Paving
Proximity warning and struck-by safety technology field guide for paving crews
Worker tags, cameras, and radar reduce the struck-by risk around heavy equipment, but they are the last layer. Separate people from equipment first, run the traffic plan and the spotter, and let the tech catch what slips through.
Direct answer
Struck-by safety technology is the proximity warning, camera, and radar systems that detect when a worker on foot enters the danger zone around heavy equipment and alarm the operator and the worker. It backs up the separation plan, it does not replace it. Separate people from equipment first, then let the tech catch what slips through.
Key takeaways
- Struck-by is one of OSHA's construction Focus Four, and roughly three out of four struck-by fatalities involve heavy equipment, most often backing or turning.
- Separate people from equipment first with an internal traffic control plan, spotter, and high-vis; proximity warning, camera, and radar are the last layer.
- If the operator cannot see the spotter, the machine does not move; the spotter is the control and the tech is the backup.
- Alarm fatigue is the single biggest reason proximity tech fails; fix it by tuning the detection zone, not by muting or disabling the system.
- Tag-only systems miss the untagged visitor, dead battery, or tag left in the truck; cameras and radar see whatever is there, tagged or not.
What struck-by safety technology is, and where it sits
Struck-by safety technology is the set of detection systems that warn an operator and a worker when the two get too close around heavy equipment. It covers proximity warning systems where the worker wears a tag the machine senses, cameras that show the operator the blind sides, and radar or ultrasonic sensors that flag an object behind or beside the machine. Some systems go further and slow or stop the machine when a person is detected in the danger zone.
The reason any of it exists is the proximity between people on foot and machines that cannot see them. Struck-by is one of the construction Focus Four, the four hazard groups OSHA ties to most construction deaths, and roughly three out of four struck-by fatalities involve heavy equipment. The killer is a worker standing in the spot the operator cannot see while the machine moves, most often backing or turning.
Here is the part that gets lost the moment a vendor demo lights up. The technology is the last layer, not the first. The thing that keeps people alive is separating them from the equipment in the first place, with a traffic plan, a spotter, high-vis, and eye contact. The tech is there to catch what slips through that, and a system bolted onto a site with no separation plan is a screaming alarm sitting on top of a hazard nobody engineered out.
The technology backs up the plan, it does not replace it
This is the one idea to carry out of the whole guide. Proximity warning and camera systems support the basics. They do not replace them. The hierarchy is separate people from equipment, run the traffic plan and the spotter, put everyone in high-vis, and then add the tech as a backup for the moment the plan breaks down.
The failure mode is predictable. A crew buys detection systems, mounts them on the fleet, and quietly relaxes the discipline that was actually keeping people alive. The spotter wanders off because the camera is there now. People cut through the equipment path because the tag will beep if it matters. That is exactly backward. Technology that breeds complacency makes a site less safe than no technology, because it moved the trust from a control that works to one that has a battery, a range limit, and a tag somebody might leave in the gang box.
Treat every detection system as a seat belt, not a substitute for driving sober. It is the thing that helps when everything else has already gone wrong. If the conversation on your site is which system to buy before the conversation about how to keep people and machines apart, the order is wrong and the money is going to the wrong layer first.
Why struck-by is a killer on a paving and earthwork site
Struck-by means a worker is hit by a moving or falling object: a vehicle, a piece of equipment, a swinging load, or something dropped from above. On a paving or earthwork job it is the moving iron that does the damage. Struck-by sits among the construction Focus Four and is one of the leading causes of construction death, behind falls, and the leading cause of nonfatal injuries on top of that.
The geometry is what makes it deadly. A dozer, a loader, a roller, or a haul truck weighs many tons and the operator sits high and forward, so a person standing close and low behind or beside the machine is simply not in view. The machine does not have to be fast. A roller idling backward at walking pace kills a worker who stepped into its path to grab a tool, and the operator never feels the difference.
Backing is the worst of it. A large share of equipment struck-by deaths happen while the machine is reversing or turning, with the worker in a blind spot, no spotter assigned, and no separation between the foot path and the equipment path. Those three gaps show up again and again, which is why the controls in this guide aim straight at them. The exact figures and the hazard categories are laid out in OSHA's struck-by materials, and the company safety plan should reflect them.
Separating people from equipment is the primary control
If a worker is never in the path of the machine, the struck-by cannot happen. That is the whole game, and it is why separation sits at the top of the controls and the tech sits near the bottom. Engineer the people and the equipment into different space and the hazard is gone, not just watched.
