Electronics-handling AGVs typically span 50–15,000 kg of configurable payload, with chassis footprints customisable from 600×800 mm unit-load carts up to 6000×2400 mm pallet trucks [S2]. For SMT-line feeders, PCBA totes, and semi wafer-carrier moves, the realistic working envelope is the lower half of that range (50–500 kg).
Selection logic differs sharply from bulk palletising: the bottleneck is static-discharge control, repeatability of pick-up at sub-10 mm tolerance, and shared-aisle coexistence with manual operators on the same material handling floor. Treat navigation, payload, and ESD as a single decision, not three sequential filters.
Payload Class and Chassis Footprint Mapping
Unit-load AGVs in the 50–1500 kg bracket dominate electronics lines because the dominant flow unit is a tote or a stack of magazines rather than a pallet [S2]. A 600×800 mm footprint fits one or two standard SMT magazines; 1200×1000 mm covers a four-magazine stack on a single roller deck.
Lightweight unmanned forklifts in the 1.5 t class (e.g. the LSXT15-D platform) use multi-line LiDAR 3D SLAM for natural navigation without magnetic tape or QR codes, and integrate with WMS/WCS via a multi-machine scheduler that re-plans routes between waypoints on each cycle [S3]. For finished-goods moves out of the SMT cell, that payload class is the practical ceiling; above 2 t, you cross into pallet-truck and counterbalanced forklift categories that the same LiDAR SLAM stack still covers but at larger turning radii.
Navigation: Magnetic Tape vs LiDAR SLAM vs QR/2D Code
Three navigation families compete in electronics plants. Magnetic-tape and QR/2D-code systems give deterministic ±10 mm accuracy on a fixed path, but every layout change requires re-laying tape or re-printing codes; multi-line LiDAR 3D SLAM systems build the map on first pass and only need incremental re-mapping when racks or conveyors move [S3].
For a 1.0–1.5 t AGV robot on a stable SMT line, the decision usually resolves to: QR/2D-code where floor space is premium and the route is permanent; LiDAR SLAM where the plant is rented, frequently re-balanced, or runs multiple product mixes. Repeatability of 3D-SLAM systems in dynamic warehouses is reported at centimetre-class for pallet-pickup, which is sufficient for tote handoff but borderline for direct PCB tray mating without a mechanical funnel.
ESD, Cleanroom, and Floor-Quality Constraints
Electronics AGVs must dissipate static charge through ESD-rated wheels (typical 10⁶–10⁹ Ω) and grounded brush contact at the chassis, otherwise a single stop near an exposed PCB can take out a batch. Conductive polyurethane wheels and a verified ground path are the two specs to lock into the RFQ; rubber-tyred stock units are not acceptable on ESD-protected floors. [S1]
Cleanroom compatibility drives a different filter: ISO Class 7 (10,000) lines tolerate most LiDAR-equipped AGVs with non-particle-shedding wheels, but Class 6 (1000) zones need sealed drive enclosures and HEPA-friendly motor cooling. Floor flatness matters more than people expect, since 3D-SLAM localiser accuracy degrades on deflected or polished concrete beyond a 2 mm/m slope variance [S3].
Safety Stack: LiDAR Zones, Bumpers, ISO 3691-4
AGV safety is now governed by ISO 3691-4, which mandates two independent safety channels, defined protective fields around the vehicle, and a category-3 PL d architecture for the motion-control safety function. The LSXT15-D class unit ships with a 3D protective LiDAR that establishes multi-level regional protection, plus mechanical buffer strips and a system-anomaly detection module as a third fallback [S3].
For shared aisles with line-side operators, specify a dual-LiDAR field set: a 270° warning zone at 3 m, a 180° slow-down zone at 1.5 m, and an emergency-stop zone at 0.5 m, each with independent channel monitoring. Standard mechanical e-stops must be reachable from all four corners, and audible/visual alarms should drive forward and reverse. Anything below this stack fails most plant safety audits on first pass.
Fleet Scheduling and WMS/WCS Integration
Beyond the vehicle, selection hinges on how the storage handling layer schedules traffic. A multi-machine scheduler integrated with the WMS, ERP, and WCS re-plans optimal routes between waypoints in real time, avoiding deadlocks at conveyor junctions and charger bays [S3]. This is the layer that turns a 5-vehicle pilot into a 50-vehicle plant without proportional traffic-engineering effort.
For electronics, traffic density is high but predictable: SMT line-side routes converge at three or four choke points. Specify the traffic manager with deadlock detection at minimum, and verify it handles battery hand-off (auto-dock charging at <60 V DC is the typical safe choice on an ESD floor). The same scheduler also feeds the articulated robot cells when PCBA totes hand off to AOI or robotic assembly.
Selection Criteria Matrix: Three Realistic Builds
For a sub-200 kg PCBA tote loop on a permanent SMT route, a QR/2D-code AGV with magnetic-tape fallback, 600×800 mm footprint, ESD wheels, and dual-channel safety LiDAR is the lowest-capex fit. For a 1.5 t finished-goods or magazine-stack flow on a frequently rebalanced warehouse line, the collaborative robot-adjacent LiDAR-SLAM lightweight forklift (LSXT15-D class) is the mainstream answer, with WMS/WCS scheduling as the critical dependency [S3]. For mixed payload under 15 t on a cleanroom line, a custom 6000×2400 mm chassis with the same multi-line LiDAR stack but sealed IP54 drive enclosures is the build that survives Class 7 audits [S2].
For a deeper dive into regulated-industry AGV selection, the GxP-focused checklist in AGV selection for pharmaceutical distribution covers payload-ROI and audit gates that overlap directly with electronics plant validation.
Limitations and Failure Modes
3D-SLAM performance degrades in long, featureless corridors, in front of mirrored stainless-steel equipment, and under direct sunlight bleeding into dock doors; the workarounds are reflective fiducials and dock-door light curtains, not algorithm changes. Magnetic-tape and QR systems fail differently: tape lifts under heavy forklift crossings, QR decals fade within 12–18 months under UV from ceiling LED panels. [S3]
ESD failures are usually a wheel spec issue, not a vehicle one. A AMR robot on conductive casters can still carry 100 V of triboelectric charge if the ground brush is missing or oxidised. Specify the wheel and brush as a single line item and audit it quarterly. Final-stage validation: walk every planned route once with a manned tow tractor at full payload, mark any floor deviation >3 mm/m, and resolve those spots before the AGV first run.
Trackable signal: the next revision of ISO 3691-4 guidance on mixed-traffic zones, and any 2026 plant-floor deployments that switch from QR-coded to 3D-SLAM navigation in rented or multi-tenant electronics facilities.