AS-RS Systems

Automated Storage and Retrieval Systems (AS/RS) and Goods-to-Person (G2P) represent the high end of warehouse automation investment — and the highest potential return per square foot. They require the cleanest data, the most capital, and the most capable implementation teams. Year-1 performance runs 75-85% of design case; systems that hit design case at go-live are the exception.

---

AS/RS Technology Landscape

Unit-Load AS/RS (Pallet-Level)

Stacker cranes on floor and ceiling rails store and retrieve full pallet loads in dedicated aisles.

Performance:

• 60-120+ pallet moves/hour per crane (combined in/out)
• Dual-command cycle: crane deposits a pallet and retrieves another in one trip
• 99%+ availability (well-maintained); 85-95% (aging or poorly maintained)
• Very strict pallet quality requirement — damaged or overhanging loads cause faults

When to use:

• Land-constrained sites where vertical density justifies the cost
• 24/7 automated throughput with minimal labor
• Cold/frozen storage (labor in -20°F makes economics compelling)
• Pharmaceutical, chemicals, or regulated environments requiring traceability
• High inventory cube (2,000+ pallet positions) with relatively low SKU count

Cost benchmark: $3,000-10,000 per pallet position installed. A 5,000-position system: $25-50 million installed. System life: 20-25 years structural; cranes refurbished at 10-15 years; controls replaced at 15-20 years.

---

Mini-Load AS/RS (Tote/Carton-Level)

Same crane principle as unit-load but scaled for totes, cartons, and trays (2-80 lbs).

Performance:

• 150-500 tote moves/hour per crane
• Tote specs: typically 400-800mm L × 300-600mm W × 150-350mm H

When to use:

• High-velocity each-pick operations requiring dense, accurate storage
• Defined SKU catalogs with stable item dimensions
• Pharmaceutical, cosmetics, high-value goods requiring FIFO and lot control

Cost benchmark: $800-1,500 per tote position. A 10,000-tote system: $10-20 million.

---

Shuttle Systems

Shuttle vehicles run on rails within rack structure, moving totes/pallets within levels. Multiple shuttles operate in parallel.

Performance:

• Up to 1,000+ tote moves/hour for large multi-level systems
• Shuttles redistributable across levels to match throughput demand
• Scalable: add shuttles without adding rack structure

Cost: Higher per position than mini-load, but throughput advantage at scale justifies for very high-speed operations.

---

Cube Storage (AutoStore-Type)

Bins stacked directly on top of each other in a dense grid. Robots on top of the grid retrieve bins by digging down through the stack.

Key vendors: AutoStore (market leader), Ocado, Swisslog CarryPick, KNAPP Open Shuttle.

Performance:

• Storage density: 4-6× denser than conventional shelving; 50-60% footprint reduction
• Robot throughput: ~200-300 bin movements/hour per robot
• System scales with robot count; large systems: 1,000-5,000+ deliveries/hour
• Bins stacked 5-16 high; taller stacks increase density but slow retrieval

FIFO limitation: Not inherently FIFO — bins are buried. Requires software management.

AutoStore economics work when:

• E-commerce processing >5,000-10,000 orders/day
• Very space-constrained; land/lease cost justifies premium
• High SKU count with small, lightweight items
• Labor turnover and cost are primary business drivers

What AutoStore doesn’t handle: Large/heavy items (bins limited to ~30-50 lbs), bulk quantities of fast-moving items, very high peak-to-average ratios (robot fleet is fixed).

Cost benchmark: $5-15 million for 5,000-10,000 bin systems; $15-35 million for 15,000-40,000 bins.

---

Goods-to-Person (G2P) Picking Stations

Every AS/RS system serves a G2P picking station where an operator (or robotic pick arm) extracts items from the delivered tote.

G2P pick rate benchmarks:

Station Type	Lines/Hour	Notes
Basic G2P (mini-load)	300-450	Operator + scan confirm
Optimized G2P (light-directed)	450-600	Minimized wait; ergonomic position
G2P + robotic pick arm	400-800+	Highly SKU-dependent
AutoStore standard port	300-500	Port presentation time included
AutoStore dual-port	500-800	Operator alternates between 2 ports

vs. conventional RF pick: 80-175 lines/hour. G2P is a 3-4× improvement in real operations.

Multi-port advantage: Dual-port stations eliminate wait time for tote delivery. Single port: 300-400 lines/hr. Dual port: 450-600 lines/hr. Triple port: marginal gain with added ergonomic complexity.

Ergonomic requirements:

• Tote presentation: 28-42 inches from floor (golden zone)
• Screen at eye level, anti-glare
• Pick confirm: light-directed fastest; scan adds 1.5-3 seconds per pick

---

Analytical Sizing — Bozer & White and FEM 9.851

For unit-load and mini-load crane cycle time sizing, see AS-RS Sizing Methods for the full derivation. Key results:

Dual-command is ~2× as productive as single-command for a typical rack shape factor. Insist on dual-command throughput in any RFQ — a vendor demonstrating compliance via single-command-only modes is not meeting the real requirement.

FEM 9.851 is the European standard for cycle time procurement specifications. Include the applicable test case in the RFQ and require witnessed acceptance testing against that case before SAT sign-off.

Double-deep throughput penalty: 10–20% at 50% fill; 25–30% at 80% fill. Size the system with this penalty included, not as a post-hoc adjustment.

85% fill ceiling: AS/RS systems operating above 85% storage fill experience measurable throughput degradation — longer WMS search time for put-away locations, fewer dual-command pairing options, longer average crane travel. Design to 85% maximum operational fill.

AutoStore robot-to-port sizing rules: see AS-RS Sizing Methods for the worked example. Quick reference: robots = peak bin deliveries/hr ÷ 30; ports = peak picks/hr ÷ 300; robot-to-port ratio should land 2–5 (target 3–4).

---

System Sizing Methodology

Step 1: Inventory Profile

Total tote positions = (Peak active SKUs × Avg totes per SKU) ÷ Occupancy target (85%)

Example:
3,000 SKUs × 1.5 totes/SKU = 4,500 occupied totes
4,500 ÷ 0.85 = 5,294 total positions

Step 2: Required Throughput

Required throughput (moves/hr) = Peak pick lines/hr ÷ Lines per tote retrieval

Example:
1,200 lines/hr at 3 stations ÷ 1.5 lines per tote = 800 moves/hr

Step 3: Technology Selection

At 800 moves/hr and ~5,294 positions:

• Mini-load crane: 150-500 moves/hr per crane → need 2-6 cranes
• AutoStore: ~150-200 moves/robot/hr → need 4-6 robots

Both viable. AutoStore wins on footprint density; mini-load crane wins on per-position cost.

Step 4: Footprint Comparison

AutoStore at 5,294 bins stacked 8-high:

Grid cells = 5,294 ÷ 8 = 662 cells
At 600mm pitch: area = 662 × 0.36 m² = ~238 m² = ~2,560 SF

Equivalent conventional shelving: 8,000-12,000 SF for same SKU count.

---

Real-World vs. Vendor Claims

Specification	Vendor Claim	Real-World Typical
System availability	99%+	92-97% (year 1-3); 97%+ (mature, well-maintained)
G2P pick rate	600-800 lines/hr	350-550 lines/hr (mixed SKU, real operators)
Throughput ramp-up	Immediate at go-live	60-80% design at launch; 90%+ at 90-120 days
Software integration	3-6 months	6-18 months for complex WMS/WCS/ERP stacks

Design your business case at 75-85% of design capacity for year 1. Budget for the gap.