SLAM-based AMRs no longer need reflectors, magnetic tape, or QR codes on the floor: reflector-free laser SLAM units use only onboard sensors to map and navigate the warehouse [S7]. That single fact reframes the procurement question, because the three legacy guidance families now compete on infrastructure cost and route-change frequency, not on raw navigation capability.
Magnetic tape and inductive wire still dominate brownfield AGV retrofits because they are deterministic, low sensor cost, and indifferent to lighting and dust [S1][S5]. Laser triangulation with wall reflectors is the legacy free-path compromise, and natural-feature SLAM (Lidar or visual) is the method the major AMR vendors ship as standard on new fleets [S2][S9]. The decision is really about how often your routes change and how dirty the floor gets.
How each navigation family actually works on a moving vehicle
Magnetic tape and inductive wire are fixed-path methods: the vehicle senses a magnetic field from tape glued to the surface or a wire embedded in a floor slot, then steers to stay centred. RFID tags or magnetic markers give periodic position updates along the loop [S1][S5]. The path is physically fixed, which is the main reason these systems are still chosen for safety-critical loops where the vehicle must never leave a defined corridor.
Laser triangulation with reflectors mounts retro-reflective targets on walls or columns at known coordinates; an onboard 2D laser scanner triangulates against them to localise the vehicle. Natural-feature SLAM drops the reflectors: the Lidar or stereo camera builds a map of the existing environment and localises against that map, updating it continuously as the scene changes [S1][S3][S7]. The mechanical difference is whether you install anything on the floor or walls; the algorithmic difference is whether the map is curated or auto-generated.
Decision matrix: infrastructure cost, flexibility, accuracy, environmental tolerance
Across the four main options the trade-off sits on the same axis: less installed infrastructure means more software complexity and more sensitivity to the environment. Magnetic tape and inductive wire carry the lowest sensor cost and tolerate dust, condensation, and lighting variation, but every route change means cutting floor or relaying tape, branching is limited, and throughput is capped by the fixed network [S1][S5]. Laser triangulation sits in the middle: it needs reflectors on the walls but no floor work, and accuracy in clean industrial settings is well documented.
Natural-feature Lidar SLAM is the most flexible: floor-based routes are kept in software rather than the building, layout changes are config edits, and modern systems can achieve accuracy levels below 10 mm in steady environments [S1][S3]. The cost shows up in higher upfront vehicle price, more demanding commissioning, and degradation in dusty or highly reflective spaces (glass, polished metal, fog). Visual SLAM extends the same free-path idea using cameras and AI-based feature extraction, which works in cluttered human spaces but is more sensitive to lighting changes than Lidar [S9].
Hybrid systems are now a deliberate procurement category, not a workaround. A robot might use Lidar SLAM for general navigation and switch to a floor-based anchor for high-precision docking or charging, and vendors are explicitly marketing this split [S1]. One recent commercial example pairs a Lidar-plus-camera "Hybrid SLAM" stack with a proprietary UV-fluorescent reflector tape that can be applied to any surface: the tape only fluoresces under UV, so the system still works like a free-path SLAM unit day-to-day, but uses the tape as a precision anchor when docking or crossing contaminated floor zones [S4].
Where each method is the right call, and where it fails
Magnetic tape and inductive wire are the right call when routes are permanent, the floor is concrete, dust or condensation is present, and you need deterministic safety validation with the simplest possible sensor stack. They are the wrong call when you expect more than a handful of route changes per year or when branching is needed, because the cost of re-cutting the floor or relaying tape quickly dwarfs the original install cost [S5][S8]. A useful sanity check from integrators: if a route changed more than twice in the last 24 months, fixed-path guidance is almost always the wrong choice.
Reflector-based laser triangulation is the right call when walls are clean, the layout is stable, and you want a free-path system without the commissioning overhead of SLAM. It is the wrong call in dusty plants, in spaces with lots of glass, or in warehouses where racking gets moved often, because reflectors have to be re-surveyed and re-mapped each time [S1][S3]. Reflector-free laser SLAM is the right call for most greenfield warehouse and shop-floor AMRs: no floor anchors, no wall reflectors, route changes are software edits, and the vehicle can detect and avoid obstacles that fixed-path AGVs would simply stop in front of [S2][S7].
Visual SLAM is the right call when lighting is controlled, the environment is feature-rich (cluttered retail backrooms, hospital corridors), and you want a camera-only payload. It is the wrong call in outdoor or rapidly changing lighting, on long featureless corridors, or wherever you cannot accept the higher compute footprint [S9]. The same accuracy caveats apply as for Lidar SLAM: dust on the lens, water on the floor, or strong backlight will degrade performance, and a hybrid anchor (magnetic tape, reflector tape, or QR markers) is often used to recover precision at docking stations [S1][S4][S6].
Standards, maintenance reality, and what to ask the vendor
There is no single ISO or IEC standard that picks a navigation family for you; safety validation is driven by the application standard (for example ISO 3691-4 for driverless industrial trucks) and the risk assessment on your site. What you should demand from any vendor is a written accuracy figure with the test conditions, a documented re-mapping procedure after layout changes, and a maintenance plan for the guidance infrastructure (tape replacement, reflector cleaning, lens cleaning) [S1][S5].
Maintenance reality is where the procurement case is won or lost. Fixed-path tape needs periodic re-gluing and replacement in high-traffic aisles, typically measured in years but accelerating in fork-truck crossings. Reflectors on walls need cleaning and re-surveying after any structural change. Lidar optics need compressed-air cleaning in dusty plants and will fail in fog or heavy steam. Visual SLAM cameras need lens cleaning and lighting checks. Budget maintenance hours per vehicle per month, not just the unit price, when you compare quotes [S1][S5].
Selection criteria, shortlisting, and a procurement checklist
A defensible shortlist starts with five questions: how often does each route change per year, what is the worst-case dust or condensation level on the floor, what docking accuracy do you need at pick and drop stations, what is the acceptable re-commissioning time after a layout change, and what is the recurring maintenance budget per vehicle per year. Map these answers against the matrix above and only then call vendors in [S5][S8].
For the typical industrial plant with monthly route tweaks and dirty floors, the modern default is reflector-free laser SLAM with a magnetic-tape or reflector-tape anchor at charging and high-precision pallet stations, a combination that several vendors now ship as a stock configuration [S1][S4]. For a fixed-loop pallet line that has not changed in five years, inductive wire or magnetic tape remains the cheapest, most robust choice, and traditional AGVs on these systems still run at higher vehicle speeds than AMRs on free-path networks, so throughput is not a penalty [S2][S5]. Treat any vendor claim of "universal best" navigation with the same scepticism an experienced integrator would: there is no best method, only the best fit for your routes, your floor, and your maintenance budget [S3][S8].
Trackable signals to watch over the next two quarters: more AMR OEMs releasing hybrid SLAM-plus-floor-anchor stacks as standard (Thira's UV-fluorescent tape is one early example) [S4], visual SLAM moving into 3D feature mapping for cluttered human spaces [S9], and magnetic-tape vendors pushing back with thinner, more durable tape and quicker re-laying workflows aimed at brownfield retrofits [S5][S8].
For component-level specifications, see amr robot, warning tape, and magnetic material.
For related coverage, see Capacity Reservation vs Take-or-Pay: Spec-Level Differences for Industrial Off-Take.