Separation is physical and procedural together. You set exclusion zones around working equipment that people on foot do not enter while it runs. You lay out one-way equipment routes and separate walking paths to and from the work. You mark no-go areas behind and around machines that back or swing. On a paving train you keep the ground crew clear of the roller path and the truck backing zone, and you make the access points to the mat deliberate instead of wherever someone feels like crossing.
The OSHA hierarchy of controls puts this in order: eliminate the exposure, then engineer it out, then control it with work practices and a plan, then fall back on PPE and warning devices. Separation is the elimination and engineering end. Spotters and the traffic plan are the work-practice layer. High-vis and proximity tech are the bottom. Spend your attention top down, because a control near the top removes the hazard while a control near the bottom only warns you about it. The specifics belong in the site safety plan and follow the OSHA framework.
The internal traffic control plan (ITCP)
An internal traffic control plan is the drawing and the rules that route equipment and people through the work area so the two stay apart. It is the engineered version of separation, written down before the shift instead of sorted out in the dust. The core of it is a diagram of the activity area showing where machines travel, where people walk, and where the access and egress points are.
A usable ITCP names the equipment routes and keeps them one-way where it can, so machines are not meeting and backing into each other or into the crew. It maps the foot paths to and from each operation and keeps them off the equipment routes. It marks the blind-spot zones and the no-go areas, sets where trucks back and who controls that move, and posts the signs and the access control that hold it together. On roadwork it ties into the external traffic setup, but the ITCP is specifically the internal side, the people and iron inside the work space.
The plan is worthless if the crew has not seen it. It has to be communicated to everyone on the site, including the truck drivers coming in, before work starts and again whenever it changes, because a driver who does not know the route is the one who backs into the lane the crew is standing in. ITCP guidance from FHWA and OSHA covers the layout and the practice; build yours to the site safety plan and the conditions of the actual job.
The blind spots and the kill zone around the machine
Every piece of heavy equipment has zones the operator cannot see, and those zones are where people die. On most machines the worst of it is directly behind and along the rear quarters, where the body of the machine, the engine deck, or the load block the view entirely. The area close to the machine on the ground, low and tight to the tires or tracks, is the kill zone, because it is both unseen and exactly where a person on foot ends up when they approach to work or to talk to the operator.
The blind spots differ by machine and they are larger than operators think. A haul truck backing has a deep blind area behind it that swallows a person standing in it. A loader has the bucket and the articulation point cutting the view. A dozer working in reverse covers ground the operator is craning to watch over a shoulder. A roller looks small but the operator is focused forward on the mat and the drum edge, and the back of the machine is an afterthought.
Map the blind spots for the machines on your job and treat the map as part of the plan. Walk each machine, sit in the seat, and have someone stand where the operator loses sight of them so the crew sees the zone with their own eyes. That walk does more for awareness than any poster, and it tells you exactly where the camera and the sensors need to cover.
What is a proximity warning system?
A proximity warning system, sometimes called a proximity detection system, senses when a person or another object enters a defined zone around a machine and alerts the operator, the worker, or both. It is the active detection layer that sits on top of the separation plan, watching the boundary the plan drew. The point is to catch the moment a person ends up where they should not be before the machine reaches them.
The common approaches split into a few families. Tag-based systems put a small tag on the worker, often in the hard hat or the vest, and an antenna on the machine that detects the tag inside a set range, using RFID for broader zones or ultra-wideband (UWB) for tighter, centimeter-level accuracy. Camera systems give the operator a live view of the blind sides, increasingly with AI object detection that flags a person on the screen. Radar and ultrasonic sensors detect an object behind or beside the machine without needing a tag at all. Many sites run a hybrid, a tag system for the crew and a camera or radar for the visitor who is not wearing a tag.
All of it is a warning layer. The system tells you a person is close; it does not move them, and on most installs it does not move the machine either. Hedge every system claim to the manufacturer's specification, because range, accuracy, and behavior vary widely between products, and a vendor's demo range is not your range in steel, dust, and weather.
Tag-based systems: the worker wears it, the machine senses it
In a tag-based system the worker carries a small transponder and the machine carries the detection antenna. When the tagged worker comes inside the machine's zone, the operator gets a visual and audible alert in the cab, and the worker's tag warns back, often by vibrating, so both people in the equation know at the same moment. That two-way alert matters, because the worker is the one who can step clear fastest.
