TCAS to ACAS X, the Climb-Climb-Now Voice in the Cockpit, and the Optimized Logic That's Teaching Collision Avoidance to Turn Left Instead of Just Up

SUMMARY: How aircraft collision avoidance works, why crews must “fly the RA,” and how ACAS X is replacing TCAS for a sky full of drones and air taxis.

When two aircraft are converging on the same piece of sky, a collision avoidance system issues coordinated commands to each cockpit - telling one crew to climb and the other to descend - so they pull apart safely even when neither pilot can see the other. Today that system is TCAS (Traffic Alert and Collision Avoidance System), and it’s being replaced by a smarter, more flexible successor called ACAS X. The shift matters because the old logic can only send aircraft up or down, while the new system can also command turns - the half of the sky TCAS was never able to use.

What Is TCAS and How Does It Work?

The system in most airliners flying today is the Traffic Alert and Collision Avoidance System, or TCAS. The most common misconception about it is fundamental: TCAS does not use radar, and it cannot see out the window. It has no eyes at all.

What it uses is the transponder. Every aircraft in controlled airspace squawks on a transponder, replying to ground interrogations with its altitude and identity. TCAS flips that arrangement around. Instead of only answering the ground, your aircraft begins interrogating others, listening for the replies from every transponder-equipped aircraft nearby, and calculating range, closing speed, and altitude from the timing of those replies.

In effect, the aircraft builds its own radar-like picture entirely out of everybody else’s transponders. This also exposes the system’s blind spot: if a target is not squawking - no transponder, or the transponder is off - TCAS cannot detect it. It can only protect you from aircraft it can hear.

Why TCAS Cares About Time, Not Distance

Once TCAS is tracking a target, it doesn’t just measure position. It measures how fast the gap is closing and calculates time to the closest point of approach.

Distance alone is meaningless without speed. Two aircraft a mile apart and drifting slowly are nothing. Two aircraft ten miles apart and closing head-on are an emergency. What matters is time, and the system’s alerts work in tiers on a clock.

At roughly 40 seconds from the closest point of approach, TCAS issues a Traffic Advisory (TA) - the “traffic, traffic” call. It commands nothing. It tells you to get your eyes outside and get ready.

If the geometry keeps developing, at roughly 25 to 30 seconds out, TCAS escalates to a Resolution Advisory (RA). Now it stops suggesting and starts commanding: climb, descend, maintain vertical speed, or level off - precise instructions for the pitch input needed to miss the other aircraft.

How Do Two Aircraft Avoid Being Told to Climb Into Each Other?

This is the most elegant piece of the engineering. When two aircraft both carry TCAS and both are about to issue an RA against each other, the two systems negotiate.

Through the Mode S datalink, one transponder talks directly to the other, and in a fraction of a second they agree on who climbs and who descends. That coordination is why you never get the nightmare scenario of both aircraft being told to climb into one another. The maneuvers are always complementary - one up, one down - and neither crew ever knows the negotiation happened. It’s invisible, and it just works.

The Überlingen Collision: Why Pilots “Fly the RA”

The importance of that coordination was written in blood on July 1, 2002, over Überlingen, in southern Germany near the Swiss border, at night.

A DHL Boeing 757 cargo jet and a Tupolev Tu-154 charter carrying a group of Russian schoolchildren ended up on converging paths at the same altitude. Both aircraft had TCAS, and both systems worked exactly as designed. They coordinated: the 757 was told to descend, and the Tupolev was told to climb. Had both crews followed their Resolution Advisories, the aircraft would have missed cleanly.

But a single air traffic controller was working alone, overloaded, with equipment partly out of service. At almost the same moment TCAS told the Tupolev to climb, the controller told the Tupolev to descend. Trained - as many crews were in that era - to trust the controller, the Russian crew descended. The DHL 757 was already descending, following its own RA.

Both aircraft descended into the same point in the sky. Seventy-one people died, most of them children.

Out of that catastrophe came one of the most important rules in modern aviation: when you get a Resolution Advisory, you follow the aircraft, not the controller. The RA wins. The two boxes are coordinated with each other; the controller, however skilled, is not part of that negotiation and cannot see what your TCAS agreed to with the other aircraft. Today, “fly the RA” is doctrine everywhere. Überlingen is why.

What Are the Limitations of TCAS?

When everyone follows it, TCAS works beautifully. But it has real limits.

It only maneuvers vertically. TCAS will tell you to climb or descend, but it will never tell you to turn. Vertical guidance is simpler and more predictable, and aircraft change altitude in cleaner, more measurable ways than they change heading. But it means the system is throwing away half the sky - sometimes the safest escape is a turn, and TCAS cannot offer one.

It generates nuisance alerts. In busy airspace, with aircraft legitimately leveling off a thousand feet above or below each other, the older logic can get jumpy. It sees a fast climb toward your altitude and doesn’t fully trust that the other aircraft will stop where it should, so it alerts. Every false alarm chips away a little at how much crews trust the box.

