Connected jobsite, IoT, and RTLS field guide

The connected jobsite, and what it actually is

The connected jobsite puts sensors on the four things a project manages: the people, the equipment, the materials, and the environment they all work in. Those sensors feed data back to a platform, so the team manages by what the data shows instead of by walking the site and guessing. That is the whole idea. A superintendent who used to find a missing welder or an idle excavator by driving the site now sees both on a screen before the first coffee.

The categories are real-time location for assets and people, environmental sensors, equipment telematics, and wearables, all tied into one platform. The hard part is getting connectivity onto a temporary site that started with bare ground and no infrastructure. The sensitive part is putting sensors on the people. Both of those, not the sensors themselves, decide whether the system pays.

This guide covers the sensing, the network, the platform, and the privacy. The data it produces feeds the daily report and the schedule, and it sits next to the coordinated model that drives the install, so the BIM coordination and daily-report guides are the siblings to read with it. Where they end, this one picks up: the model is the plan, the daily is the record, and the connected jobsite is the live feed in between.

Why the value is the decision the data drives, not the data

The single number that decides whether a connected jobsite pays is not how many sensors you hang. It is how many decisions the data changes. A dashboard nobody acts on is wasted money, and the industry is full of them. Fleets run proven idle-time and theft dashboards that return their cost inside a month, and the gap is almost never the technology. It is whether anyone turns the reading into a call.

So the rule before you buy anything: collect what you will act on. If the data tells you a generator sat idle for 200 hours last month and nobody moves it, the sensor cost you money and saved you nothing. If it tells you the same thing and you send that generator to the job that is paying rent for one, the sensor paid for itself the first week.

Work backward from the decision. Name the call you want to make better, the lost tools, the idle iron, the missed pour temperature, then put in the one sensor that informs it. Data you do not act on is data you are paying to store and ignore. Every section that follows is a tool for a decision, and the question to keep asking is which decision it serves.

The categories: what each sensor type senses

Sorted by what they sense, connected-jobsite technology falls into a handful of categories, and a real deployment mixes several. The point of sorting them this way is that each category answers a different question, and you pick the category by the question you need answered, not by the gadget a vendor is selling.

Cameras for visual progress are the fifth category, and they earn their own treatment because the data they produce feeds the daily report more than any other sensor. The location, environmental, telematics, and wearable categories are the ones this guide works through in depth.

What is RTLS on a construction site?

Real-time location systems (RTLS) track where things are by putting a tag on the asset and reading that tag against fixed reference points. Tag a tool, a piece of equipment, a material lift, or a person, and the platform shows the position on a map of the site. The four jobs it does are find the asset, measure how much it gets used, flag when it crosses a boundary, and time how long work takes at a location.

Find-the-asset is the case that sells it. A crew burns real hours every week hunting the impact wrench, the total station, the welder that walked to another floor. Put a tag on the high-value, high-loss tools and that search collapses to a phone screen. The utilization case is quieter but worth more over a year: the data shows what actually gets used, which tells you what to stop renting and what to stop buying.

Tags read by radio, and the technology splits by range, accuracy, and cost. The common families are RFID, BLE, UWB, and GPS, covered next. The right tag depends on whether you need a yes-or-no at a gate or a sub-meter position indoors, so match the tag to the job and confirm the read range and battery life with the manufacturer for your site conditions.

RTLS tag technology: RFID, BLE, UWB, and GPS

There is no single best tag. There is the right tag for the read you need, and the trade is always range against accuracy against cost. Pick the cheapest technology that answers the question, because every step up in precision is a step up in tag price, infrastructure, or both.

RFID is the cheap, short-range workhorse: a passive tag costs cents and needs no battery, but it only tells you the asset passed a reader, so it suits check-in and check-out at a gate or gang box. BLE gives room or zone level, roughly 1 to 3 m, on a coin-cell battery that lasts years, which makes it the volume choice for tools and materials. UWB gives the precise position, on the order of 10 to 30 cm, but it needs fixed anchors and the tags cost more, so it is reserved for work where the exact spot matters. GPS covers the wide outdoors at a few meters, ideal for equipment and vehicles, but it is blind inside a structure where the steel blocks the sky. The figures below are typical, and the real numbers depend on the product and the site, so verify them with the manufacturer.

