One-Lane, Two-Way Traffic Control: The Five Methods and When Each Fits

Every traffic control company runs one-lane, two-way operations, and most of them run one method for everything because that is what the trucks are set up for. That is usually two flaggers, sometimes for years, on jobs where a signal or a pilot car would have been cheaper, safer, or both.

The national standard actually recognizes five methods plus a no-control condition. Knowing all six changes what you can bid, what you staff, and what you tell a customer when the job turns out to be longer than the plan assumed.

This walks through each, what the standard permits and limits, and the operational trade. Requirements vary by state, county, and city, and they change. Many states adopt their own manual or a supplement, and your contract documents and the sealed plan for the job govern on top of all of it. Verify with the agency having jurisdiction.

The governing rule

Chapter 6E of Part 6 of the Manual on Uniform Traffic Control Devices sets the frame. When traffic in both directions must use a single lane for a limited distance, movements from each end must be coordinated. That is the requirement; everything else is how you satisfy it.

Guidance is that provisions should be made for alternate one-way movement through the constricted section, and it names the methods: flagger control, a flag transfer, a pilot car, traffic control signals, or stop or yield control. It adds a practical siting note — control points at each end should be chosen to permit easy passing of opposing lanes of vehicles. A control point on a blind curve or in a spot with no room for two vehicles to clear each other creates a standoff.

Version matters. The 11th Edition was published in December 2023 and took effect January 18, 2024, with Revision 1 published and effective March 5, 2026, per the FHWA MUTCD site. States get two years from an effective date to adopt an edition, so at any moment some agencies enforce the current national manual and others a state document that has not caught up. The broader tour of Part 6 is in our MUTCD Part 6 overview.

Why any of this matters in dollars and in injuries: FHWA records 821 fatal work zone crashes in 2022, with rear-end collisions accounting for 21 percent of them, per FHWA's work zone facts and statistics. A one-lane, two-way operation manufactures the exact conditions that produce rear-end crashes — a stopped queue on a road where drivers do not expect one.

Method one: flagger control

The workhorse. Traffic is controlled by a flagger at each end of the constricted section, one of them designated the coordinator, and the two able to communicate orally, electronically, or by manual signals — with the caveat that those manual signals should not be mistaken for flagging signals. Somebody waving at the far flagger with the paddle is how a release gets misread.

A single flagger is permitted when the zone is short enough that one person can see from one end to the other. That flagger should be stationed on the shoulder opposite the constriction or work space, or in a position where good visibility and traffic control can be maintained at all times. When one station cannot maintain that, the standard sends you back to a flagger at each end.

The drift problem is real: a zone that was short enough at 7 a.m. gets longer through the day, and nobody re-checks whether the single flagger can still see the far end. That is a supervisor check on a schedule, not a one-time decision at setup.

Where the flaggers stand is its own subject with its own requirements — stopping sight distance, escape routes, advance warning signs, and night illumination. We covered it separately in flagger station setup and escape routes.

Cost shape: two people for the duration, plus radios. Highest labor cost of the five methods, and the most flexible — a flagger can pass an emergency vehicle, hold for a driveway, or shut the operation down instantly.

Method two: flag transfer

The driver of the last vehicle proceeding into the one-lane section is handed a red flag or other token and asked to deliver it to the flagger at the other end. When that flagger receives it, they know traffic can move the other way. A variation replaces the token with an official pilot car following the last road user through the section.

Guidance puts a hard practical bound on it: the flag transfer method should be used only where one-way traffic is confined to a relatively short length of road, usually no more than one mile.

The reason is arithmetic. Neither end can release until the token arrives, so cycle time is bounded by transit time. Stretch it and queues build on both ends while both flaggers stand there waiting on a stranger to do their job.

It also depends on a member of the public completing a task. Most do. The one who does not — who turns off at a driveway inside the zone, or drives off with your flag — stalls the operation until somebody drives the token back.

Cost shape: same two flaggers, but it removes the need for reliable radio communication. Genuinely useful in terrain where radios do not work, which is exactly where crews tend to improvise something worse.

