Fixed Intralogistics Transformed with WMS and AGV Integration for Dynamic Warehouse Routing and Safety Optimization

Overview

A distribution center’s intralogistics ran on fixed tugger routes and manual pallet moves, which created congestion and unpredictable handoffs. Tasks were triggered by radio calls, static paths overlapped, and safety spotters paused traffic frequently. Intelligex integrated the Warehouse Management System (WMS) with an Automated Guided Vehicle (AGV) fleet manager to assign moves dynamically, route around bottlenecks, and enforce safety zones with governed overrides. Moves became predictable, manual interventions tapered, and aisles operated more safely—without replacing WMS, forklifts, or existing material handling equipment.

Client Profile

• Industry: Distribution and fulfillment
• Company size (range): Large regional DC with high-mix pallet and case flow
• Stage: Established WMS and radio dispatch; limited pilot AGVs without WMS orchestration
• Department owner: Operations & Manufacturing
• Other stakeholders: Warehouse Operations, Transportation, Maintenance/MHE, EHS/Safety, IT/OT Security, Facilities

The Challenge

Fixed intralogistics patterns were designed for steady volume, not real-time conditions. Tuggers and pallet jacks followed loops that crossed at choke points. When inbound spiked or outbound waves overlapped, queues formed in narrow aisles and near dock doors. Operators used radios to reshuffle priorities, redirect carts, and call for help when aisles stalled. AGVs from an earlier pilot ran isolated missions with static routes; when paths were blocked, floor teams paused the mission or took the load manually.

WMS knew which pallets were staged, which waves were releasing, and where replenishment was needed, but that intelligence did not guide vehicle dispatch. The AGV fleet manager knew traffic and battery status, yet it had no view of WMS task priority or due times. Safety policies were strong, but zone changes, seasonal one-way aisles, and temporary no-go areas were updated on paper maps. Any improvement had to bring these signals together, respect safety procedures, and keep familiar workflows in place for forklift drivers and pick teams.

Constraints were clear. The site could not rip and replace WMS or its MHE. Network and OT teams required segmented, read-only access to controls, and EHS needed approvals for any changes to routes, speed limits, or pedestrian zones. Carriers and dock schedules could not be disrupted during rollout, so the approach had to run alongside current operations and phase in gradually.

Why It Was Happening

Root causes were siloed dispatch and static routes. WMS released work based on waves and replenishment rules, while drivers and AGVs executed paths that did not adjust to traffic or door status. The fleet manager optimized within its own map, but it did not know which task mattered most to shipping or pick completion. Safety zones lived in binders and floor markings rather than in a system that could enforce dynamic routing and slow-down areas. As a result, congestion built in predictable places, and the only relief came from radio triage and manual reroutes.

Ownership was diffused. Operations managed waves, MHE teams managed routes, EHS set zone policies, and IT/OT guarded interfaces. Without a single, governed queue and a shared map of constraints, each group optimized locally. The floor worked hard, but pallets waited at intersections and dock doors, and AGVs were underused because they did not fit into a WMS-driven flow.

The Solution

Intelligex implemented an orchestration layer that connected WMS task creation to the AGV fleet manager’s dispatch and traffic control. WMS pushed prioritized move requests with context—source, destination, due horizon, load type—and the fleet manager accepted or deferred tasks based on live traffic, battery state, and safety zones. Digital maps defined one-way aisles, slow-down areas, and temporary no-go zones; changes required approval and were time-bound. When conditions changed, routes were recalculated, and tasks were reassigned automatically, with human-in-the-loop controls for exceptions.

• Integrations: Connected WMS platforms such as SAP EWM, Manhattan, Blue Yonder, or Oracle WMS Cloud to the AGV fleet manager via secure APIs. Optional reads from dock scheduling and conveyor PLCs were exposed through Open Platform Communications Unified Architecture (OPC UA); reference: OPC UA.
• Canonical task model: Standardized move requests with attributes for load, dimensions, priority, required window, source and destination, and handling constraints. Prevented duplicate missions and ensured inventory status updates on completion.
• Dynamic dispatch and routing: The fleet manager assigned missions based on priority and shortest safe time, routing around active bottlenecks and temporary blocks. Battery-aware scheduling queued charging during low-priority windows.
• Safety zones and policies: Digitized slow-down areas, one-way aisles, pedestrian crossings, and event-driven no-go zones. Changes followed EHS approval and effective dating, aligned to guidance such as ISO 3691?4 for driverless trucks.
• Human-in-the-loop controls: Dispatchers could freeze zones, reroute missions, or force handoff to manual moves with reason codes. EHS approved map edits and speed policy changes. All overrides were logged.
• Exception handling: Stuck or blocked vehicles raised alerts with suggested actions. Low battery, lost localization, or dropped load events triggered pre-defined recovery steps and, if needed, converted work back to manual with a clean WMS handoff.
• Dashboards and alerts: Live views showed open missions, aging, blocked aisles, and zone status. Notifications flowed to supervisors, dispatch, and EHS in Microsoft Teams for fast coordination.
• Security and segmentation: Used least-privilege service accounts, network segmentation between IT and OT, and read-only access to controls where applicable. All data movement was encrypted and auditable.