RFID and UWB are the two common technologies, and they trade off range against precision. RFID has a long track record across construction, quarrying, and waste work and is good at covering a broader zone. UWB resolves position to the centimeter, which lets you tune tighter zones, separate a warning ring from a stop ring, and log exactly where the close calls happened. The choice depends on the machine, the spacing of the work, and what the manufacturer's system actually does, so match it to your situation rather than the brochure.
The weak point of any tag system is the tag. It only protects the worker who is wearing it and wearing it correctly, with a live battery, in the spot the system expects. A tag in a pocket, a tag left on the bench, a dead tag, or a visitor with no tag at all is invisible to the machine. That is the gap a camera or radar layer is meant to cover, and it is the reason the tag discipline has to be enforced like PPE, not treated as optional.
Cameras and radar on the blind sides
Camera and radar systems work on the machine instead of the worker, which is their advantage: they see whatever is there, tagged or not. A camera, often a rear view or a stitched 360-degree view, puts the blind sides on a screen in the cab so the operator can see the person the mirrors miss. Newer camera systems add AI object detection that picks a person out of the scene and flags them, which cuts the load on an operator who cannot watch a screen every second while running the machine.
Radar and ultrasonic sensors detect an object in the zone behind or beside the machine and warn the operator by sound and light, scaling the warning as the object gets closer. They do not care about dust on a lens or a worker who forgot a tag, and they hold up in conditions that blind a camera. What they give back in range and selectivity they lose in detail: a radar sensor tells you something is there, not that it is a person, so it will flag a cone, a pile, or a wall the same as a worker.
Most serious installs combine them. A camera so the operator can see and judge, radar so something is watching when the operator's eyes are forward, and on a tagged crew a proximity tag layer on top. The 360 view covers the close-in geometry around the machine; the rear and side sensors cover the deep blind zone behind. Set the coverage to the blind-spot map you built, not to a default the installer ships.
Systems that slow or stop the machine
The most active form of the technology does not just alarm, it intervenes. On detection of a person in the inner zone, the system slows the machine or brings it to a stop through an interlock to the controls, removing the assumption that the operator will react in time. This is the top of what the tech can do, and on the highest-risk machines and interactions it is worth the cost.
It also carries the most caution. An intervention system that stops a loaded haul truck or a working dozer at the wrong moment can create its own hazard, so the zones, the speeds, and the override have to be set with care and to the manufacturer's specification. There is usually a controlled override for the cases where the machine has to keep moving with a person knowingly present, and the override has to be a deliberate, logged action, not a switch the operator leaves flipped to make the nagging stop.
Most sites do not start here. Alarm-only systems are simpler, cheaper, and lower-risk to deploy, and for many interactions a loud, clear two-way warning is enough. Reserve slow-and-stop for the equipment and the moves where the consequence is worst and the warning alone has the least margin, and let the manufacturer and the site safety plan drive how the interlock and override are configured.
What is alarm fatigue, and why does it kill the system?
Alarm fatigue is what happens when a detection system alarms so often, and so often for nothing, that the operator stops reacting to it. This is the single biggest reason proximity technology fails in the field. A system that cries wolf every time a cone, a sign, a passing truck, or a worker fifty feet away trips a too-wide zone trains the operator to tune it out, mute it, or disable it. A muted system is no system, and now the crew is relying on a control that has quietly been switched off.
The root cause is almost always the detection zone, not the hardware. A zone set too large or too undifferentiated fires constantly during normal work, and the warning loses all meaning. The fix is tuning: set the zone to the real danger geometry of the machine, separate a wider awareness zone from a tight danger zone so the urgent alarm only fires when it should, and adjust it to how the machine actually works on this site. Tag systems with finer position resolution make this easier, but every system needs it.
Manage the nuisance alarms or the technology is worthless, and treat a disabled or muted system as a stop-work issue, not a quirk. The whole value of the layer is that the operator believes the alarm, and the first false alarm too many is the one that taught the operator not to. Tune it, watch the alert data for the zones that fire too much, and fix the configuration rather than letting the crew fix it by turning it off.
The spotter and the communication
A spotter is a dedicated worker, on the ground and clear of the path, whose only job in that moment is to guide a machine through a move the operator cannot see safely on their own. For a blind backing move, a swing in tight quarters, or a load near people, the spotter is the control, and the technology is the backup to the spotter, not the other way around.