Method three: pilot car

A pilot car may be used to guide a queue of vehicles through the zone or a detour. Three things are required rather than suggested:

• The PILOT CAR FOLLOW ME sign must be mounted on the top or on the rear of the pilot vehicle.
• The pilot car operation must be coordinated with flagging operations or other methods of control at each end of the one-lane section. The pilot car is not a replacement for control at the ends; it is added to it.
• If an AFAD is used in a pilot car operation, it must be operated by a flagger positioned near and within the line of sight of the device, and it must not be left unattended at any time it is in use.

Guidance adds that the pilot car should prominently display the name of the contractor or the contracting authority, and that where temporary signals are used in pilot car operations and drivers will face long waits, consideration should be given to signing the wait time or the pilot car operation.

That last point is underrated. A driver who knows they are waiting six minutes behaves very differently from a driver who thinks the road is broken. Signed wait times reduce the u-turns, the shoulder runs, and the argument at the flagger station.

Pilot cars earn their keep on long zones, on shifting alignments where the correct path is not obvious, on gravel or milled surfaces, and any time the through-path changes as work progresses. They also set the speed of the platoon, which is a safety benefit no cone provides.

Cost shape: a vehicle plus a driver on top of end control. More expensive per hour, and often cheaper per job than the alternative, because platooning moves more vehicles per cycle.

Note the interaction with automated flagger assistance devices: the AFAD does not save the headcount in a pilot car operation, because the standard requires an attending flagger anyway. What it buys is distance between that person and the traffic stream.

Method four: temporary traffic control signals

Traffic control signals may be used to control vehicular movements in a one-lane, two-way zone. The manual treats this as an option and points to its signal chapter for the device requirements.

This is the method most under-used by contractors and most liked by agencies on long-duration work. Once a portable signal pair is set and timed, it runs without a person in the traffic stream, through the night, in weather, without breaks, and without the fatigue curve that degrades flagging late in a shift.

The trade-offs are honest ones. Signals need power and monitoring. They need timing that matches actual zone length and travel speed, and a bad cycle time creates queues a flagger would have prevented by simply looking. They do not adapt — a flagger sees the funeral procession, the school bus, the ambulance; a signal does not. And they are capital, so they only pencil on jobs long enough to amortize the mobilization.

The duration categories in Section 6N.01 are a useful screen. Work occupying a location more than three days is long-term stationary; more than one daylight period up to three days, or nighttime work over an hour, is intermediate-term. Once you are in long-term territory on a two-lane road, a signal usually beats standing two people in the road for weeks. More on the device side in portable traffic signals in work zones.

Cost shape: capital and mobilization up front, near-zero labor after. Inverts the economics of long jobs.

Method five: stop or yield control

STOP or YIELD signs may be used to control traffic on low-volume roads at a one-lane, two-way zone when drivers are able to see the other end of the operation and have sufficient visibility of approaching vehicles. If the sign is installed for only one direction, guidance is that it should face the road users driving on the side of the roadway closed for the work activity area.

Every condition in that sentence is load-bearing. Low volume. Drivers can see the other end. Sufficient visibility of approaching vehicles. Strip any one of those and the method stops being appropriate — this is not the cheap default for a road that happens to be quiet at 10 a.m.

Cost shape: the cheapest of the five. Two signs and no standing labor, on the narrow set of sites that qualify.

The sixth condition: self-regulating

The manual also allows for no positive control at all. If the work space on a low-volume street or road is short and road users from both directions are able to see traffic approaching from the opposite direction through and beyond the worksite, movement through the constriction may be self-regulating.

Worth knowing because it is a legitimate option a plan may call for, and worth treating carefully because "the neighbors will figure it out" is not the same as meeting the stated conditions. Short. Low volume. Mutual visibility through and beyond the work space.