Environmental sensors: concrete, dust, noise, and water

Environmental sensors watch the conditions that affect the work and the people, and the readings turn into a quality decision or a compliance record. Temperature and humidity are the base measurement. From there the high-value ones are concrete maturity, respirable silica, noise, and water.

Concrete maturity is the strongest case, because it replaces a guess and a broken cylinder with a number. An embedded thermal sensor logs the in-place temperature history, and the maturity method, the framework in ASTM C1074, converts that history into an estimated strength so you can strip forms or post-tension on data instead of on a calendar. The concrete and post-tensioning guides go deeper on the maturity method itself.

Dust monitoring matters because respirable crystalline silica is regulated to a hard exposure limit under the OSHA silica standard, and a continuous particulate monitor is what proves you stayed under it. Noise monitors do the same job for local ordinances in dense areas. Water-leak sensors earn their keep on a data center, where a leak in or near a data hall is a catastrophe, so a sensor that catches the first drip pays for itself the one time it fires. Calibrate these against a known reference and trust the manufacturer's drift interval, because an environmental reading is only as good as the last calibration.

Equipment telematics: hours, idle, fuel, and location

Equipment telematics is the data the machine reports about itself: GPS location, engine hours, idle time, fuel burn, and fault codes off the engine controller. Most iron built in the last decade ships with a telematics unit, and the question is rarely whether the machine reports, it is whether anyone reads it.

The two decisions it drives hardest are utilization and maintenance. Idle time is the one that surprises people, because a machine can log more hours warming the seat than doing work, and every idle hour burns fuel and racks up hours against the next service. Fault codes and hour-based service intervals turn maintenance from a fire drill into a schedule, which is where the reported 30 to 50 percent cuts in unplanned downtime come from. Location closes the loop into theft and recovery.

The catch on a mixed fleet is that every manufacturer speaks its own data dialect, so pulling Cat, Deere, Komatsu, and rental units into one view means dealing with a standard or a middleman. The AEMP telematics standard, now carried as ISO 15143-3, exists to make that data portable across brands. Confirm what your machines and your platform actually support before you assume a mixed fleet shows up in one place, because the gaps are common.

Wearables: gas, falls, heat, and worker location

Wearables put the sensor on the worker, in a vest, a badge, a clip, or a hard-hat sensor, and they sense the things that hurt people: gas exposure, a fall, heat strain, and location for response. The output is a leading indicator and a faster rescue, not a productivity report, and keeping that distinction clear is what keeps the program alive.

A connected gas monitor that detects a hydrogen sulfide or low-oxygen atmosphere can alarm the wearer and the office at once, so a man-down in a vault is known in seconds instead of at the next head count. Fall detection and man-down sensors do the same for a worker who goes still. Heat-strain monitoring matters as the standards around heat illness tighten, because it catches the early signs before someone collapses.

The location side of a wearable is exactly where the privacy question gets real, covered later in this guide. The safety case is strong enough to carry the program, but only if the data is used for safety and response and not quietly repurposed into a timecard. The detailed safety-wearable selection and the proximity and avoidance systems belong with the broader safety-technology material; here the point is that wearables are a sensing category with a different rulebook than tracking a tool.

Connectivity is the hard part

The sensors are the easy part. Getting their data off a temporary site is the hard part, and it is where most connected-jobsite plans quietly fail. A jobsite starts as bare ground with no infrastructure, no wired network, often weak cellular, and no permanent power, and every sensor you hang is useless if its data cannot get back to the platform.

Two constraints drive every choice. The first is coverage across a site that is large, changing, and full of steel and concrete that block radio. The second is power, because most sensors run on batteries and the radio is the biggest battery drain, so a high-bandwidth link that dies in a week is worse than a low-bandwidth one that lasts a year. A camera streaming video and a temperature sensor sending a number every hour have completely different network needs, and one network rarely serves both well.

