Tube counts vs. radar counters: which one fits your link?

Most capacity studies still come down to a choice between two field methods: lay tubes across the lane, or set up a radar unit on the shoulder. Both get you volumes and some speed data. Neither is free, and the wrong pick for a given link wastes a field day you won't get back before the study deadline.

What pneumatic tube counts actually get you

Tubes are the default for a reason. Two rubber hoses stretched across the travel lanes, a counter box chained to a sign post, and you've got axle strikes converted to vehicle counts by class, with per-lane and per-direction breakdowns if you stagger the tubes right. On a rural two-lane or a collector with low pedestrian conflict, a crew can lay a station in twenty minutes.

The catch shows up on anything with real traffic. Tubes shift under truck axles, kids walk off with them, and a crew working the shoulder of a 45-mph arterial is a safety exposure your agency's traffic control plan has to account for, not skip. Multi-lane counts get noisy too: axle hits from adjacent lanes bleed into each other unless you offset the tubes carefully, and class counts on anything with heavy truck mix need a second pass to sanity-check. You also lose the count the day someone's plow catches a tube edge, which happens more than vendors like to admit.

Where radar detectors earn their keep

Radar units solve the "crew on the shoulder" problem. Mount it on a pole or a tripod set back from the curb, point it down the corridor, and it reads speed and presence per lane without anyone touching the pavement. That makes it the better call on signalized arterials, ramps, and anywhere a tube install means a lane closure or a flagger.

Radar has its own blind spots. Stopped or very slow traffic, the exact condition you're often trying to characterize at a signalized intersection during peak, is where Doppler-based units struggle most: a car idling at a red light doesn't generate the return a moving one does. Lane discipline near a merge or a wide shoulder can also put a vehicle in the wrong bin. And you're still buying or renting the unit, still sending someone to place it and come back for it, and still troubleshooting a unit that went dark overnight because the battery pack didn't hold a charge in cold weather.

The gap neither method covers well

Both methods assume the link is worth the field trip: the mobilization, the traffic control, the equipment that might walk off or die mid-count. For a link in your network that's never been counted, carries modest volume, and isn't scheduled for a signal retiming this cycle, that math rarely pencils out. So it stays uncounted, and the next capacity memo either borrows a count from a similar-looking link three miles over or flags the segment as a data gap and moves on.

That's the real cost of both tube counts and radar vehicle detectors: they're built for links where someone already decided the count mattered enough to justify a crew. For the long tail of links that nobody's instrumented and nobody wants to send a truck to, a one-pass vehicle count built from a single scheduled image gets you a tally by link and direction without laying a tube or installing a sensor, and without the lane closure paperwork either method requires.

A field note on picking between the two

If you're choosing between tubes and radar for a scheduled study, the deciding factors are usually mundane: lane count, posted speed, whether you need vehicle classification or just volume, and whether your traffic control crew has a slot open that week. Tubes still win on cost and classification detail for low-volume, low-speed roads. Radar wins when the install itself is the risk. Neither one is going to get cheaper to deploy on a link that was never going to get a crew in the first place, which is where it's worth checking what a single overhead pass can cover instead.