The spotter works on agreed hand signals or radio, stays in the operator's view the entire move, and stops the move the instant they lose sight of the operator or the operator loses sight of them. The rule that keeps spotters alive is simple: if the operator cannot see the spotter, the machine does not move. The spotter never stands in the path or the pinch point to give a signal, and the operator never assumes the spotter is clear, they confirm it. Eye contact before the machine moves is not a courtesy, it is the verification.
The most dangerous moment is the one everyone treats as routine, the quick backing move where nobody bothered to set a spotter because it would only take a second. That is the move that kills, because it is the one where the operator is alone with the blind spot. A camera helps the operator see, but it does not replace the second set of eyes on the ground who can stop the move and pull a worker clear.
High-visibility PPE is the baseline
High-visibility clothing is the floor that everything else stands on. A worker the operator can actually pick out of the background has a chance the operator who sees a gray shape in the dust does not. High-vis does not separate anyone from anything, so it is near the bottom of the hierarchy, but it is cheap, it is always on, and it has no battery to die.
The garment class is matched to the exposure. Class 2 is the common baseline for workers on foot around traffic and equipment in daylight, and Class 3, with more material and sleeves, is used for night work, high-speed zones, and low visibility, where it is commonly required. Keep the garments clean and replace them when the reflective material is worn or grimed over, because a faded vest at night is most of the way back to no vest at all.
High-vis is the easiest control to take for granted and the easiest to let slip. It only works if it is worn, worn fully, and seen, so it is worth the same enforcement as a hard hat. Follow the site safety plan and the applicable high-visibility standard for the class and the conditions of the work.
The wearables: tag, high-vis, and beyond
The worker carries more of the safety system than they used to. The high-vis garment makes them visible. The proximity tag makes them detectable to the machine. Between those two, a worker on a modern site is both seen and sensed, and the two cover different failures: high-vis for the operator's eyes, the tag for the moment the eyes are pointed the wrong way.
Other wearables show up on heavy-equipment sites for related hazards. Airbag-style vests exist for some applications, and heat-stress monitoring matters on summer paving where the mat and the sun stack up and a worker in distress is a worker who stops paying attention to the machine behind them. These are adjacent to struck-by rather than central to it, but they ride on the same body and the same comfort and compliance problem.
The hard truth about wearables is that they only work when worn, charged, and used right. A vest left in the truck, a tag with a dead battery, a monitor someone took off because it chafed, all of them fail silently. Build the wearable check into the daily routine the way you check any other PPE, because a safety device the worker is not wearing is protecting the gang box.
The hierarchy of controls, applied to struck-by
The hierarchy of controls is the order you reach for solutions, most effective first, and it is the frame that tells you where the technology belongs. Eliminate the exposure where you can, engineer the separation where you cannot eliminate it, control what is left with administrative work practices and the plan, and put PPE and warning devices at the bottom as the last line.
Mapped onto struck-by it reads cleanly. Eliminate the proximity by keeping the work and the equipment in separate space and time. Engineer the separation with the internal traffic control plan, the exclusion zones, and the routes. Administer it with the spotter, the signals, the access control, and the training. Then, at the bottom, the PPE and the proximity tech: high-vis, the worker tag, the cameras, the radar, the slow-and-stop.
The proximity systems everyone wants to talk about live at the bottom of that list. That is not a knock on them, it is their correct place: they are the catch for what the higher layers missed, and they are most effective on a site where the higher layers are already strong. A site that spends at the bottom and skimps at the top has the hierarchy upside down. Build top down and let the framework, anchored in the OSHA hierarchy, drive the spend.
The operator side: seat belts, ROPS, and inspection
Struck-by is mostly about the person on foot, but the operator's own safety rides on the same machine, and the controls overlap. The seat belt and the rollover protective structure, the ROPS, are what keep an operator alive in a rollover or a tip, which is its own line in the equipment fatality count. The belt only works buckled, and the ROPS only works if it has not been modified or damaged, so both belong on the daily check.
A machine that is not maintained is a machine more likely to put someone in a struck-by situation: a backup alarm that quit, a camera caked over, a brake that grabs late, a horn that does not sound. The daily walkaround that catches those is the same one that keeps the fleet running, and the detail of running that program lives in the equipment fleet maintenance guide rather than here. The point for struck-by is that the safety devices on the machine are only as good as the inspection that confirms they still work.