Implementation

• Discovery: Mapped high-frequency lanes, choke points, and recurring exceptions. Cataloged WMS task types, priority rules, and inventory update events. Documented safety policies, pedestrian routes, and seasonal one-way practices.
• Design: Defined the canonical move request schema, priority logic, and acceptance rules. Built the initial digital map with zones, one-way aisles, and temporary event overlays. Established EHS approval flows and dispatcher override policies.
• Build: Implemented WMS and fleet manager connectors, created task orchestration with duplicate prevention, and configured routing policies and safety zones. Stood up dashboards, alerting, and audit logs.
• Testing/QA: Ran in shadow mode: WMS generated move requests while AGVs simulated acceptance and routing; drivers continued manual moves. Compared suggested missions to actual flows, tuned priority and zoning, and conducted a human-in-the-loop review with Operations and EHS.
• Rollout: Activated low-risk lanes first, then expanded to cross-dock and replenishment. Kept manual dispatch as a controlled fallback. Introduced time-bound one-way and no-go zones during live shifts with EHS approval to validate behavior.
• Training/hand-off: Delivered role-based training for dispatchers, supervisors, drivers, and EHS on overrides, zone edits, and recovery procedures. Updated SOPs for AGV interactions and handoffs to manual. Transferred ownership of maps, policies, and thresholds to Operations and EHS under change control.

Results

Move execution aligned to real demand and aisle conditions. Pallet missions were released from WMS with clear priorities, AGVs routed around congestion, and manual drivers interacted with predictable traffic patterns. Bottlenecks cleared faster because the system throttled flow into crowded zones and automatically rebalanced routes. Radio calls to triage traffic decreased, and inventory updates posted cleanly as missions completed.

Safety and governance strengthened. One-way aisles and slow-down zones became part of the dispatch logic rather than taped arrows and reminders, and changes were approved and time-bound. When exceptions occurred, they were logged with context and resolution steps. The site kept its WMS, forklifts, and existing AGVs; the difference was a coordinated flow managed by shared rules, not shift-by-shift workarounds.

What Changed for the Team

• Before: Fixed loops and radio calls set priorities. After: WMS tasks fed a governed queue and the fleet manager assigned missions dynamically.
• Before: AGVs stalled on blocked paths. After: Routes adjusted in real time, and missions were reassigned or handed back to manual cleanly.
• Before: Safety zones lived on paper maps. After: Digital zones enforced one-way, slow-down, and no-go areas with approvals and effective dates.
• Before: Dock congestion required constant spotters. After: Traffic throttled into door areas, and moves sequenced to appointment windows.
• Before: Exceptions were ad hoc. After: Stuck vehicle, low battery, and dropped load events followed standard recovery steps with audit trails.
• Before: Inventory updates lagged. After: Completions flowed from the fleet manager to WMS automatically with verified scan events.

Key Takeaways

• Let WMS set the what and a fleet manager set the how; a shared queue with context and constraints keeps flow predictable.
• Replace static routes with digital maps and policies; zoning and one-way rules should live in the system, not on the floor.
• Keep humans in the loop; dispatch and EHS approvals for overrides and map edits sustain safety and trust.
• Start small and expand; validate behavior in shadow mode and on low-risk lanes before scaling to cross-dock and replenishment.
• Integrate, don’t replace; use APIs to connect WMS and fleet manager, leaving forklifts, scanners, and dock processes intact.
• Instrument exceptions; standardized recovery for stalls, low battery, and dropped loads reduces downtime and rework.

FAQ

What tools did this integrate with? The orchestration connected WMS platforms such as SAP EWM, Manhattan, Blue Yonder, or Oracle WMS Cloud to the site’s AGV fleet manager through secure APIs. Optional signals from conveyors or dock doors were read through OPC UA. Dashboards and alerts surfaced in the organization’s standard BI tools and Microsoft Teams.

How did you handle quality control and governance? Move requests followed a canonical schema and priority rules approved by Operations. Safety zones, one-way aisles, and speed policies were maintained under EHS change control with effective dates and audit trails, aligned to guidance such as ISO 3691?4. Dispatcher overrides and recovery steps required reason codes, and all actions were logged.

How did you roll this out without disruption? The system ran in shadow mode first while drivers continued manual moves. Low-risk lanes were enabled before expanding to busier areas. Manual dispatch remained as a controlled fallback during early cycles. No changes were made to core WMS workflows, and existing forklifts and procedures stayed in place.

How were safety zones and pedestrian interactions enforced? Digital maps defined slow-down, one-way, and no-go areas for specific times and conditions. The fleet manager respected these zones in routing and speed control. Changes required EHS approval, and temporary events—like maintenance or high pedestrian traffic—could be activated with a time-bound profile. Visual cues on the floor remained, and training covered coordinated behaviors between drivers, pedestrians, and AGVs.

How did dynamic routing decide which move to run next? WMS provided move requests with priority and due windows based on wave release, dock appointments, and replenishment needs. The fleet manager combined that context with live traffic, battery levels, and zone constraints to assign missions. If conditions changed, missions were rerouted or reassigned automatically, and WMS inventory updates posted on completion through confirmed scan events.