AGV Robot

What is an AGV Robot

An Automated Guided Vehicle is a driverless, battery-powered industrial transport vehicle that moves loads along predefined, pre-engineered paths inside warehouses, factories, and distribution centers without an onboard operator. It belongs to the Storage & Warehouse Equipment family within Logistics & Packaging, and it is one of the foundational building blocks of intralogistics automation alongside conveyors, stacker cranes, and shuttle systems. The single defining trait of an AGV is that it follows a fixed, physically defined guidepath or a pre-mapped fixed route and takes its missions from a central fleet controller, rather than freely planning its own route around obstacles the way a fully autonomous robot does.

In safety-standard language, AGVs are not called "robots" at all. They are formally termed "driverless industrial trucks" under ISO 3691-4 and "driverless, automatic guided industrial vehicles" under ANSI/ITSDF B56.5. This matters for procurement, because the certification, type-examination, and conformity documents a supplier provides are written against those formal categories, not the marketing word "robot." When a tender specifies AGV safety compliance, it is the truck-safety standards that apply.

The newer term AMR (Autonomous Mobile Robot) refers to free-roaming SLAM-based robots that build their own map of the environment and re-plan their path on the fly. Modern AGVs and AMRs increasingly share natural-feature and SLAM navigation, so the line between the two is genuinely blurring; many vehicles sold today as "AGVs" are in fact SLAM-based. Despite that convergence, classic AGVs still imply fixed routing and central dispatch, and that distinction drives both the navigation technology you choose and the fleet-management software you will run.

Functionally, an AGV combines five subsystems working together. (1) A navigation and guidance subsystem determines where the vehicle is and which path to follow. (2) A drive subsystem moves it, using steered or differential wheels, or omnidirectional Mecanum wheels on advanced units. (3) An onboard controller or PLC sequences the mission. (4) A safety subsystem of scanners, bumpers, and emergency-stop devices protects people and equipment. (5) A wireless link connects to a central fleet-management and dispatch system that assigns missions and manages traffic, intersections, and battery scheduling. The vehicle continuously senses its position relative to the guidepath or map, corrects steering, decelerates and stops at programmed points, performs load transfer, and reports status.

Photo: MakoGomez90, CC BY-SA 4.0, via Wikimedia Commons

The operating cycle is worth tracing once in full, because it is what every subsystem exists to support. On receiving a mission, the vehicle confirms its current position against the guidepath or map, accelerates to its zone-appropriate speed, and tracks the route while its safety scanners sweep speed-dependent protective fields ahead of it. As it nears a programmed point it decelerates, stops within its rated tolerance, and executes the load transfer, whether that is rolling a tote onto a conveyor, forking a pallet into a rack slot, or lifting a cart from beneath. Throughout, it reports status, battery state, and faults back to the fleet controller, which is simultaneously sequencing every other vehicle, resolving intersection conflicts, and deciding when each unit should peel off to charge. The vehicle is therefore never acting alone; its behavior is the product of onboard control and central traffic management together.

Because an AGV is a system rather than a single device, the buying decision is never just about the vehicle. The same load can be moved by very different machines depending on whether you tow it, carry it on a deck, fork it into racking, or slide under it and lift. The chapters that follow work outward from the vehicle types to the navigation that distinguishes them, then to the materials, the spec parameters that appear on datasheets, and finally a structured selection sequence you can reuse as an RFQ template.

AGV Types by Load-Handling Function

AGVs are most usefully classified by how they handle the load, because that determines the mechanical deck, the transfer method, and ultimately which vehicle matches your dock-to-rack flow. Six load-handling families cover the great majority of industrial deployments. Choosing the wrong family is the most expensive selection mistake in this category, because the load-transfer interface, not the vehicle chassis, is what ties an AGV to your existing conveyors, stands, and racking.

Towing or tugger AGVs are the original and still very common type. They pull one or more trailers or carts in a "train" along milk-run routes, which makes them strong for high-volume, repetitive line-feed and replenishment in just-in-time manufacturing. Tow capacities span roughly 2,000 lb up to 60,000 lb and beyond, with specialty heavy tuggers rated far higher. Because the load rides on its own trailers, a single tugger can service many drop points on one loop, giving the best throughput per vehicle when the route is long and the stops are many.

Unit load AGVs carry a single discrete load, a pallet, tote, or roll cage, on a deck. They are often equipped with a powered roller, belt, or lift-deck transfer mechanism that automatically loads and unloads to belt conveyors and stands, and an RFID-tagged load ID is common so the fleet controller always knows which load is on which vehicle. Unit load AGVs are the workhorse for point-to-point pallet and tote moves between fixed stations.