Plan the connectivity before you plan the sensors, not after. This is the step that separates a pilot that works from a pile of dead tags in a gang box. Treat the carrier coverage, the gateway placement, the power source, and the battery life as a real design problem, size it for your actual site, and confirm it with the network provider, because the conditions that decide it vary site to site.

Jobsite connectivity options compared

The options trade coverage against data rate against power, and most real sites run more than one. Low-power sensors ride a long-range, low-data network, while cameras and gateways ride cellular or a private network with the bandwidth to carry them. The split below is the usual shape, and the right mix depends on the site size, the carrier coverage, and the data each sensor sends, so size it with the provider.

LoRaWAN is the one that makes hundreds of simple sensors practical, because it sends small packets a long way on a battery that lasts years, and a gateway with a cellular backhaul puts a network on a site with no wired infrastructure. Private LTE or 5G is the heavier answer for a large site that needs site-wide, controlled coverage and is willing to stand up the infrastructure. Public cellular and Wi-Fi cover the high-bandwidth devices, and satellite is the fallback where there is no cell at all.

The data platform: one pane, not ten apps

The platform is where the sensors stop being a science project and start being a management tool. Its job is to pull every sensor type into one place, the locations, the environmental readings, the telematics, the wearables, so the team works from a single view instead of logging into ten vendor apps that do not talk to each other. Ten disconnected dashboards is the most common way a connected jobsite turns into a chore nobody does.

What the platform owes you is a single pane that shows the site, alerts that reach you when something needs a decision, and an API so the data flows into the systems you already run. The integration is the work. A location feed that does not reach the people who dispatch tools, or a telematics feed that does not reach the people who schedule service, is data sitting in a silo.

This is where a field tool like FieldOS earns its place, by being the one app the field already opens, so the sensor data lands next to the daily report, the photos, and the tasks rather than in a separate window. The test for any platform is simple: does the person who has to make the decision see the data in the tool they already use, or do they have to go find it. If they have to go find it, they will not.

Dashboards and alerts: tell me when something is wrong

A dashboard with a hundred charts is a dashboard nobody reads. The useful pattern is exception-based: most of the time the system stays quiet, and it speaks up when a reading crosses a line that needs a decision. The asset left the geofence. The concrete hit strength. The gas level spiked. Those are the moments worth an interruption, and the rest is reference you pull up when you have a question.

Design the alerts as carefully as the sensors, because alert fatigue kills more connected-jobsite programs than bad hardware does. Send an alert for everything and people mute the channel within a week, and then the one alert that mattered gets muted with the noise. Set the thresholds where a human actually needs to act, route each alert to the person who can act on it, and prune the ones that fire without a follow-up.

The charts still have a place. They are for the weekly look at utilization trends and the monthly review of idle and fuel, where a pattern over time drives a fleet or staffing decision. Live operations run on alerts. The trends run on the dashboard. Mixing the two, putting the slow pattern in an alert or the urgent event in a chart nobody checks, is how both get ignored.

What the connected jobsite is used for

The use cases are where the categories turn into money, and they rank loosely by how fast they pay. Utilization and theft tend to return their cost first because the savings are large and easy to see. The rest pay over a longer horizon in fewer surprises, fewer claims, and fewer failed inspections.

Pick one of these to start, not all of them. The table is a menu, not a checklist, and a deployment that tries to serve every row at once is the deployment that serves none of them well.

Asset utilization: you own and rent too much idle gear

Most contractors own and rent more equipment than the work needs, and they cannot prove it because nobody measures what actually runs. Asset utilization is the first place a connected jobsite usually pays, because the data turns a gut feeling into a fleet decision. The skid steer that you were sure was busy logged 12 percent active time last month. That is the number that ends an argument.

Utilization data drives three calls. It right-sizes the owned fleet, so the machine that sits gets sold or shifted to a job that needs it instead of bought again. It informs the rent-versus-own line, because a machine that runs constantly is a buy and one that runs a few weeks a year is a rent. And it cuts double-renting, where one job rents iron that another job has parked.

