AMR and AGV: not a contest.
Most plants end up running both — a guided vehicle on the lane that never moves, an autonomous one where the layout keeps changing. The decision is per route, not per site.
- AGV
- Follows a path installed on the floor
- AMR
- Builds a map, plans its own path
- Infrastructure
- Tape or QR markers vs no markers
- Route change
- Re-lay the floor vs redraw the map
- Typical fleet
- Both, on different routes
Two answers to “where am I?”.
AGV — automated guided vehicle
It follows a path that was physically installed for it: magnetic tape, QR markers on the floor, reflector strips on walls, or an inductive wire. Position comes from reading that path, so the path has to exist before the vehicle moves and stay intact afterwards.
AMR — autonomous mobile robot
It carries a map of the site and a laser scanner. It matches what it sees against that map to work out where it is, then plans a route through whatever is in the aisle. Nothing is installed on the floor; the map is the infrastructure.
Ten differences that change the quote.
| Criterion | AGV | AMR |
|---|---|---|
| Navigation | Magnetic tape, QR markers, reflectors or wire | SLAM against a stored map |
| Floor work at install | Markers along every route | Charger position and network coverage |
| Changing a route | Re-lay markers, re-teach, re-test | Redraw the route on the map |
| A person in the aisle | Stops and waits for it to clear | Re-plans around them |
| A pallet left in the lane | Blocked until someone moves it | Routes around it or reports the blockage |
| Repeatability | High, but tied to marker spacing | Around ±10 mm, map-dependent |
| Unit price | Lower | Higher |
| Cost over five years | Rises with every layout change | Flat after installation |
| Best fit | Stable layout, fixed lanes, high volume | Changing layout, mixed traffic, shared aisles |
| Typical failure | Forklift damages tape, pallet covers a QR code | Map drift after a major change, dirty scanner glass |
Navigation
- AGV
- Magnetic tape, QR markers, reflectors or wire
- AMR
- SLAM against a stored map
Floor work at install
- AGV
- Markers along every route
- AMR
- Charger position and network coverage
Changing a route
- AGV
- Re-lay markers, re-teach, re-test
- AMR
- Redraw the route on the map
A person in the aisle
- AGV
- Stops and waits for it to clear
- AMR
- Re-plans around them
A pallet left in the lane
- AGV
- Blocked until someone moves it
- AMR
- Routes around it or reports the blockage
Repeatability
- AGV
- High, but tied to marker spacing
- AMR
- Around ±10 mm, map-dependent
Unit price
- AGV
- Lower
- AMR
- Higher
Cost over five years
- AGV
- Rises with every layout change
- AMR
- Flat after installation
Best fit
- AGV
- Stable layout, fixed lanes, high volume
- AMR
- Changing layout, mixed traffic, shared aisles
Typical failure
- AGV
- Forklift damages tape, pallet covers a QR code
- AMR
- Map drift after a major change, dirty scanner glass
The rows describe two navigation approaches, not two competing products. Repeatability and price bands differ by model, and either approach can be specified on a different route in the same plant.
Two navigation types, one floor plan.
The question is rarely which technology to standardise on. It is which routes are worth changing later — and those are the routes that get the autonomous vehicle.
A route that has run unchanged for years
Raw material from the same dock to the same cell, all day. A guided vehicle does this cheaper, and there is nothing for autonomy to solve.
A route that follows the product mix
Changeovers, seasonal lines, temporary cells. This is where a map edit beats re-laying tape, and where the extra unit cost pays back.
Shared aisles with people and forklifts
Dynamic avoidance is a navigation property, not a dispatch one. A path-following vehicle stops; an autonomous one re-plans around the obstruction.
One dispatcher for both
A fleet manager assigns tasks across both types, arbitrates shared intersections and keeps one traffic rule set. Two technologies, one queue of work.
Mixing is a scheduling question before it is a technology question. The survey maps every route, then assigns each one a navigation type.
Four questions, asked route by route.
Ask these of one route at a time. A plant with six routes can legitimately answer four ways one way and two the other.
Will this route change in the next two years?
If yes, autonomy. Every later change becomes a map edit instead of a re-lay, and that is where the cost difference actually lives.
Is this route fixed and its cycle tight?
One lane running all day is where a guided vehicle still wins: simpler, cheaper per unit, and predictable. Keep it, and spend the autonomy budget on the routes that move.
Do people and forklifts share this aisle?
Shared aisles force dynamic avoidance. A path-following vehicle stops; an autonomous one goes round.
Can this floor be modified?
Sealed floors, cold stores and clean rooms often cannot take tape or inlays. That rules the guided option out before price is discussed.
What buyers ask about the comparison.
Answers describe the two navigation approaches. Which one suits a route is confirmed in the survey, with routes and cycle times measured rather than assumed.
01Are AMR and AGV alternatives to each other?
Not usually. They answer different questions. A guided vehicle answers "how do we move this same load along this same lane, cheaply and repeatably". An autonomous one answers "how do we keep moving when the lane, the load or the layout changes". Both answers can be right in the same building.
02Can AMR and AGV run in the same fleet?
Yes. A fleet manager dispatches both types, arbitrates shared intersections and applies one traffic rule set. In practice that is what most large sites do: keep the guided routes that already work, add autonomy where the layout moves.
03Do we have to replace working AGVs to add AMRs?
No. Existing guided routes can stay as they are. The usual sequence is to add autonomous vehicles on the routes that change most often, then decide later whether the fixed lanes are worth converting — often they are not.
04Is an AMR always better than an AGV?
No. On a fixed lane with a fixed cycle, a guided vehicle is cheaper, simpler and easier to justify. The advantage of an AMR is change, not speed.
05Does an AMR need floor marking?
No tape, QR codes or reflectors. The site work is a charging position, network coverage and a mapping run.
06Is ±10 mm precise enough?
For pallet and trolley handover, usually yes. For machine tending with tighter tolerances, add a fixed docking station or a vision assist.
07Is a QR-code robot an AGV or an AMR?
Guided, by navigation. The path lives on the floor, so the route is only as flexible as the labels.
08What happens when the layout changes?
The map is re-surveyed and the routes redrawn. No floor work and no re-teaching of every stop.
09Will it hit people?
Laser scanners slow and stop the vehicle, with bumpers as the last layer. The safety configuration is set per site and confirmed in the survey.
10How much more does an AMR cost?
The unit price is higher. The total depends on how many times the layout will change, so we quote both against your actual route plan rather than a list price.