KUKA has introduced iiQKA.AI Vision, a new software solution that brings artificial intelligence-based image processing directly into its robot ecosystem. It supports both 2D and 3D vision. The aim is to make vision-guided robotics accessible to integrators and end users who don’t have specialist machine vision know-how.
Vision applications without expert knowledge
The software covers some of the most common vision-guided tasks in industry: bin picking, assembly, quality inspection and depalletizing. Users don’t program image processing from scratch. Instead, they build applications with drag-and-drop tools, ready-made templates and step-by-step guided workflows. This cuts the time and expertise needed to get a vision cell up and running.
Behind the scenes, the AI separates parts from the background, works out suitable gripping points and sends the position data straight to the robot controller. According to KUKA, the system keeps working reliably when parts vary or lighting conditions change, two of the classic weak points of conventional rule-based vision systems.
Edge computing on NVIDIA Jetson Orin
iiQKA.AI Vision runs locally on the KUKA AI Extension Platform, which is built on the NVIDIA Jetson Orin module. No additional industrial PC, cloud connection or third-party vision system is needed. Because all processing stays at the edge, KUKA highlights three benefits:
- Low-latency AI inference for fast cycle times
- Production data that stays on site, which improves data security
- A lower total cost of ownership thanks to fewer hardware components
Integrated in iiQWorks
Vision applications are configured inside iiQWorks, KUKA’s engineering environment. This gives a single, consistent workflow from engineering through commissioning to operation. Applications can be managed and monitored centrally through a web interface.
Live at NVIDIA GTC Berlin
KUKA will show iiQKA.AI Vision in action at NVIDIA GTC Berlin, 20–22 October 2026, where visitors can see how AI-driven vision and robotics work together in real applications.