Forklift or fork AGVs are automated counterbalance, reach, or straddle forklifts that pick, transport, and stack pallets both at floor level and up into racking. They use high-resolution encoders and fork-height sensing to place loads precisely into rack positions. Fork AGVs are the right answer when you need to interact with conventional pallet racking rather than purpose-built transfer stands, and they often replace a manned forklift on a repetitive shuttle route.

Under-ride or latent (lift) AGVs are low-profile units that drive beneath a cart or rack, lift it, and carry it. This is the "goods-to-person" pattern popularized by the Kiva model, in which a fleet of compact robots brings shelving units to stationary pickers. Because the vehicle tucks entirely under the load, this family achieves very high storage density and is well suited to high-throughput e-commerce fulfillment.

Heavy-load or carrier AGVs are custom platforms built for very large loads, such as coils, molds, energy-storage racks, and aerospace fuselage sections. Payloads here can reach tens to hundreds of tonnes, far beyond standard pallet vehicles, and the chassis, drive, and steering are engineered per project. Assembly-line AGVs are a distinct application of the carrier idea: they serve as moving assembly platforms that carry a product, for example an automobile body, through workstations at a controlled pace, replacing a fixed conveyor line with a reconfigurable fleet.

Photo: AGVExpertJS, CC BY-SA 3.0, via Wikimedia Commons

Navigation and Guidance Technologies

Navigation is the core differentiator between AGV models, and it is the single most consequential technology choice in the category. The guidance method decides how much you must modify the floor, how easily you can reroute when the layout changes, how accurately the vehicle stops, and how much it costs to commission. Seven guidance technologies dominate the market, ranging from fixed physical guidepaths to fully infrastructure-free mapping.

Wire-guided (inductive) navigation runs a low-frequency, low-voltage current through a wire buried in the floor; the resulting magnetic field is tracked by an onboard induction coil. It is extremely robust and immune to surface contamination, but the path is fixed, and re-routing requires re-cutting the floor to move the wire. It is best for permanent trunk routes that will not change for years.

Magnetic tape or magnetic spot guidance has the AGV follow adhesive magnetic tape laid on the floor, while discrete magnetic spots or RFID tags provide absolute position references and command points for stop, turn, and transfer actions. It is low cost and easy to lay, which makes it popular for first deployments, but the tape wears under traffic and rerouting means relaying tape. Optical or line-following guidance is a related low-cost method in which a camera or photosensor follows a painted or taped colored line on the floor.

QR-code or 2D-code grid navigation places floor-mounted QR codes that give discrete absolute positions, while inertial dead-reckoning bridges the gaps between codes. This is the common pattern in goods-to-person fulfillment grids, where a dense matrix of codes lets a large fleet of under-ride robots localize cheaply and reliably across a structured floor.

Laser triangulation (reflector) navigation uses an onboard rotating laser scanner that measures angles to fixed retro-reflective targets mounted on walls and columns; the controller triangulates the vehicle's position from at least three reflectors. It delivers high accuracy and flexible routing without modifying the floor, but it requires installing and surveying the reflectors, and it depends on clear lines of sight to them. Inertial (gyroscope) navigation combines a gyroscope with wheel odometry and is usually paired with floor magnets, transponders, or QR codes to correct the drift that accumulates in any dead-reckoning system.

Natural-feature or SLAM navigation (laser or 3D-vision) is the most flexible method: the vehicle builds and matches against a map of existing walls, racks, and features, using no reflectors or tape at all. This enables free routing and easy layout changes, and it is the convergence point with AMRs; many vehicles sold today as AGVs are SLAM-based. The trade-off is that performance depends on a feature-rich, reasonably stable environment, and large open or constantly changing spaces can challenge map matching.

In practice the choice follows the facility. A simple, permanent route favors wire or tape for cost and robustness; an accuracy-critical fixed plant favors reflector triangulation; a structured high-density fulfillment grid favors QR codes; and a facility whose layout changes often favors infrastructure-free SLAM. The decision is rarely about which technology is "best" in the abstract, but about which matches your floor condition, accuracy target, and rate of change.

Materials, Media, and Power

An AGV's chassis, wheels, and energy system have to survive years of continuous duty under load, so material and power choices are as important to total cost of ownership as the navigation technology. The chassis and frame are typically welded carbon steel or other structural steel for load rigidity. Where the application demands it, stainless steel or aluminum is used instead: stainless for cleanroom and hygiene duty, aluminum where weight is the constraint. The choice trades rigidity and cost against corrosion resistance and mass.

Drive wheels use polyurethane (PU) treads, which give good grip while protecting the facility floor from marking and wear. The wheel and drive configuration is itself a design variable: vehicles use differential, tricycle, quad-steer, or omnidirectional Mecanum arrangements, with Mecanum reserved for advanced units that must maneuver in very tight spaces and move sideways without turning.

