Platform
The technology surface behind the robot workforce.
Deeper reference for technical buyers. Vision, motion, and ERP/MES — described honestly, including the constraints.
Vision
- Short-range 3D RGB-D, fixed overhead eye-to-hand mount
- One-time hand-eye calibration is the key accuracy step
- Classical CV — one-stem-per-bin makes the vision problem tractable
- Post-grasp re-image measures grip and protrusion before placement
Motion
- Direct Python control of the collaborative arm
- Waypoint + inverse kinematics baseline for open cells
- Optional full motion planner for cluttered layouts
- Force-mode insertion, scaled per stem fragility
ERP & MES functions
- Work orders & batch release from Sunflower
- Real-time consumption and backflush of components
- Substitution and re-cost when a variety runs short
- Quality and exception capture per unit
- Full traceability and unit-level costing
- Production analytics and utilization reporting
Throughput & honest constraints
- The post-grasp re-image trades a few seconds per stem for grip certainty
- Wired low-latency network is required, not optional
- Matte backdrop and controlled lighting recommended
- SKU family selection matters — start where the geometry is stable