The reported fuel savings from cutting idle, often in the range of 15 to 25 percent, are real but they are the smaller prize. The bigger one is the capital you stop tying up in iron that does not work. Run the utilization report for a quarter before you approve the next equipment purchase, because the report usually finds the machine you already own.

Theft and recovery: the geofence alert

Jobsite theft is a large, quiet loss, with industry estimates putting equipment theft in the hundreds of millions to over a billion dollars a year in the United States and recovery rates low, commonly cited under 25 percent without tracking. A GPS or cellular tag on the machine changes both ends of that. It cuts the loss with a geofence and it lifts the recovery odds with a location.

The geofence is the practical tool. Draw a boundary around the site, and the platform alerts you the moment a tagged machine crosses it, especially after hours when nothing should be moving. That alert is the difference between knowing a loader is gone at 2 a.m. and finding out at 6 a.m. when the crew shows up to an empty pad. Speed is everything in recovery, because stolen iron moves fast and gets parted out or repainted within days.

Hide the tag and consider a second one. A thief who knows the game looks for the obvious tracker and pulls it, so the visible tag plus a hidden backup is the pattern that survives a determined theft. Confirm the device keeps reporting on its own power when the machine battery is cut, because a tracker that dies with the ignition is a tracker that helps the thief.

Safety: gas, proximity, man-down, and heat

The safety use cases are the ones that justify themselves the one time they fire, and they work as leading indicators rather than after-the-fact reports. A gas sensor that alarms before a worker enters a bad atmosphere prevents the incident. An after-incident log only documents it.

The high-value cases are gas detection in confined spaces and vaults, proximity and avoidance between people and moving equipment, man-down detection for a worker who falls or goes still, and heat-strain monitoring as the rules around heat illness tighten. Companies running connected safety systems have reported injury reductions in the 10 to 20 percent range, and the value of one prevented serious injury swamps the cost of the whole program.

The discipline that makes safety sensors work is acting on the leading indicator instead of filing it. A proximity system that beeps and gets ignored trains people to ignore it, the same failure as alert fatigue. The safety-technology and proximity guides cover the device selection and the response procedures in depth. The point here is that safety is a category with the clearest payback and the most direct path from a reading to a decision, which is a worker who goes home.

Tracking people is the sensitive part

Tracking equipment is a business decision. Tracking people is a trust decision, and getting it wrong does more than annoy the crew, it can break the law. The governing principle is plain: track the work, not the worker. Locate the zone for a safety response, log the presence for a compliance record, but do not turn a safety wearable into a surveillance tool, because the day the crew decides it is surveillance is the day they defeat it.

The legal exposure is real and it is not yours to wing. In the United States, monitoring that chills protected concerted activity can run afoul of the National Labor Relations Act, the EEOC has put employers on notice about wearable data, and biometric and location data carry their own state-by-state consent rules. A unionized site usually has to bargain the who, what, where, and when of any tracking before it goes in. None of that is settled by a vendor brochure.

So the hard hedge: this is a question for your labor counsel and, where a union is involved, the bargaining table, not for a project manager and a tech vendor. Write a policy, get informed consent where it is required, limit the data to the safety and operational purpose, and put it in front of the lawyers before the tags go on people. The technology is the easy part. The consent and the law are the part that ends programs.

Worker buy-in: explain the why

Even where tracking people is legal, it only works if the workers accept it, and acceptance is earned by explaining the why. A crew told nothing assumes the worst, that the tag is there to time their breaks and build a case to fire them. A crew told the truth, that the wearable calls for help if you go down in a vault and the location clears people away from a swinging load, has a reason to wear it instead of leaving it in the gang box.

The framing is safety and efficiency, not surveillance, and it has to be true to hold. Limit what the system records to the stated purpose, opt in where you can rather than mandate, and put the policy in writing so the rules do not drift after the program starts. Research on construction wearable adoption keeps landing on the same factors: perceived usefulness, the influence of peers, and privacy. Win those and people wear the device. Lose them and they defeat it.

Defeating it is easy and quiet. The tag goes in the gang box, the wearable stays in the truck, the sensor gets covered. You will not get an argument, you will get bad data, which is worse than no data because you will trust it. Buy-in is not a soft nicety on top of the rollout. It is the difference between a system that reflects the site and one that reflects an empty gang box.

