Physical AI Platform

Meet
Odin

A full-stack physical AI system that turns any forklift into a fully autonomous machine - using LiDAR and edge compute for reliable, precise, all-environment operation.

Live across warehouses, fulfilment centers and factories.

No new equipment Any forklift or tow truck No infra changes 3-6 wk deployment LiDAR safety Edge AI inference
System architecture

A full-stack physical AI platform

Odin is not a single module - it's a layered AI system from physical sensors all the way to cloud fleet management, designed to be dropped onto any vehicle in any environment.

LiDAR first sensing
LiDAR provides precise 3D spatial depth and range. That highly detailed data is used to classify pallets, trolleys, workers, obstacles - giving Odin the most accurate sensing modalities.
On-device edge compute
All perception, planning and control decisions execute locally on the forklift. No cloud dependency for safe operations - it runs even if the network goes down.
Cloud fleet orchestration
Every unit streams telemetry to a central dashboard. Assign missions, view real-time maps, review incidents, and track throughput from any device.
Sensor Array
3× Safety LiDARs · IMU · Equipment feedback sensors
Hardware
Edge Compute
Ruggedized on-bot compute · all inference runs locally
Hardware
Navigation & Controls
Live 3D mapping, low-level controls of throttle, obstacle avoidance
Autonomy
Path Planning
Dynamic real-time path planning, obstacle avoidance, rereouting
Autonomy
Perception Engine
LiDAR + AI · Object detection & classification
AI
Mission Orchestration
Task assignment, route planning, multi-unit coordination
Planning
Fleet Management Dashboard
Web & mobile · live position, missions, safety alerts
Cloud
Built general purpose

One robotics platform.
Starting with Warehouses and Factories.

Odin wasn't built only for forklifts. It's a general-purpose physical AI stack for autonomy in real-world environments. We're deploying it in warehouses and factories first, where the ROI is fastest and the need is most urgent.

Any vehicle Any environment Warehouses & factories first
The sensor advantage

Why LiDAR
beats visual AI

Other autonomous systems pick one or the other. Odin uses LiDAR for precise spatial depth, while also using AI for rich semantic understanding. The result is a system that works in every environment, every time.

Camera-only systems
Vision without depth

Cheap to deploy, but blind to exact distances. Struggles in low light, fog, dust, and reflective surfaces common in warehouses.

No accurate depth measurement
Fails in low light & dust
Struggles with reflective floors
Poor at sub-cm positioning
High false-positive rate
Flux Auto Odin — LiDAR + AI
Depth + intelligence

LiDAR maps the exact 3D geometry of your space, and then classifies every object. This is the most reliable perception stack for any controlled environment.

Spatial 3D point cloud
Odin AI Models run on 3D point cloud data
Works in dark, dusty & close-proximity conditions
Sub-centimetre location accuracy
Near-zero false positives
Detection capabilities

Odin sees everything
on your floor

Four core detection scenarios - all handled in real time,
on-device, without cloud latency.

Pallet Detection
Dense LiDAR point cloud identifies pallet geometry, orientation and load height with ±2cm precision
Worker Safety
Automatic stop within when a pedestrian enters the defined safety and stop zone. Zero false negatives
Mixed Traffic Navigation
Operates alongside manual forklifts and foot traffic. No segregated aisles, no exclusion zones required
Low-Light Operation
LiDAR doesn't need ambient light. Same detection performance at 2 AM as 2 PM
Deployment timeline

Operational in 3-5 weeks

From first conversation to full autonomous operation. Here's exactly what to expect.

Pre Kickoff
Assessment & proposal

We walk your facility (in-person or via video), review your fleet, document your workflows, and deliver a tailored scope of work and ROI estimate. No commitment required.

Week 1
Hardware installation

Our field engineers mount the Flux ARK hardware kit with sensors, edge compute unit, and control interface on your equipment. Typically 1–2 days per unit. All equipment remain manually driveable throughout.

Week 1–3
LiDAR mapping & Odin AI calibration

We build a dense 3D LiDAR map of your facility and Odin starts training on your floor specifically — your pallet types, your rack configurations, your aisle widths, your dock layouts — through supervised runs, not a one-size-fits-all model.

Week 4
Supervised autonomous operation

Gradual handover to full autonomy with your team monitoring. We co-manage the system, refine edge cases, add features, and train your operators on the robot.

Week 5+
Full autonomous deployment

Your fleet runs 24/7 without supervision. The Flux team provides ongoing remote monitoring, software updates, and a dedicated support team to ensure your fleet is delivering every single day.

Ready to see Odin
in action?

Book a technical demo. We'll walk through the solution and show you real deployment footage.

Book a technical demo ← Back to overview