Energy storage has shifted decisively from lead-acid to lithium-ion, predominantly Lithium Iron Phosphate (LiFePO4, also written LFP). LFP is favored for its cycle life, fast charging, and crucially its support for opportunity and auto-charging without a dedicated off-board battery room. That single change has reshaped AGV operations: instead of swapping heavy lead-acid batteries in a ventilated charging room, modern fleets top up in place during natural dwell time, which sharply raises uptime. The same chemistry also tolerates partial-state-of-charge cycling far better than lead-acid, which is exactly what opportunity charging demands.

The "media" an AGV handles is the unit load itself rather than any process fluid: pallets, totes, roll cages, racks, coils, molds, and fixtures. This framing matters for selection because the load's weight, dimensions, fragility, and transfer interface drive both the deck design and the safety-field configuration. A vehicle that carries a stable palletized load has very different requirements from one that lifts a tall, top-heavy rack, even at the same nominal payload.

The shift to LiFePO4 also changes how a fleet is sized over its life. Because LFP supports opportunity and auto-charging, a fleet no longer needs one spare battery per vehicle plus a staffed swap room; instead it needs charging contacts positioned at natural dwell points along the route. That moves cost out of consumable batteries and floor space and into infrastructure that pays back through higher availability. The practical consequence for procurement is that battery chemistry, charging strategy, and route design should be specified together rather than as separate line items, since each constrains the others.

Material and power decisions interact with the environment in ways that should be settled early. Cold-chain and freezer duty constrains battery performance and demands appropriate seals and lubricants; cleanroom and ESD areas drive the stainless or aluminum chassis and conductive wheels; hazardous or ATEX zones impose explosion-protection requirements on the whole vehicle; and outdoor or sloped operation affects traction, gradeability, and ingress protection. Settling these constraints before model selection prevents the common and costly mistake of specifying a vehicle that cannot survive its own operating environment.

Key Specification Parameters

Reading an AGV datasheet is a core skill for purchasing engineers, because suppliers describe the same vehicle in different ways and a handful of parameters do most of the selection work. The ones that truly drive a decision are payload, travel speed, positioning accuracy, battery and charging strategy, drive configuration, and the governing safety standard. The table below summarizes the typical ranges; always verify against the specific model, since several parameters are highly model-dependent.

Payload and load capacity span an enormous range and depend entirely on type. Light unit-load AGVs carry roughly 100 to 500 kg; standard pallet AGVs handle about 1,000 to 1,500 kg; and heavy-load and tugger variants go far higher, with heavy-duty carrier platforms reaching tens to hundreds of tonnes. Tugger tow ratings are specified separately as a tow capacity, commonly from 2,000 lb to 60,000 lb and beyond. When comparing models, confirm whether a figure is a deck payload or a tow rating, because the two are not interchangeable.

Travel speed is commonly up to about 1.0 to 1.5 m/s, which is roughly 60 to 90 m/min. A representative production unit, the Sharp XF series, tops out at 60 m/min, and some laser-guided units run to about 1.4 m/s, equivalent to 5 km/h. Crucially, the rated top speed is not the operating speed everywhere: the vehicle automatically reduces speed near people, at intersections, and at load-transfer points, so throughput modeling must use realistic zone speeds rather than the headline maximum.

Positioning and stopping accuracy follows the guidance method. Magnetic and optical guidance achieve roughly plus-or-minus 10 mm at stop points, while laser-SLAM and laser-triangulation systems achieve about plus-or-minus 5 to 10 mm. Where docking or transfer must be tighter than the free-navigation tolerance, mechanical guides at the stand can constrain the final position. Specify the accuracy you need at the pick and drop, not just the vehicle's free-running accuracy.

Battery and charging parameters center on a 24 V or 48 V bus, with higher voltages on heavy units, and LiFePO4 chemistry increasingly standard. Runtime ranges from about 8 to 24 h depending on duty cycle. Charging is typically handled in one of three ways: auto-charge completed in two hours or less, opportunity charging during dwell, or a battery swap done in a few minutes. Drive configuration may be differential, tricycle, quad-steer, or omnidirectional Mecanum for tight-space maneuvering. Finally, gradeability, ground clearance, turning radius, and operating temperature are all model-specific and must be matched to the facility rather than assumed.

