General Robotics Advances Robot Deployment With Self-Engineering Grid Platform
KUALA LUMPUR, Sept 10 (Bernama) -- General Robotics, a company developing an intelligence layer for physical artificial intelligence (AI), has advanced its GRID robot intelligence platform to automate key stages of robot deployment, reducing the time and specialised expertise needed to bring robots into production.
The company in a statement said GRID can now use AI to automate processes ranging from robot onboarding and AI model integration to the creation and deployment of new robotic skills.
General Robotics founder and chief executive officer, Ashish Kapoor said the platform brings robotics knowledge spanning research, software, AI models, simulations, data and hardware into a single agent-first system, enabling intelligence to compound with each deployment.
The latest development has reduced robot onboarding time from one month to as little as two hours, while model ingestion has been cut from three days to as little as 20 minutes. Skill transfer across different robot form factors now takes as little as 1.5 hours, while new skill creation and deployment can be completed in as little as two days.
General Robotics said GRID uses knowledge graphs to convert each deployment, task, model and failure into structured, reusable intelligence while protecting customer data and intellectual property.
The platform supports robots from different manufacturers and form factors, allowing customers to select systems suited to specific tasks. Its growing ecosystem includes industrial robotics maker FANUC and bimanual mobile manipulation company Galaxea Dynamics.
The company said GRID customers include global companies in automotive manufacturing, port operations, energy generation and food and beverage production, as well as government agencies.
General Robotics said the robotics industry faces shortages of specialised talent and a fragmented development stack that requires bespoke integration of software, communications protocols and programming systems, limiting the transition from pilot projects to production-scale deployments.
GRID is designed to address these constraints through a modular library of robotic skills, foundation models and classical techniques that can be used across different robot types and applications.
The latest version of GRID can determine the skills required for a task, select and combine relevant models and approaches, create and run simulations, and deploy and monitor AI skills. The platform then uses real-world performance and failures to determine further refinements, creating a continuous development loop across connected robots and tasks.
-- BERNAMA