Why AMRs Fail in Real Factories: Oil, Skewed Pallets, Pedestrians, Layout Changes, and Mixed Traffic

AMR performance at a supplier’s demo center — smooth running, precise docking, efficient obstacle avoidance — is the result of ideal conditions. Flat floor, adequate lighting, no personnel interference, fixed material formats, single vehicle type. These conditions are almost impossible to fully replicate at a production site.

Commissioning engineers have noted that many manufacturing companies think deployment completion means they can relax — but on-site commissioning takes a huge amount of work. Another practitioner observed that logistics equipment cannot be judged by paper specs alone; whether it can adapt to the messy, ever-changing real workshop conditions is the real measure.

This article breaks down the lab-vs-real-factory differences into verifiable problems. Cases mentioned are illustrative observations from industry practitioners and buyer feedback — not industry statistics and not tied to specific factories or brands.

Oil, Debris, and Floor Changes

Lab conditions: Clean, flat epoxy floor, stable friction coefficient.

Real factory:

Buyers have reported “navigation unreliable, no positioning sustained” messages when floor texture degrades. Others report that floor codes break often, resin coating gets dirty, and PET film gets crushed by forklift traffic — floor marker maintenance is a continuous cost.

Mitigation: Floor condition assessment is a pre-deployment must. Oily areas need speed reduction, regular cleaning, or rerouting. Floor markers need a maintenance plan and backup options.

Skewed, Damaged, and Non-Standard Pallets

Lab conditions: Standard pallets, intact, precisely placed at marked positions.

Real factory:

Commissioning records list pallet recognition errors — 3D vision algorithm recognition rate drops under real lighting. Lab-perfect lighting and pallet posture do not exist in real workshops.

Mitigation: Pallet standardization is a deployment prerequisite. If standardization is not possible (customer-supplied material), the system needs to handle non-standard cases — speed reduction, manual confirmation, or diversion.

Personnel, Forklifts, and Temporary Obstacles

Lab conditions: No personnel interference, no other vehicles, no temporary obstacles.

Real factory:

Commissioning records list frequent personnel crossing as a core lab-vs-real difference.

Mitigation: Person-vehicle separation is ideal but not always feasible. Safety zones, speed limits, and personnel detection are needed. Temporary obstacles need an operations SOP — not “the robot handles it” but “who is responsible for clearing.”

Shift Changes, Line Changeovers, and Layout Drift

Lab conditions: Fixed layout, single product flow, stable takt.

Real factory:

Commissioning records list temporary line changeover as a lab-vs-real difference. Lab-stable production lines are dynamic in reality.

Mitigation: Maps and tasks need quick update capability. Layout changes need a change management process — not “change first, deal with it later” but “assess AMR impact before changing.”

Mixed Robot Types and Intersection Conflicts

Lab conditions: Single vehicle type, no multi-vehicle conflicts.

Real factory:

Commissioning records note multi-vehicle intersection path conflicts (deadlock during multi-type mixed running) as a core commissioning issue. Mixed fleets increase integration and traffic-authority complexity; achievable performance depends on the fleet-control architecture, interoperability support, traffic rules, and commissioning quality.

Mitigation: Mixed vehicle types need a clearly defined traffic authority or coordination architecture. If different types use different dispatching systems, intersection conflicts cannot be effectively managed. Standards for multi-brand interoperability exist (such as VDA 5050), but standards do not replace on-site commissioning. If mixed running is necessary, thoroughly test the dispatching system’s multi-vehicle management capability during the pilot phase, especially intersection traffic rules and deadlock detection/recovery.

Docking Offset and Local Calibration

Lab conditions: Docking positions precisely installed, markers intact, floor flat.

Real factory:

Commissioning records list assembly line docking positioning offset as a commissioning-stage issue.

Mitigation: Docking positions need periodic inspection and calibration. Markers need a maintenance plan. Docking accuracy needs to be verified with actual payload during SAT.

What to Include in Pilot, Commissioning, and Exception Logs

Lab vs. Factory Risk Matrix

Commissioning Exception Log Template

Risk Mitigation Inputs

1. Deployment complete does not equal project complete — on-site commissioning is a major effort, not “go-live and done.”
2. Oil, debris, and floor wear are persistent challenges — floor assessment and maintenance are long-term costs.
3. Pallet standardization is a deployment prerequisite — non-standard pallets need diversion handling.
4. Mixed personnel traffic needs safety design — not “the robot handles it” but “who is responsible for separation.”
5. Layout changes are the norm — maps and tasks need quick update capability; changes need a management process.
6. Mixed fleets increase integration complexity — achievable performance depends on architecture, interoperability, traffic rules, and commissioning.
7. Docking offset and marker wear are ongoing maintenance — not one-time installation.
8. Exception logs are the foundation for improvement — no recording means no root cause analysis.