It was built for a world of airliners. TCAS assumes big aircraft, transponders, professional crews, and predictable performance. It was never designed for a sky about to fill with drones and electric air taxis.

How ACAS X Is Different: Optimized Logic Instead of Hand-Written Rules

The logic inside classic TCAS is essentially a giant set of hand-written rules. Over decades, engineers coded “if this geometry, then this response.” It’s deterministic and it works, but it’s rigid and extremely hard to extend - every new rule can have side effects on all the others.

So engineers asked a different question: what if we stopped writing the rules by hand? The answer is ACAS X (Airborne Collision Avoidance System X), and it thinks in a fundamentally different way.

Instead of thousands of if-then rules, ACAS X is built around a giant precomputed lookup table. Engineers modeled collision avoidance as a problem of decision-making under uncertainty. They can’t know exactly what the other aircraft will do next, but they can assign probabilities - the chance it keeps climbing, the chance it levels off. For every possible situation, they calculate the response that best balances two competing costs: the catastrophic cost of a collision against the disruptive but survivable cost of a nuisance maneuver.

That optimization runs offline, on powerful ground computers, across millions of simulated encounters. The result is a lookup table. In the air, the system simply measures the situation and reads the pre-calculated best action. The heavy thinking has already been done.

Much of this optimized-logic work traces back to research at MIT Lincoln Laboratory and to engineers like Mykel Kochenderfer, who helped show that framing collision avoidance as a probability problem could beat decades of hand-tuned rules. The results are striking: better protection and fewer unnecessary alerts at the same time - a trade engineering usually forces you to make in one direction only.

The ACAS X Family: Xa, Xo, Xu, and Xr

ACAS X isn’t a single system. It’s a family, and that’s the whole point.

ACAS Xa - “a” for active. The mainline, direct replacement for TCAS on airliners.
ACAS Xo - “o” for operations. Tuned for specific procedures such as closely spaced parallel approaches, where older logic panics at aircraft that are supposed to be near each other.
ACAS Xu - “u” for unmanned. Collision avoidance for drones, and the breakthrough here is that Xu can command horizontal maneuvers - it can tell an aircraft to turn. A drone flying low or operating in tight airspace often needs to turn, not just climb, to avoid a threat.
ACAS Xr - “r” for rotorcraft, increasingly meaning eVTOL electric air taxis. Picture hundreds of small electric aircraft operating low and close together over cities. A system built for airliners at 30,000 feet is useless there; Xr is being shaped for slow speeds, tight spaces, vertical takeoffs, and horizontal escapes.

How ACAS X Uses ADS-B

There’s one more shift. ACAS X is designed to lean on Automatic Dependent Surveillance–Broadcast (ADS-B). Rather than only interrogating everyone’s transponders and cluttering the radio spectrum, the new system can passively listen to the ADS-B position reports aircraft already broadcast.

That’s a quieter, more efficient way to build the same picture - less interrogation traffic choking the spectrum, more aircraft sharing their own precise position. Surveillance and collision logic finally begin working as one system instead of two.

When Will ACAS X Replace TCAS?

The standards already exist. The technical documents defining ACAS X, published through RTCA - the standards body that writes this material for the FAA - are complete. The math is validated and the simulations have run. In the lab and in the standards world, this is real.

But collision avoidance is the most conservative corner of an already conservative industry, and it should be. This is a system that can command an airliner full of people to dive or climb; you don’t push a software update on a Tuesday. Every version must be tested against millions of encounter scenarios, certified, and carefully rolled into a global fleet where the old and new systems have to coexist and coordinate for years during the transition.

Realistically, this is a gradual changeover measured across the better part of a decade, not a switch that flips. The mainline airliner version (Xa) comes first, while the drone and air taxi versions (Xu and Xr) mature alongside the aircraft and regulations they’re meant to serve.

The deeper reason ACAS X matters is that the old system had hit a wall. You cannot hand-write rules forever, and you cannot stretch a vertical-only, airliner-only system to cover a sky full of drones and air taxis that need to turn, not just climb. ACAS X is the architecture that finally scales to the airspace we’re actually building - and in a sense, it started over Überlingen, with the hardest possible lesson about what happens when the machine and the human disagree.

Key Takeaways

TCAS uses transponders, not radar, to track nearby aircraft - so it is blind to any aircraft that isn’t squawking.
The system alerts on time to closest approach: a Traffic Advisory around 40 seconds out, a Resolution Advisory around 25–30 seconds out, with two aircraft’s boxes coordinating complementary climb/descend commands over the Mode S datalink.
The July 1, 2002 Überlingen collision killed 71 people and established the global rule to always “fly the RA” rather than a conflicting controller instruction.
ACAS X replaces hand-written rules with a precomputed, probability-optimized lookup table, delivering better protection and fewer false alerts at once.
The ACAS X family scales collision avoidance to drones (Xu) and eVTOL air taxis (Xr), adding horizontal turn commands and leaning on ADS-B - capabilities TCAS never had.