Tie the proximity equipment into that same inspection. A detection system is PPE for the whole interaction, and like any PPE it gets verified before the shift: the camera shows a clear picture, the sensors power up and self-test, the backup alarm sounds, the operator's tag display is live. A system nobody confirmed is a system you are trusting on faith.
The data: alerts, near-misses, and the close-call map
The detection systems generate something the old controls never did: a record of every time a person and a machine got too close. The proximity alert log, the near-miss reports, and on the more capable systems a heat-map of where the close calls cluster are a leading indicator, a way to see the danger building before it becomes an incident instead of reading about it afterward.
Use the data to fix the plan, not to file it. If one corner of the site lights up the alert log every shift, the traffic plan routed people and iron through the same pinch and the fix is a route change, not a louder alarm. If one machine trips its system constantly, either the zone is mistuned, the false-alarm problem is starting, or that machine's interaction with the crew is genuinely too tight. The pattern in the data tells you which higher control to go back and strengthen.
Capture the alerts, the near-misses, and what you changed in response in a field record that travels with the job, so the next crew and the next shift inherit the lesson instead of relearning it. A tool like FieldOS can hold the alert log, the near-miss reports, and the close-call map alongside the rest of the job record so the leading indicator is something you actually look at, not data trapped in a box on a machine.
The work-zone side: public traffic and the road
On roadwork and paving in live lanes, the struck-by hazard has a second source: the public vehicles passing the work. This is the external traffic side, and it is its own discipline on top of the internal one. The car that drifts into the work space at speed is a different problem from the loader backing inside it, and it needs its own layer of protection.
The external controls are the temporary traffic control setup: the advance warning, the taper, the channelizing devices, the positive barrier where the speed and exposure justify it, and the flagger controlling the traffic at the access. Positive barrier is the engineered separation between the public traffic and the crew, the same idea as the internal exclusion zone but aimed at the road. The flagger is the human control at the boundary, exposed and reliant on the drivers seeing and obeying, which is why flagger positioning and an escape path matter.
Keep the two plans straight. The internal traffic control plan separates the crew from the construction equipment inside the work space; the temporary traffic control plan separates the work space from the traveling public. A paving job in a live corridor needs both, and the proximity tech on the machines does nothing about the car in the next lane. Build the external plan to the agency's traffic control standards and the site safety plan.
Training the operators and the ground crew
The technology is only as good as the people running it and working around it, and that is a training problem before it is an equipment problem. Operators have to know what their system covers and what it does not, how to read the alert without fixating on the screen, and that disabling or muting it is not their call to make. The ground crew has to know the traffic plan, the exclusion zones, the tag discipline, and the rule that a tag or a camera does not earn them a shortcut through the equipment path.
The no-disable rule deserves its own line in the training, because it is the rule most likely to be broken quietly. An operator annoyed by a system that cries wolf will turn it off, and the training has to make clear that the answer to a nuisance alarm is to report it and get the zone tuned, not to silence the layer the crew is counting on. Pair that with the discipline that the spotter and the plan stand whether or not the tech is working.
Put the struck-by hazard and the day's equipment moves into the regular toolbox talk and pre-task routine so it gets refreshed at the work face, not just at orientation. The mechanics of running a good job hazard analysis and toolbox talk are covered in the JHA and toolbox-talk guide; the point here is that the proximity plan has to ride along in that daily briefing, because the hazard and the layout change with every shift.
Where does proximity technology actually fail?
Proximity technology fails in ordinary, foreseeable ways, and an honest crew plans for them instead of trusting the system blind. The battery dies. The tag gets left in the truck or worn wrong. The range is shorter in steel and dust than the demo suggested, or the zone is set wrong. The camera lens cakes over with the same dust the job throws all day. The alarm fatigue sets in and the operator stops reacting, or someone mutes it. A visitor with no tag walks in invisible to a tag-only system.
Every one of those failures is silent. The system does not announce that it has stopped protecting anyone; it just stops, and the crew keeps acting as if it works. That is the specific danger of a bottom-of-the-hierarchy control: it warns, and when it quietly cannot warn, the false confidence it built is still there. A worker who trusted the tag and stopped watching the machine is worse off than one who never trusted it.
So treat the tech as a backup to the plan, never a substitute, and never let it relax the basics. Separation, the traffic plan, the spotter, the high-vis, and the eye contact have to hold whether the system is working or not, because some shift it will not be working and nobody will know until it matters. The technology buys you a margin on the days it works. It does not buy you the right to stop separating people from equipment.
