HEALTH AT WORK

Workplace Safety

Warehouse Robot Collision Risk for Pickers Working Alongside Automated Guided Vehicles

14 October 2026 · Health at Work · AI-assisted draft

Photo: ROCCO STOPPOLONI / Pexels

The collision risk between warehouse pickers and automated guided vehicles is not a theoretical concern - it is a documented, growing cause of serious workplace injury. As fulfilment centres deploy autonomous mobile robots and AGVs at scale to meet rising order volumes, the workers who pick, pack, sort and replenish alongside that equipment are operating in an environment that was not designed with their movement in mind. Understanding exactly how these collisions happen, what injuries result, and what effective controls look like is now a core occupational health obligation for every employer running a mixed human-robot operation.

What the Collision Risk Actually Looks Like in a Working Warehouse

An automated guided vehicle typically weighs between 500 kilograms and several tonnes when loaded. It travels at speeds ranging from a slow walking pace to roughly 2 metres per second depending on the system and zone. A picker moving between aisles, stepping backwards from a shelf, crouching to retrieve a low-level item, or turning quickly to check a handheld scanner is not always visible to the vehicle in time for its sensors to stop it safely. The physics are simple and unforgiving: a loaded AGV at 1.5 metres per second cannot stop instantly, and a human body in the path of that mass will absorb the force that the braking system cannot dissipate.

The collision scenarios that appear most frequently in incident reports are not dramatic runaway-robot events. They are ordinary moments: a picker steps into a shared corridor from an aisle opening just as a vehicle rounds a corner; a worker kneels on the floor to retrieve a fallen item in a zone the AGV does not expect a human to occupy; a picker walks against the flow of AGV traffic in a narrow lane because it is the shorter route to their next pick location; or a vehicle resumes movement after a brief sensor-triggered stop and the worker who triggered the stop has moved into the vehicle's new travel path rather than away from it.

Autonomous mobile robots that navigate dynamically - meaning they plan their own routes rather than following fixed tracks - introduce a different risk profile from classic fixed-path AGVs. Their ability to reroute around obstacles is a safety benefit in open areas, but in dense pick-face environments where humans and robots share every square metre, their path changes can be unpredictable to workers who have learned to read the fixed-path vehicle's behaviour.

Who Is Most Exposed and in Which Tasks

Every worker on a mixed-floor operation is exposed to some degree, but exposure is not uniform. Order pickers working batch or zone picking in high-density racking are at greater risk than workers stationed at fixed pack benches, because pickers are in continuous motion and their routes intersect repeatedly with AGV travel lanes. Workers on induction and replenishment tasks who carry bulky items that restrict their forward sightline are at elevated risk because they cannot see AGV approach signals. New starters are consistently over-represented in near-miss data, because they have not yet developed the spatial awareness of the floor layout that experienced workers build over weeks.

Night shift and peak-period workers face compounding risk factors: fatigue shortens reaction time, AGV traffic density may be higher during surge periods, and communication between floor supervisors and the warehouse management system controlling the vehicles may be less responsive. Workers wearing hearing protection in noisy zones who also rely on audible AGV approach warnings lose a significant layer of protection without a compensating visual alert. Short-statured workers can be below the effective range of proximity sensors calibrated for average adult height.

Injuries: What the Early Signs and Late Consequences Look Like

Not every incident between a picker and an AGV is a full-speed collision. The injuries that go unreported most often are the low-energy contacts - a vehicle nudging a worker's heel or hip at slow speed, a worker stepping aside suddenly to avoid a vehicle and wrenching a knee or ankle, or a worker falling when startled by a vehicle appearing in their peripheral vision. These events produce soft tissue injuries that workers frequently dismiss as minor. Over time, workers in high-AGV environments report increased rates of lower limb musculoskeletal pain that may be partly attributable to repeated avoidance movements rather than to single collision events.

Higher-energy collisions produce blunt force trauma to the lower legs, pelvis, and torso, crush injuries where a worker is trapped between a loaded vehicle and a fixed structure such as a racking upright or a wall, and fall injuries where the worker is knocked off balance and strikes the floor or racking. Crush injuries to the foot and ankle from vehicle wheels are among the most severe outcomes and can result in fractures, nerve damage, and permanent disability. Psychological consequences following a serious near-miss or witnessed collision include acute stress reactions, hypervigilance, and in some cases post-traumatic presentations that affect a worker's ability to return to a floor environment comfortably.