The taper at the ends

Whichever method you pick, the geometry at the control point is the same. The one-lane, two-way traffic taper runs a minimum of 50 feet and a maximum of 100 feet under the national taper criteria, and unlike a merging or shifting taper it does not scale with the computed taper length L, per Part 6 of the 11th Edition. Its purpose is to define a stopping point, not to move traffic laterally at speed. If you want the full taper arithmetic, it is in taper length and buffer space explained.

Advance warning is what stops the rear-end crash, and on higher-speed roads it has to start early. The manual's guidance for rural highways is that the first warning sign should be placed 8 to 12 times the speed limit in feet, with the advance warning area extending 1,500 feet or more for open highway conditions.

Choosing, in practice

A rough decision order that matches how the economics fall out:

1. Does the site meet the stop-or-yield conditions? Low volume, mutual visibility. If yes, it is the cheapest compliant answer.
2. Is this long-duration? More than a few days on a two-lane road, and a temporary signal usually beats standing labor.
3. Is the through-path long, shifting, or unobvious? Pilot car.
4. Do radios work here? If not, and the zone is under a mile, flag transfer is a real option instead of improvising.
5. Otherwise, flagger control — and check whether one flagger can genuinely see end to end before staffing two.

The reason this matters commercially is that most companies quote method-blind. The estimate assumes two flaggers because that is the habit, and then a three-week job carries six weeks of flagger labor that a signal pair would have covered. Getting the method into the quote — and into the rate sheet, so the customer sees why a signal job prices differently — is the difference between a bid that wins and one that wins and makes money. Our notes on quoting jobs faster and estimating basics cover the mechanics.

Three things every method has to handle

Whichever of the five you run, the same three situations will show up and the method has to have an answer before they do.

Emergency vehicles. An ambulance or fire apparatus arriving mid-cycle has to get through, and the answer differs sharply by method. A flagger sees it coming and can hold both ends. A signal cannot — somebody has to physically intervene, which means a signal operation on a route with meaningful emergency traffic needs a person available and a documented procedure. Stop-or-yield control leaves it entirely to drivers. This belongs in the pre-job discussion with the agency, not in the moment.

Driveways and side roads inside the zone. Every one is a place a vehicle can enter the one-lane section without passing a control point, or leave it carrying your flag transfer token. On residential and commercial frontage this is the single most common reason a clean method degrades into improvisation. Count the accesses during the site visit, not on the morning of.

Queue length. A one-lane operation manufactures a stopped queue on a road where drivers do not expect one, and the back of that queue is the exposure. If the queue reaches back around a curve or over a crest, the advance warning that was correct at setup is now protecting the wrong point. Somebody needs to be watching queue growth, and the plan needs a threshold at which you change something — longer cycles, an additional advance sign, or shutting down during peak. Queue-warning technology exists for exactly this problem on larger jobs, covered in smart work zones and queue warning systems.

Where mobile work sits

One-lane, two-way control assumes a fixed constriction. Work that moves is a different regime — Section 6N.01 treats mobile operations as work that moves intermittently or continuously, with the advance warning area moving along with the work area. That includes the rolling roadblock, a method that uses pacing vehicles to create a gap in traffic so work can proceed downstream, normally run by law enforcement during off-peak hours. If your job is really a moving operation rather than a fixed closure, start with moving operations and the rolling closure and short-duration and mobile work zones instead.

Recording what you ran

Whichever method the plan called for, the thing that matters afterward is whether you can show what was actually on the ground. Which method, who was on each end, when the operation started and stopped, whether the pilot car ran, whether the signal was timed and by whom, and dated photographs from the driver's approach.

That record is data, and on federal-aid work it sits inside a larger obligation — the traffic control plan is a component of a required transportation management plan under 23 CFR 630 Subpart J, with both the agency and the contractor required to designate a trained person responsible for implementing it. A signed, GPS-stamped daily ticket carries the time and place it was made, which a paper ticket filled in on Friday does not. Additional work zone crash and safety data for toolbox talks is at the National Work Zone Safety Information Clearinghouse.

Five methods, one requirement: coordinate the movements from each end. The method you pick is an engineering decision, a safety decision, and a margin decision at the same time, and most companies only ever treat it as the first two.