Behind the datasheet sit the governing standards that define what "safe" means for these vehicles. ISO 3691-4:2023, titled "Industrial trucks — Safety requirements and verification — Part 4: Driverless industrial trucks and their systems," is Edition 2, published June 2023; it replaced and withdrew the first edition ISO 3691-4:2020. As EN ISO 3691-4:2023 it is the harmonized reference for CE marking of AGVs and AMRs in the EU and has effectively superseded the former EN 1525. It requires formal hazard identification and risk assessment aligned with ISO 12100, and it mandates that safety-related control functions meet defined Performance Levels (PLr) per ISO 13849-1.

In the US and Canada, the reference is ANSI/ITSDF B56.5-2024, "Safety Standard for Driverless, Automatic Guided Industrial Vehicles and Automated Functions of Manned Industrial Vehicles," which replaced the 2019 edition. It covers design and construction, performance testing, installation and commissioning, automatic braking, speed control, collision avoidance, pathway design, warning devices, and emergency stops. It addresses control-system safety, but it does not use the ISO 13849 PLr framework that ISO 3691-4 mandates, which is a key difference when a project must satisfy both regions. Separately, VDA 5050 is a communication-interface standard from VDA and VDMA, not a safety standard, defining a universal master-control-to-vehicle interface so AGVs and AMRs from different makers can run in one fleet; it uses JSON messages over MQTT, and the current version, VDA 5050 3.0, was published by VDA/VDMA in March 2026, adding zones for traffic management and superseding version 2.1.

The safety functions an AGV typically must provide are concrete and checkable: speed-dependent protective fields from safety laser scanners, rated to an appropriate Performance Level (commonly PLd); emergency-stop devices; contact bumpers; audible warnings, often around 85 dB, plus visual warnings; automatic braking; overspeed and overload protection; and a safe stop on sensor fault or low voltage. Confirm each of these against the supplier's type-examination and conformity documents rather than the brochure.

Selection Decision Factors

To turn the preceding five chapters into a specific model, work through the decision sequence below. As with most capital equipment, selection mistakes rarely come from a single wrong step; they come from deciding the model before the load profile, navigation fit, and environment are settled. These eight factors can serve as a fixed RFQ template.

1. Load profile: Define weight, dimensions, and load type (pallet, tote, cart, or coil), then the load-transfer method (deck, roller, fork, lift, or tow). This single factor narrows the candidate vehicle types from Chapter 2 before anything else.
2. Navigation fit: Assess floor condition and your willingness to modify the floor (tape or wire) versus installing reflectors versus going infrastructure-free with SLAM. Weigh how often the layout changes, because frequent changes strongly favor SLAM, while a permanent route favors tape or wire.
3. Required accuracy: Specify the accuracy needed at pick, drop, and docking, distinguishing the plus-or-minus 5 to 10 mm class from coarser tolerances. Tighter docking may require mechanical guides beyond the vehicle's free-navigation accuracy.
4. Throughput and route: Count vehicles, traffic density, and intersections, and identify elevator and door interfaces. Evaluate fleet-management capability, and for mixed-vendor fleets confirm VDA 5050 support so vehicles from different makers can run under one master controller.
5. Environment: Settle temperature (freezer or cold-chain), cleanroom or ESD requirements, hazardous or ATEX areas, indoor versus outdoor operation, and floor slope. These constraints can disqualify otherwise-suitable vehicles outright.
6. Power strategy: Choose between opportunity/auto-charging and battery swap, select the battery chemistry (LFP for uptime), and match runtime to the shift pattern. The charging model must fit the natural dwell time in your flow, not the other way around.
7. Safety and compliance: Require ISO 3691-4 conformity for CE/EU or ANSI B56.5 for the US, confirm the safety-scanner Performance Level rating, and verify integration with existing safety infrastructure. Projects spanning both regions must satisfy both frameworks, including the ISO 13849 PLr requirement.
8. Integration: Confirm WMS, WCS, and ERP connectivity, MES line-feed timing where the AGV feeds an assembly line, and safe coexistence with human traffic. The software interface is often the deciding factor between two vehicles with similar mechanical specs.

One dimension that buyers consistently underweight is fleet software and serviceability. A fleet controller that cannot model your traffic, manage intersections, and schedule charging will throttle throughput no matter how capable the individual vehicles are; and a vehicle whose maker supports VDA 5050 protects you from being locked into a single supplier as the fleet grows. Among the cross-verified market leaders are Daifuku (Japan), JBT (US), KION Group / Dematic (Germany), Toyota Industries / Toyota Material Handling (Japan), KUKA (Germany), Oceaneering International (US), Hyster-Yale, E&K Automation, SSI Schaefer, Scott, and Seegrid; the top five of these (Daifuku, JBT, KION, Toyota Industries, KUKA) are frequently cited as collectively holding roughly 40 to 50 percent of the global AGV market. Brand reputation is a useful filter, but the eight factors above, not the logo, should decide the model.