Start with one painful problem

The way connected-jobsite programs fail is by trying to do everything at once. The way they succeed is by solving one painful problem first. Do not try to instrument the whole site on day one. Pick the one thing that costs you real money and real sleep, the tools that keep walking off, the equipment you suspect is idle, the pours you cannot strip on schedule, and put sensors only on that.

Run it as a pilot with a number attached. Decide before you start what success looks like, the search time you will cut or the rental you will drop, and measure against it. A pilot that proves a dollar figure is what funds the next phase, and a pilot that proves nothing because it tried to prove ten things is what gets the whole idea killed in a budget meeting.

Then scale to the next problem with what you learned. The connectivity you stood up for the first use case usually carries the second. The platform you chose for tools handles the materials. Growth from a proven base beats a big-bang rollout that strands you with sensors nobody reads and a network that does not reach. One problem, proven, then the next. That is the path that survives contact with a real jobsite.

Data hygiene: garbage data, garbage decision

A connected jobsite runs on the assumption that the data reflects reality, and that assumption breaks fast without maintenance. Garbage data drives a garbage decision, and a confident garbage decision is worse than no data at all, because you act on it. The work of keeping the data clean is unglamorous and it is the difference between a system you trust and one you quietly stop believing.

Three things go wrong. The tag ends up on the wrong asset, so the report tracks the trailer that is actually the generator. The sensor drifts out of calibration, so the temperature or the gas reading is wrong in a way nobody notices until it matters. And tags or wearables go dark, batteries dead, devices removed, so the map shows an empty site that is actually full. Each one produces data that looks fine and is not.

Build the upkeep into someone's job. Audit the tag-to-asset mapping on a schedule, follow the manufacturer's calibration interval for every environmental sensor, and run a report of devices that have stopped reporting so a dead tag gets caught instead of trusted. The decision is only as good as the data, and the data is only as good as the last time someone checked it.

Where the connected jobsite pays

The return shows up in a few places, and naming them up front keeps the program honest about what it is supposed to deliver. Utilization is usually the largest line, from selling or shifting idle iron and cutting rental you do not need. Theft avoided and theft recovered is the next, and it is lumpy: nothing for months, then a loader you did not lose.

Labor is real but harder to bank, because the search time you save the crew gets reabsorbed into other work rather than showing up as a smaller payroll. Safety pays in incidents that did not happen, which is the hardest number to put on a spreadsheet and often the largest when an insurance and claims year is added up. Environmental compliance pays as fines and stop-work orders avoided, and predictive maintenance pays as the unplanned breakdown that did not strand a crew.

The honest version of the ROI case ties each dollar to a decision the data changed, not to the existence of the data. A program that can point to the machine it sold, the theft it stopped, and the pour it stripped early has a number. A program that can only point to a dashboard has a cost. Build the business case on the decisions, and reverify it after the pilot with the real figures, because the savings vary by fleet and by site.

Tie the data into the workflow

A connected jobsite reaches full value only when its data flows into the work the team already does. The sensor feed is an input, and the outputs that matter are the daily report, the schedule, and the model. Left in its own app, the data is a curiosity. Wired into the workflow, it becomes the record and the plan updating themselves.

The cleanest loops are concrete. Concrete maturity and weather readings flow into the daily report as the conditions of record, so the log carries sensor data instead of a guess, which is exactly the contemporaneous, specific record the daily-report guide argues wins disputes. Equipment hours and presence inform the schedule, so the look-ahead reflects what is actually on site. And as-built location data from RTLS can be checked against the coordinated model, tying the field position back to the plan the BIM coordination guide describes.

Aim for the closed loop: the model sets the plan, the sensors report what is happening, the platform raises the exceptions, and the daily report captures the record, each feeding the next without re-keying. That loop is the connected jobsite working as a system rather than a pile of separate gadgets. The integration to a single field tool is what closes it, which is why the platform choice earlier in this guide is the decision the rest depends on.

