraft; traffic awareness in GA panels comes from advisory systems and ADS-B traffic displays. The TCAS-to-ACAS X transition is a background story for GA operations today.

What This Means for Every Pilot Flying in Shared Airspace Now

The commercial aircraft above you are getting a rebuilt collision avoidance brain. The unmanned systems that will share your airspace are being engineered with ACAS Xu compatibility from the start.

The lesson from Überlingen doesn’t change with this transition - it sharpens. When the algorithm generating the advisory has fewer nuisance alerts, more precise geometry, and better position data than its predecessor, the case for trusting it unconditionally is stronger. Fly the advisory. The system sees the geometry more clearly than you do in that moment. That was true in 2002. It will be true when ACAS X is issuing the command.

Key Takeaways

TCAS has prevented countless mid-air collisions since its 1993 U.S. mandate, but its fixed lookup table cannot adapt in real time, integrate GPS position data, or function on unmanned aircraft
ACAS X uses dynamic programming - a real-time optimization updated continuously - replacing the stored table with an algorithm that adapts to actual encounter geometry and incorporates ADS-B position data
The ACAS X family includes Xa (commercial successor to TCAS), Xo (closely-spaced parallel operations), and Xu (unmanned aircraft with autonomous or remote-operator response)
Published simulations show ACAS X producing fewer nuisance alerts than TCAS while maintaining equivalent or better collision avoidance performance - a direct safety improvement, not just a performance upgrade
ACAS Xa certification on new commercial type certifications is expected in the second half of the 2030s; the architecture is being built now for an airspace that includes aircraft with no pilot aboard