Rolling it out without crying wolf
Do not blanket the fleet on day one. Start with the highest-risk equipment and the highest-risk interaction, usually the machines that back blind around people the most, the haul trucks and the loaders working tight to the crew. Pilot the system there, learn how it behaves in your dust and your geometry, and tune the zones before you scale it across the fleet.
The tuning is the work, and skipping it is how a rollout dies. Set the zones to the real danger geometry of each machine, separate the awareness alarm from the danger alarm, and run it long enough to find the spots that false-alarm before the operators decide for themselves that the system is junk. A pilot that earns operator buy-in because it alarms when it should and stays quiet when it should not is what carries the rollout. A pilot that screams all day teaches the whole crew to distrust the layer before it is even deployed.
Bring the operators into it early. They know where the machine is blind, where the nuisance alarms will come from, and what would actually help versus what just nags. A system the operators helped tune is a system they will keep on. A system imposed on them is a system they will quietly defeat, and the make-or-break is whether the people in the seats believe the alarm.
What to record
The record is what proves the plan existed before the incident, not after. Keep the internal traffic control plan and any revisions, the blind-spot maps, the proximity systems on each machine and their inspection and self-test results, the training and the tag-discipline sign-offs, and the alert and near-miss log with what you changed in response.
Tie it to the job rather than a binder nobody opens. A field tool like FieldOS can hold the ITCP, the equipment list and inspection checks, the toolbox and training records, and the proximity alert and near-miss data in one place that travels with the work, so the close-call map is something the next shift inherits and the documentation is there when an inspector or an investigator asks for it.
| Item | Requirement | Note |
|---|---|---|
| Internal traffic control plan | Drawn and communicated before work, updated on change | Routes, foot paths, access, no-go zones |
| Blind-spot map | One per machine type on the job | Walked with the crew, not just posted |
| Proximity systems | Listed per machine, inspected and self-tested per shift | Configure to manufacturer specification |
| Detection zone settings | Tuned to the machine and reviewed against alerts | The fix for alarm fatigue lives here |
| Spotter assignments | Named for blind and tight moves | Signals and stop-rule briefed |
| High-vis class | Class to match conditions, day and night | Replace when reflectivity is gone |
| Training and tag discipline | Operators and ground crew, no-disable rule | Refreshed in the toolbox talk |
| Alert and near-miss log | Captured and reviewed, with actions taken | Leading indicator, drives plan fixes |
Common mistakes
- Relying on proximity technology instead of separating people from equipment in the first place.
- Letting nuisance alarms breed alarm fatigue until the operator ignores, mutes, or disables the system.
- Running a site with no internal traffic control plan, so people and iron share the same space by default.
- Skipping the spotter on a quick blind backing move because it would only take a second.
- Treating the tag as optional, so workers leave it off or carry a dead battery and go undetected.
- Mounting tag-only detection and forgetting the untagged visitor who walks in invisible.
- Collecting the proximity alert data and never using it to fix the route or the plan.
- Letting the technology relax the basics, so high-vis, eye contact, and the plan quietly erode.
Field checklist
Want this checklist to run itself on every job — with photo proof and a signed record crews can hand the customer? That's FieldOS.
Standards and references
Struck-by is one of OSHA's construction Focus Four, and OSHA's struck-by materials and the construction standards set the framework for the hazard and the controls. The general layout of separating workers from equipment, the spotter, and the high-visibility apparel follow OSHA's requirements and guidance, with the high-vis class matched to the conditions. Treat OSHA as the floor and the company safety program and site safety plan as the specifics, since the plan is written to the actual job.
The internal traffic control plan has guidance from FHWA and OSHA, and on roadwork the external temporary traffic control follows the governing agency's traffic control standards and the state DOT specifications. The ITCP covers the internal separation of crew and construction equipment; the temporary traffic control plan covers the work space against the public traffic. A live-corridor paving job is built to both.
The proximity warning, camera, radar, and slow-and-stop systems are manufacturer products, and their range, accuracy, zones, interlock behavior, and override all follow the manufacturer's specification and installation instructions. Configure and inspect them to that specification, not to a generic default, and verify their performance on your site rather than trusting the demo. Hedge every system capability to the manufacturer, and let the site safety plan and the OSHA hierarchy of controls decide where each layer fits.