Field checklist

What to document

The connected jobsite produces records that matter long after the sensor stops reporting, and the program needs its own paper trail beyond the live data. The decision case, the consent, and the network design are the documents that get pulled when a worker objects, an auditor asks, or a phase fails to scale, so capture them deliberately rather than after the fact.

Record what you deployed and why, who agreed to be tracked and on what terms, how the network is built, and what the program saved against the target it set. The data feeds the daily report and the maintenance log as it flows, so the documentation here is the system around the data, the part that proves the decisions were sound.

Common mistakes

• Collecting data nobody acts on, so a dashboard runs but no decision changes.
• Hanging sensors with no connectivity plan for a temporary site, so the data never gets off the ground.
• Running ten disconnected vendor apps instead of integrating to one platform.
• Tracking people without buy-in, consent, union review, or a privacy policy.
• Bad data hygiene, a tag on the wrong asset or an uncalibrated sensor, driving a confident wrong decision.
• Trying to instrument the whole site at once instead of solving one painful problem and proving it first.
• Treating a safety wearable as a productivity tracker, which kills the trust that makes it work.

Standards and references

Connected-jobsite technology pulls from several standards bodies and there is no single code that covers it, so cite the one that governs the specific point. For equipment telematics across a mixed fleet, the AEMP telematics data standard, carried as ISO 15143-3, defines how machine data moves between brands and platforms. For concrete maturity, the maturity method is framed in ASTM C1074. For respirable crystalline silica exposure, the OSHA silica standard sets the enforceable limit that dust monitoring is proving compliance against.

The radio technologies have their own references. RFID, BLE, UWB, and the LoRaWAN specification are defined by their respective alliances and standards groups, and the real range, accuracy, and battery life come from the product datasheet for your conditions, not from the headline number. For cellular and private LTE or 5G, the carrier and the network provider define coverage and capacity for your site. Hedge all of it to the manufacturer and the provider.

The privacy and labor side is governed by law, not by a technical standard, and it varies by jurisdiction. The National Labor Relations Act, EEOC guidance on wearable technology, and state biometric and location-privacy statutes can all apply to tracking people, and a unionized site carries bargaining obligations. Confirm the current rules with qualified labor and privacy counsel for your location, because this area is moving and the exposure is real. Data security across the whole platform follows your owner's and the project's information-security requirements.

Units, terms, and definitions

Connected-jobsite work mixes vocabulary from IT, surveying, and the trades, so the same idea shows up under different names across a vendor sheet, a spec, and a field conversation. The definitions below fix the terms this guide uses.

Location accuracy is given in meters or centimeters, range in meters or kilometers, and battery life in months or years. RTLS, RFID, BLE, UWB, and GPS describe location technologies by how they sense position. LoRaWAN and private LTE describe how the data travels. Keep the sensing and the network straight, because a tag's accuracy and a network's reach are separate questions that often get blurred together.

Connected jobsite / IoT

A site instrumented with internet-connected sensors on people, equipment, materials, and the environment, feeding data to a platform

RTLS

Real-time location system, tracking the position of tagged assets or people against fixed reference points

RFID vs BLE vs UWB vs GPS

Location technologies ordered by precision and cost: RFID is cheap read-or-no-read, BLE is zone level, UWB is sub-meter, GPS is outdoor

Environmental sensor

A device measuring site conditions such as temperature, humidity, concrete maturity, silica dust, noise, or water

Equipment telematics

Machine-reported data including GPS location, engine hours, idle time, fuel, and fault codes

LoRaWAN / private LTE

LoRaWAN is a long-range, low-power network for small sensor packets; private LTE or 5G is a high-capacity network you control on site

Data hygiene

Keeping the data accurate: correct tag-to-asset mapping, calibrated sensors, and live devices, so decisions are sound

Asset utilization

How much an owned or rented asset is actually used versus sitting idle, used to right-size the fleet

Related tools

Calculators and readiness checks for this work

Compare your options

FAQ

People also ask

• What is BIM coordination?
• What goes in a construction daily report?
• What should you photograph on a jobsite?
• What is construction robotics?
• What is the OSHA noise action level?