A plant should size its AMR fleet from material demand, cycle time, payload, and recovery capacity, because a line with frequent resupplying can require two delivery units plus one container position for every replenishment cycle [S1].
The calculation covers line-side material delivery for assembly operations, not finished-goods transport or warehouse picking, and it must be validated against route congestion, charging, empty-container returns, and the failure of any single robot.
Start with the resupply cycle, not a robot count
Line-side delivery is the interface between assembly operators and internal logistics, so the design must reconcile the production takt with the supplier's delivery frequency and the space available at the line edge [S1].
For a line whose resupplying interval is 10, 15, 20, 30, or 60 minutes, the stated capacity rule is 2U + 1 container, where U represents the material unit used for the resupply [S1]. In practical terms, the system needs two units in active delivery plus one container available to absorb a timing mismatch, rather than sizing the fleet from the number of assembly stations alone.
That rule should be translated into measured plant data: units consumed per interval, average round-trip time, load size, shift pattern, and the number of parallel delivery routes. A line receiving kit-based assemblies rather than thousands of individual parts also changes handling frequency, container dimensions, and the value of autonomous transport [S2].
Define the unit before counting vehicles
The container, not the robot, is the basic capacity unit in line-side design, because every component requires a defined package and every package must be accessible without creating a separate handling operation [S1].
A plant should record at least three physical variables for every route: the number of load units per delivery, the usable payload per AMR, and the time needed to move one unit from the source to the line. These values determine whether one robot can complete a round trip inside the available interval or whether additional vehicles are unavoidable.
Where standard bins or totes fit the material, the operator can access them within a stated ergonomic window of 700–1300 mm. The edge of the line should also remain below 1,300 mm so that operators can see the work area and supervisors can identify problems without entering a restricted zone [S1]. These dimensions affect replenishment speed because an inaccessible container adds picking and repositioning time to every cycle.
For palletized or fork-carried material, the AMR selection must follow the physical load rather than the packaging concept alone. Published line-side examples include an autonomous forklift rated for 3,000 lb with a 60-inch lift height and 6.7 mph travel speed, and a pallet-handling platform rated for 10,000 lb with a 4.5 mph speed [S3]. Those figures are model examples, not universal AMR requirements.
Compare AMR types against operating criteria
AMR, AGV, and manual cart delivery differ in routing flexibility, labor exposure, and fleet economics, so the preferred option depends on the operational constraint rather than on the automation label [S2].
The comparison must include more than purchase cost. Manual delivery can adapt to unusual material requests, while scheduled automation can lose effectiveness when a fixed route cannot be changed because of an obstruction or production change [S2]. For a fleet decision, compare at least four criteria: deliveries required per hour, load payload, round-trip time, and route recovery time after an interruption.
Autonomous guidance also removes the need for magnetic strips and pre-programmed routes in the workflow described by ManufacturingTomorrow, allowing a self-driving vehicle to move outside a strictly programmed path [S5]. This flexibility is useful only if the plant has a traffic plan, reserved loading positions, and rules for mixed pedestrian and robot movement.
Calculate fleet demand from measurable work
The minimum fleet is the number of robots required to cover all productive and recovery time during the planned production window, and it must also satisfy the line-side capacity rule at every replenishment point [S1].
Use the following calculation sequence:
1. Record units consumed per minute for each line or zone. 2. Record the average and peak units consumed during 10, 15, 20, 30, and 60-minute intervals where those intervals apply. 3. Convert consumption into the required delivery-unit capacity for each interval. 4. Subtract the line-side buffer implied by the 2U + 1 container rule. 5. Divide the remaining recurring transport demand by the safe payload and net cycle time of one AMR. 6. Add recovery capacity for charging, blocked routes, empty-container returns, and failed missions.
For example, if demand during one interval is 3U and the line-edge buffer provides 1U, the recurring delivery requirement is 2U. If one AMR can safely complete 2U per interval, one robot may satisfy the arithmetic, but it leaves no capacity for charging or disruption. The justified fleet would then be 2 AMRs if both units must remain productive during a mission or 1 only if a proven manual or automated contingency can cover the missing capacity.
This is a sizing method, not a claim that every plant must buy the theoretical minimum. Production-critical lines normally need a defined recovery margin, while low-risk side flows may share a pooled fleet with other routes.
Account for the empty-container return
A delivery fleet that ignores empty-container movement will appear sufficient on paper while failing in operation, because the vehicle must complete an outbound delivery and then recover the carrier before its next assignment [S1].
The return process should be treated as a separate flow, with a dedicated exit path so that empty containers do not block inbound material. Line-side design guidance calls for an empty-container return line, or “all down the drain,” and states that empty returns should not become saturated between two internal-logistics passages [S1].
Include the empty travel distance, transfer time, and staging requirement in the AMR cycle time. If a robot can carry a pallet one way but cannot handle its empty carrier, the practical fleet is larger than a payload-only calculation suggests. The line-side system must also prevent operators from bending to retrieve parts: the documented ergonomic window is 700–1300 mm, with large pieces placed at the bottom of the container [S1].
For a plant already using an automated storage and retrieval system, line-side AMRs can serve as the final distribution layer between storage and assembly. The AMR count should then reflect the ASRS output frequency, tote compatibility, and whether downstream delivery can be pooled with other internal traffic.
Test the plan against failure modes
A fleet sized only for normal travel time is not production-ready, because every AMR cycle can lose capacity to charging, congestion, blocked aisles, payload changes, and empty-carrier recovery [S8].
The validation plan should record at least six operating conditions: scheduled charging time, payload variation, route obstruction response, station waiting time, container exchange time, and failed-mission recovery. A useful stress case is a robot removed from service during the highest-consumption interval. The remaining fleet must still meet the 2U + 1 container requirement at the affected line [S1].
Manual delivery remains the fallback in some operations, but its performance depends on labor availability, safety constraints, training, and route variation [S2]. Automated fleets therefore need an operational escalation path, not merely a software dashboard. Site planning should define robot-only routes, shared aisles, crossing points, pedestrian procedures, and the location of charging or battery-change activity.
Real-time material visibility can support replenishment, but the operating target remains the right material, quantity, and location at the required time [S4]. For a conveyor sorting line, the comparable calculation is based on presenting sequenced totes at the pickup point; for an AMR fleet, the same principle becomes a question of how many load units each robot can present before the line consumes them.
Choose the business case by workload
The strongest business case appears when a plant has repeatable routes, recurring station demand, a measurable labor burden, and enough variation that fixed-path automation becomes difficult to redeploy [S2][S8].
A factory with one low-volume station and irregular kitting needs may be better served by a cart and a shared labor resource. A plant with multiple assembly lines, common replenishment intervals, and synchronized takt demand gains more from a dedicated or pooled AMR fleet because utilization and routing can be managed centrally.
Before approving the purchase, build a baseline from four measured quantities: 10-minute peak unit consumption, average AMR round-trip minutes, percentage of cycle time spent waiting or empty-returning, and labor hours per shift assigned to line-side movement. The installed fleet should then pass three checks: normal demand is covered, one-robot loss is survivable, and the 2U + 1 container buffer remains available at critical lines [S1].
Track cycle completion, station stockouts, empty-container saturation, charger occupancy, and mission failures after commissioning. If deliveries meet the planned interval but empty returns saturate, add fleet capacity or redesign the return route. If utilization stays low, consolidate routes before buying additional robots.
The trackable procurement output is not a generic “robots per line” factor; it is a documented fleet calculation with route-level demand, usable cycle time, contingency capacity, and a confirmed buffer for every critical line-side delivery point [S1].