Berlin-headquartered Micropsi Industries is helping manufacturers reduce downtime with addresses variance in its second-generation Computer Vision software that uses AI to improve robot accuracy in position, shape, color, lighting and background in production processes.
Launched in March, the MIRAI 2 software builds on the flagship product’s first iteration to allow machines to share skills based on training with factory data and to make changes to robotic-arm poses and skills without the use of Force/Torque sensors. The upgrades improve reliability, deployment and scalability, the company says, raising overall equipment efficiency (OEE) and lowering costs for production and onboarding.
MIRAI 2 runs on real factory data that gets parsed in the cloud and fed into the on-premises industrial-grade computer processing hardware and MES. The system, which can by operators and IT technicians without extensive knowledge of computer vision, eliminates the need for CAD data, controlled lighting, and predefinition of visual features.
It is the first compehensive upgrade of the company’s computer vision system that launched in 2019. Company officials said in a webinar that feedback from customers informed the enlarged suite of tools to help manufacturers and machines cope with changing production environments and materials.
These include:
- Robot skill-sharing that allows machines at the same site and elsewhere to duplicate desired tasks with  little or no additional training for the additional installations. MIRAI 2 can detect small differences in conditions by recording data from multiple installations to create more robust skills.
- Semi-automatic recording functionality that allows users to record data for skills training without the need for hand-guiding. Users can prepare training situations and corresponding robot target poses using remote devices, thus reducing time to implementation and improving safety.
- Training and running skills without a Force/Torque simplifies tool geometry and cabling setup. It makes skill applications more robust and easier to train, and reduces costs.
- Abnormal condition detection stops skills when unexpected conditions are encountered, allowing users to handle these exceptions in their robot program or alert a human operator.
- Industrial PC compatibility means that the software can now be run on a selection of industrial-grade hardware for higher dependability in rough factory conditions.
MIRAI 2’s arrival coincides with a renewed commitment to service, including pre-sale and onboarding. Micropsi services the US market from San Francisco and is augmenting its pricing structure with a subscription model.
The launch also follows the appointment of industry vetern Gary Jackson as CEO. He replaced founder Ronnie Vuine – still the company’s head of product – in December and remarked in the webinar — which features use cases that range of racking and insertion to end-of-line testing — that the new iteration sees the company moving into ‘production mode’ after a decade as a start-up.
Read Jackson’s blog post.