Forgis demonstrates voice-directed robot control on Arduino UNO Q

Date: 2026-07-13
Type: ARTICLE
Desk: Robotics
Topics: Edge Inference, Industrial Automation, Natural Language Interfaces, Robot Control
Deck: An edge model turns spoken instructions and camera observations into a pick-and-place sequence.
URL: https://korthosrobotics.com/research/forgis-arduino-uno-q-voice-robot-control-2026
Referenced profiles: Forgis (https://korthosrobotics.com/ecosystem/company/forgis); Qualcomm (https://korthosrobotics.com/ecosystem/company/qualcomm); Arduino (https://korthosrobotics.com/ecosystem/company/arduino)

Image: Forgis
Forgis has demonstrated spoken control of a robotic arm using its model on an Arduino UNO Q board powered by Qualcomm’s Dragonwing QRB2210 processor. Detailed by Qualcomm on 13 July 2026, the setup takes a voice instruction from a smartphone, identifies objects through a USB camera and converts the request into a pick-and-place sequence.
A Snapdragon-powered phone handles speech recognition and sends the transcript to the board. Forgis’s model uses the prompt and camera observations to select objects, determine their destination and calculate a motion plan. The UNO Q’s LED matrix displays the agent’s current state. Qualcomm reports 20-millisecond recognition of component count and position in this demonstration.
Forgis’s industrial work also combines robot CAD data, PLC signals and project specifications to produce structured skills and parameters for machine control. Running the demonstrated inference locally connects the language interface to the work cell without a cloud round trip. The arm then executes the planned motions through its control system.
The Zurich company disclosed $4.5 million in pre-seed funding and joined the ETH AI Center in December 2025. Founders Federico Martelli, Camilla Mazzoleni and Riccardo Maggioni developed the approach after interviewing more than 200 automation engineers about fragmented factory systems. Their earlier public work included a chess-playing robot demonstration on ETH’s Polyterrasse, preceding the voice-directed industrial handling setup.