In AGV system design, wheels are often treated as commodity parts—purchased late, specified by load rating alone, and rarely revisited until the vehicle starts drifting, missing docking positions, or showing uneven tread wear. In reality, the wheel is the physical convergence point of drive kinematics, navigation accuracy, and floor conditions. No control algorithm can compensate for a wheel that was mismatched from the start.
The Drive Architecture Dictates Wheel Loading
The first rule of wheel selection is not to read the load chart—it is to understand the drive configuration. The same polyurethane tread compound behaves very differently under a differential drive versus a steering drive.
In a differential-drive AGV, the caster wheels (or load wheels) experience lateral forces as they follow the chassis through turns. The tread must balance low rolling resistance against resistance to lateral tearing. In a steering-drive system, the drive wheel must handle traction, vertical load, and the reaction torque transmitted through the steering mechanism—plus lateral friction during steering actuation. A compound optimized for differential casters may develop tread delamination or excessive heat buildup when deployed on a steering wheel.
Mecanum wheels present a more complex case. The angled rollers simultaneously carry radial compression and axial force components, demanding high elastic recovery from the roller material. Substituting a standard polyurethane for a purpose-engineered roller compound typically leads to permanent deformation of the rollers under load.
Why Polyurethane Dominates AGV Tread Materials
Polyurethane has become the default tread material for AGV wheels for good reason. Compared with rubber, it exhibits lower compression set under sustained load. Compared with nylon, it provides the friction coefficient needed for traction while protecting the floor surface.
But “polyurethane” is a broad category. The difference between 75A and 95A durometer corresponds to entirely different application envelopes. Harder compounds (around 93A) prioritize abrasion resistance and dimensional stability under high loads—suitable for drive wheels where traction is the primary requirement. Softer compounds (around 75A) emphasize floor protection and low rolling resistance, making them better suited for cleanroom or noise-sensitive environments. A common specification error is choosing the hardest available compound for maximum life, only to induce slip on smooth epoxy floors—triggering frequent navigation corrections that accelerate wear on the very wheels the hard compound was meant to protect.
Wheel core material matters as well. Under high-frequency start-stop duty, plastic cores can expand with heat and loosen bearing fits. Metal cores—steel or aluminum—are the more robust choice. Bearing configuration directly affects startup torque: ball bearings offer significantly lower breakaway resistance than plain bearings, a difference that manifests in energy consumption over thousands of micro-adjustments per shift.
Load Calculation: The Static Average Is a Trap
Most AGV projects size wheels using static load division: a 4,000 kg four-wheel AGV gets 1,000 kg per wheel. This ignores load transfer during acceleration, braking, and cornering.
Consider a representative calculation: total mass 4,000 kg, CG height 0.5 m, track width 0.8 m, braking deceleration 0.5 m/s². The load transfer is approximately 1,250 N—about 127 kgf shifting from the inner wheels to the outer wheels. If the wheel was specified with only a 10% margin over the static value, the outer wheel may exceed its rated capacity during an emergency stop. A more defensible approach incorporates dynamic transfer into the calculation and, for four-wheel configurations, uses one-third of total load per wheel rather than one-quarter to account for the possibility of one wheel losing contact on uneven floors.
Specifying Wheels from the Application Backward
Different AGV form factors impose distinct wheel requirements. Low-profile latent AGVs demand minimal chassis height, constraining drive wheel diameters to the 100–150 mm range while still requiring substantial load capacity—a combination that tests both tread thickness and core design. Forklift AGVs sit at the opposite end: larger diameters (200–300 mm) for obstacle negotiation and abrasion resistance, typically with polyurethane hardness above 92A.
For omnidirectional platforms, Mecanum roller consistency and hub machining accuracy are the critical variables. The elastic recovery of the roller surface governs lateral motion efficiency, while the angular precision of the roller mounting directly affects kinematic coupling errors. Control algorithms matter, but they can only work with rollers that are manufactured to consistent tolerances.
AGV wheel selection is fundamentally a matching problem—between drive architecture, floor conditions, and load spectrum. There is no universally “best” wheel, only a wheel that fits the system. When an AGV starts exhibiting positioning anomalies, it is worth checking the wheels before retuning the navigation parameters: Is the tread wear even? Do the bearings still turn freely? More often than not, the problem lives in the component closest to the floor.