What Employers Are Required to Do

Under typical occupational health and safety regulations in most jurisdictions, employers introducing AGVs into a working environment have a duty to carry out a suitable and sufficient risk assessment that specifically addresses the interaction between the automated system and any human workers sharing the same space. That assessment must not be carried out as a desk exercise using the AGV manufacturer's specifications alone; it must reflect the actual layout, the actual tasks workers perform, the actual routes workers use informally, and the actual volume and speed of AGV traffic at different points in the operation cycle.

The risk assessment must be reviewed when the system changes - when new AGV models are introduced, when the warehouse layout is modified, when picking strategies change and human movement patterns shift as a result, and when incident or near-miss data indicates that the existing controls are not sufficient. Employers are also required to provide adequate information, instruction and training to all workers who share floor space with AGVs, including workers whose primary task does not directly involve the vehicles - because it is often those workers who are caught in collision events precisely because they are focused on their own task rather than on vehicle awareness.

Controls That Work, and Controls That Do Not

The most effective controls are those that eliminate or substantially reduce the shared space between humans and moving AGVs. Physical segregation - barriers, pedestrian corridors, time-separated access zones where vehicles operate in one period and workers operate in another - addresses the root cause rather than relying on human behaviour or sensor reliability. Where full segregation is not operationally possible, partial segregation through clearly marked pedestrian lanes that AGV routing software is programmed to avoid is the next most effective option.

Slow zones controlled by pressure-sensitive floor mats, light curtains, or zone-based speed reduction triggered when a human presence is detected in a defined area are well-established engineering controls. These work best when the detection technology is calibrated for the actual working environment, regularly tested, and not routinely overridden by workers or supervisors trying to maintain throughput. Speed reduction to near-walking-pace in any zone where human-robot interaction is unavoidable substantially reduces the energy involved in any contact event.

High-visibility workwear and audible approach warnings from vehicles are useful supplements but they are weak controls when used alone. High-visibility clothing does not help if the vehicle sensor system is not capable of detecting it reliably; audible warnings are ineffective in noisy environments or for workers wearing hearing protection. Relying on these as primary controls is a common and dangerous mistake. Workers who are told that the AGVs will always see them develop false confidence in a system that has real sensor limitations under real conditions.

Training must go beyond induction. New starter training that covers the basic rules of the floor is not sufficient on its own. Workers need to understand why the rules exist, what the vehicle sensor systems can and cannot detect, how to respond when a vehicle stops unexpectedly rather than assuming it is safe to walk past, and how to report near-misses without fear of blame. Near-miss reporting is one of the most valuable sources of safety information in any AGV environment, and workplaces where it is actively suppressed by productivity pressure accumulate risk silently until a serious event occurs.

What Workers Should Do if They Are Already Concerned

A worker who has experienced a near-miss with an AGV, who has noticed vehicle behaviour that seems inconsistent or unpredictable, or who has sustained any injury however minor in the course of avoiding a vehicle should report the event through whatever incident reporting system the employer operates. Where that system does not exist or is not accessible, the concern should be raised directly with a supervisor or safety representative. A near-miss that is recorded becomes data that can be acted on; a near-miss that goes unrecorded becomes a future injury.

Workers who develop lower limb pain, psychological distress, or sleep disruption linked to stress about working alongside AGVs are experiencing genuine occupational health effects and should seek assessment from an occupational health professional rather than waiting to see whether the symptoms resolve. The earlier a musculoskeletal or psychological effect is identified and linked to workplace conditions, the more effective the intervention can be and the less likely it is to become a long-term condition.

What Employers Most Often Get Wrong

The most common error is treating the AGV manufacturer's safety certification as a substitute for an employer's own risk assessment. A vehicle that is certified as safe under controlled test conditions is being deployed into a specific building, with a specific workforce, carrying out specific tasks under specific time pressures - none of which the certification covers. The second most common error is assuming that because serious injuries have not occurred yet, the current controls are sufficient. In AGV environments, near-misses frequently precede serious events by months, and a floor where near-misses are not being reported is not necessarily a safe floor. Health at Work works with employers to build structured safety monitoring, near-miss investigation, and occupational health surveillance into AGV operations so that the risks of working alongside warehouse robots are genuinely managed rather than simply assumed away.