Terms and definitions
Struck-by safety carries its own vocabulary, and the same idea shows up under different names across a safety plan, a vendor sheet, and an OSHA document. These are the terms a foreman needs straight.
- Struck-by / Focus Four
- A worker hit by a moving or falling object: a vehicle, equipment, a swinging load, or a dropped object. One of OSHA's four construction hazard groups, the Focus Four, also called the Fatal Four.
- Proximity warning system
- A detection system that senses when a person or object enters a zone around a machine and alerts the operator, the worker, or both. Also called a proximity detection system.
- Internal traffic control plan (ITCP)
- The diagram and rules that route construction equipment and workers on foot through the work area so the two stay separated, with marked routes, foot paths, access points, and no-go zones.
- Blind spot / kill zone
- The area around a machine the operator cannot see, and the close-in ground where an unseen worker on foot is most likely to be struck, especially behind a backing machine.
- Worker tag (RFID / UWB)
- The transponder a worker carries so a machine's antenna can detect them. RFID covers a broader range; ultra-wideband resolves position to the centimeter for tighter, tunable zones.
- Alarm fatigue
- The state where too many false or nuisance alarms train the operator to ignore, mute, or disable the detection system, which makes the technology worthless. Fixed by tuning the detection zones.
- Hierarchy of controls
- The order of effectiveness for controlling a hazard: eliminate, then engineer, then administrative work practices, then PPE and warning devices. Separation is near the top; proximity tech is near the bottom.
- Spotter
- A dedicated ground worker, clear of the path, who guides a machine through a move the operator cannot see safely, on hand signals or radio, and stops the move the moment the line of sight is lost.
FAQ
What is a struck-by hazard in construction?
A struck-by hazard is the risk of a worker being hit by a moving or falling object: a vehicle, heavy equipment, a swinging load, or a dropped object. It is one of OSHA's construction Focus Four, and most struck-by deaths involve heavy equipment, often a machine backing or turning with a worker in the blind spot.
What is a proximity warning system?
A proximity warning system detects when a person or object enters a defined zone around a machine and alerts the operator and the worker. It uses a worker tag the machine senses, cameras showing the blind sides, or radar, and some systems slow or stop the machine. It is a warning layer on top of the separation plan.
Does proximity technology replace a spotter?
No. Proximity technology backs up the spotter, it does not replace them. The spotter is a dedicated ground worker who guides a blind move and can stop it and pull a worker clear, while the tech only warns. Cameras and tags fail silently, so the spotter and the traffic plan have to hold whether the system works or not.
What is alarm fatigue and why does it matter?
Alarm fatigue is when a detection system alarms so often, and so often for nothing, that the operator stops reacting and mutes or disables it. It is the main reason proximity tech fails. The fix is tuning the detection zone to the real danger geometry so the urgent alarm only fires when it should, not silencing the layer.
What is the best way to prevent struck-by incidents around heavy equipment?
Separate people from equipment first. An internal traffic control plan, exclusion zones, separate foot paths, a spotter for blind moves, and high-visibility clothing keep workers out of the machine's path, which is where struck-by happens. Proximity warning, camera, and radar systems are the last layer that catches what slips through, not the first control.
RFID or UWB for a proximity detection system?
RFID covers a broader detection zone and has a long track record in construction and quarrying. UWB resolves position to the centimeter, which lets you tune tighter zones, separate a warning ring from a stop ring, and map where close calls happen. Match the choice to the machine and spacing, and to what the manufacturer's system actually does.
What is an internal traffic control plan?
An internal traffic control plan, or ITCP, is the diagram and rules that route construction equipment and workers on foot through the work area so they stay separated. It maps equipment routes, foot paths, access points, and no-go zones, and must be communicated to the crew and truck drivers before work and whenever it changes.
Why do equipment blind spots cause so many fatalities?
The operator sits high and forward, so a worker standing low and close behind or beside the machine is out of view. Backing is the worst case: a large share of equipment struck-by deaths happen while reversing, with the worker in the blind spot, no spotter, and no separation between the foot path and the equipment path.
Can a proximity system stop the machine automatically?
Some can. The most active systems interlock to the controls and slow or stop the machine when a person is detected in the inner zone, instead of only alarming. It carries more risk, since a wrong stop can create its own hazard, so the zones and override follow the manufacturer's specification, and most sites reserve it for the highest-risk machines.