Autonomous mobile robot

Also known as: AMR

On this page 7

An autonomous mobile robot (AMR) is a wheeled or legged robot that moves through its environment and plans its own path using onboard sensors, maps and computing, without physical guides such as floor wires, magnetic tape or reflectors. In robotics fundamentals, the AMR is the main example of a mobile robot that combines perception, localization, planning and control in one machine.

AMRs are most common in warehouses, factories, hospitals and retail stores, where they carry goods, tow carts or inspect facilities. Their defining capability is dynamic rerouting: when a person or pallet blocks the aisle, an AMR calculates a new path instead of stopping and waiting.

How an autonomous mobile robot navigates

An AMR runs a continuous sense-plan-act loop. Sensors measure the surroundings, a localization module estimates the robot’s pose on a map, a planner chooses a route to the goal and a controller converts that route into wheel speeds. The loop typically repeats 10 to 50 times per second.

Most AMRs build their map with SLAM (simultaneous localization and mapping) during commissioning, by driving the robot through the site once. In daily operation they localize against that map by matching 2D LiDAR scans or camera features, often fused with wheel encoders and an IMU. Planning is split into a global planner, which finds a route across the map, and a local planner, which avoids moving obstacles within a few meters.

SubsystemTypical componentsFunction
Perception2D or 3D LiDAR, depth cameras, ultrasonic sensorsDetect obstacles and features
LocalizationScan matching, encoders, IMUEstimate position and heading
PlanningGlobal and local plannersChoose and adjust the route
Control and driveMotor drivers, differential or omnidirectional baseExecute motion smoothly
SafetySafety-rated scanners, bumpers, emergency stopsSlow or stop near people
Fleet interfaceWi-Fi, fleet manager softwareReceive tasks and share traffic
Main subsystems of an AMR

AMR compared with AGV and other mobile robots

The AMR is often confused with the automated guided vehicle (AGV). An AGV follows predefined paths marked by tape, wires, magnets or reflectors and stops when an obstacle blocks the way. An AMR navigates freely within a mapped area and replans around obstacles. The distinction is about navigation autonomy, not about size or payload.

Robot typeNavigationResponse to obstaclesTypical use
AMRMap-based, free pathsReplans around themFlexible intralogistics
AGVFixed guides or routesStops and waitsRepetitive heavy transport
Service robotOften AMR-basedReplans, interacts with peopleCleaning, delivery, reception
Quadruped robotLegged, map and terrain basedSteps over or aroundInspection on rough terrain
DroneThree-dimensional flightAvoids in the airInspection and inventory scans
Comparison of AMR, AGV and related mobile robots

Many vendors now sell hybrid machines: forklifts and tuggers that use natural-feature navigation like an AMR but restrict themselves to defined lanes like an AGV, because fixed lanes simplify traffic rules and safety assessments.

Key parameters when specifying an AMR

Specifications vary greatly by application. Small delivery units carry tens of kilograms, while heavy-duty platforms carry one tonne or more. Speed in shared spaces is deliberately modest, commonly 1 to 2 meters per second, and is reduced automatically near people.

ParameterTypical rangeWhat it affects
PayloadAbout 50 kg to over 1,000 kgWhich loads and carts it can move
Maximum speedAbout 1 to 2 m/sThroughput and safety zone size
Positioning accuracyAbout ±10 mm to ±50 mmDocking at stations and racks
Runtime per chargeOften 8 to 16 hoursFleet size and charging plan
Minimum aisle widthRobot width plus safety marginWhere the robot can operate
Floor toleranceSmall steps and gaps onlySuitability of the building
Typical AMR parameters and what they affect

Safety, standards and fleet operation

Because AMRs share space with people, functional safety is central. Safety-rated laser scanners define protective and warning fields that grow with speed, and the robot slows or stops when a field is violated. Industrial AMRs are designed against standards such as ISO 3691-4 for driverless industrial trucks and the North American ANSI/RIA R15.08 series for industrial mobile robots.

Real deployments use fleets, coordinated by a fleet manager that assigns tasks, manages traffic at intersections, schedules charging and integrates with warehouse or manufacturing systems. Fleet performance usually depends more on traffic management and task allocation than on the speed of individual robots.

Mixed fleets from several vendors raise an interoperability problem, because each vendor traditionally ships its own fleet manager. Interface standards such as VDA 5050 and the MassRobotics interoperability standard define common messages for orders, states and traffic, so one master control system can coordinate robots from different manufacturers. Support for such interfaces is worth checking early in procurement, since retrofitting integration later is costly.

Limitations and deployment pitfalls

  • Changing environments: frequently moved pallets or racks can degrade localization unless the map is updated.
  • Reflective and transparent surfaces: glass walls, mirrors and shiny wrap confuse LiDAR and cameras.
  • Wireless coverage gaps: robots lose tasks or traffic coordination in areas with poor Wi-Fi.
  • Underestimated congestion: too many robots in narrow aisles block each other and reduce throughput.
  • Floor conditions: ramps, thresholds and wet floors limit traction and positioning accuracy.

A site survey and simulation before purchase, together with a pilot zone, reveal most of these issues while changes are still cheap.

A plain-language summary

An autonomous mobile robot is like a delivery cart that knows the building. It carries a map in its memory, uses sensors to see what is around it and finds its own way from one place to another.

Older factory vehicles followed a line painted or taped on the floor, like a train on rails. An AMR is more like a careful driver: if a person or a box is in the way, it slows down, steers around and continues. That flexibility is why companies can add AMRs without rebuilding their floors.

An example from practice

A hospital moves medication, lab samples and linen between departments on several floors. Porters spent much of their shifts walking the same routes, and urgent samples sometimes waited for a free person. Managers wanted transport that runs around the clock without adding staff on night shifts.

The hospital deploys a small fleet of AMRs with lockable compartments. The robots are mapped during a quiet weekend, use elevators through an integration with the building system and drive at reduced speed in patient corridors. A fleet manager prioritizes laboratory runs over linen deliveries. Porters now handle tasks that need people, and sample delivery times become predictable, with every trip logged for traceability.