US The difference between an AGV and an AMR
An AGV, or automated guided vehicle, and an AMR, or autonomous mobile robot, both move materials without a driver. The difference is how they find their way. An AGV follows guidance installed in or on the building: a wire in the floor, magnetic tape, reflectors on the walls or codes on the floor. It goes where the guidance goes and, if something blocks the route, it stops and waits.
An AMR navigates from what its own sensors see. It builds a map of the site with LiDAR and cameras, works out where it is on that map, and plans a route to each job. If a pallet or a person is in the way, it re-plans and goes around. Both kinds move pallets, carts, totes and parts, and both are managed by fleet software that assigns jobs and manages traffic.
Neither is better in every case. AGVs are predictable and well suited to repetitive routes in stable layouts. AMRs are more flexible where layouts, traffic and tasks change.
AGV vs AMR side by side
The table sets out how the two compare on the points that usually decide a project.
| Feature | AGV | AMR |
|---|---|---|
| Navigation | Wire, magnetic tape, reflectors or floor codes | SLAM: LiDAR and cameras against a site map |
| Route | Fixed, defined by the guidance | Planned per job, changes as needed |
| Obstacles | Detects them and stops | Detects them and drives around |
| Layout changes | New guidance and reprogramming | Map and route update in software |
| Installation | Floor or wall infrastructure needed | Little or none beyond mapping |
| Loads | Light to very heavy, including towing and lifting | Mostly light to medium, totes, carts and pallets |
| Onboard compute | Modest: guidance and vehicle control | Higher: mapping, localization, perception |
| Safety standards | ISO 3691-4, ANSI/ITSDF B56.5 | ISO 3691-4, ANSI/RIA R15.08 |
The line between the two has blurred. Many current AGVs use natural feature navigation, matching LiDAR scans to the building instead of following tape, while still driving fixed virtual paths. Some AMRs can be told to follow set lanes in busy areas. So it is worth asking how a vehicle navigates and how it reacts to an obstacle, rather than relying on the label.
When to choose an AGV or an AMR
The right choice depends on the route, the load and how often the site changes.
-
//
Choose an AGV for fixed, heavy or high-volume flows. Feeding a production line, moving coils or towing trains of trolleys along the same path all shift suits guided vehicles, which are very repeatable.
-
//
Choose an AMR for changing layouts and mixed traffic. Warehouses, hospitals and e-commerce sites where routes, stock locations and pedestrians change through the day suit robots that plan their own path.
-
//
Consider people on the floor. Where staff and vehicles share aisles, a robot that slows and steers around them causes fewer stoppages than one that halts at every interruption.
-
//
Consider the infrastructure. Tape and wire wear and need repair, and reflectors need a clear line of sight. AMRs avoid most of that but depend on a good map.
How AGVs and AMRs handle safety
Both kinds of vehicle carry safety-rated laser scanners that watch protective fields around them and slow or stop the vehicle when something enters a field, plus bumpers and emergency stop buttons. ISO 3691-4 sets the safety requirements for driverless industrial trucks in Europe and is widely used elsewhere, and ANSI/RIA R15.08 covers industrial mobile robots in North America. The difference is what happens next: a guided vehicle waits until the path is clear, while an AMR plans a new path. On a busy floor that difference decides how many interruptions the operation sees in a shift.
The onboard computer in AGVs and AMRs
Either way, the onboard computer decides how well the vehicle sees and reacts to what is around it. An AGV’s controller follows the guidance, drives the motors and reads the safety scanners. An AMR’s computer also runs SLAM, localization and path planning, which needs more processing.
Real-time AI on the vehicle closes much of the gap between the two. Even fixed-route vehicles benefit from recognizing pedestrians, forklifts and blocked aisles, and from telling a person from a cardboard box. A pedestrian detection system on the vehicle can slow it down before the safety scanner has to stop it. LiDAR and cameras feed the same computer, which runs both the navigation and the AI models.
Our smart factory case study shows rugged fanless Jetson computers running that inference on material handling vehicles. Our rugged Jetson computers take battery power directly, accept GMSL and PoE cameras and talk to motor controllers over CAN, and our wider range of edge AI computers covers larger GPU systems for fleet servers.
Quick answers on AGV vs AMR
AGVs and AMRs: common questions
Is an AMR better than an AGV?
Not always. An AMR is more flexible, but an AGV can be the better fit for heavy loads, fixed routes and processes where every vehicle must follow exactly the same path.
Is an AGV a robot?
Broadly, yes. It is a mobile machine that works without a driver, but it follows external guidance rather than deciding its own route, which is why the industry uses a separate name for AMRs.
What does SLAM mean for an AMR?
Simultaneous localization and mapping. The robot builds a map from its sensors and works out where it is on that map at the same time, so it can navigate without floor markers.
Can AGVs and AMRs work in the same building?
Yes. Many sites run both, with a fleet manager or warehouse system coordinating their jobs and traffic.
Which is easier to install?
Usually the AMR. It needs a site map and some configuration, while a traditional AGV needs guidance laid in or on the floor along every route.
We supply rugged Jetson and fanless computers for AGV and AMR builders as standard, with the cameras, CAN interfaces and operating system image configured and validated before the system ships.