Report
Teaser, summary, work performed and final results
Periodic Reporting for period 1 - OS for Ind robots (Disrupting industrial robots with AI software - how new AI-driven software can be used to program industrial robots to do new tasks faster, resulting in massive productivity savings)
Teaser
Generally, industrial and collaborative robots used by various industries in the world are ‘dumb’. They can be programmed to do simple repetitive tasks in environments where all objects are in known places. However, they fail in dynamic and complex environments. Adding...
Summary
Generally, industrial and collaborative robots used by various industries in the world are ‘dumb’. They can be programmed to do simple repetitive tasks in environments where all objects are in known places. However, they fail in dynamic and complex environments. Adding intelligence with a vision for the robot enables it to perform in dynamic environments. For example, in cases where parts to be processed by the robot are not in known locations but arrive in chaotic piles. Enabling robots to pick stuff is a still major challenge. In the project Cambrian developed its AI technology that provides intelligence for the robot to perform the picking tasks robustly for manufacturing purposes.
Intelligent robots that can do more and are easy to set up are very beneficial for the society. It will mean cheaper goods, more variety and a more local manufacturing. For example, manufacturing can come back from cheap labour countries to developed countries with the help of intelligent robots. And, instead of doing repetitive tasks more people can concentrate more on creative and value-adding tasks.
The main objective in the project was to increase the technology readiness level of Cambrians AI-based picking skill on robots. More specifically, develop technology to pick any kind of parts from chaotic piles in boxes with AI very robustly and at a level acceptable to the industry. Also, it was also studied what are the best ways to commercialise the technology.
Work performed
In the H2020 SME Phase I project (Jan 19 - Sep 19), the company developed and increased the technology readiness level of its AI technology with a brand name manufacturer, and also studied the best way to commercialise and productise the technology.
The technology and its features were motivated by discussions with a major electronics OEM and a large automotive OEM. More specifically Cambrian increased technology readiness level of its binpicking technology. Which is the ability to pick manufacturing parts from a chaotic piles in boxes. This kind of technology is in great demand by the industry. Cambrian also increased the speed and robustness of the picking. The technology relies fully on AI models (deeplearning) developed and researched by Cambrian.
In addition Cambrian conducted studies on how to best commercialise and disseminate the technology by having discussions with companies in the manufacturing industry in UK and in Germany.
Cambrian has begun to disseminate the technology with key partners to the industry in UK and in Germany. Other European countries and US and Asian markets will be addressed in 2020.
Final results
Cambrian has achieved a better binpicking technology relying on AI better than industry has seen so far. This is the consensus from many manufacturing and automation companies that cambrian has showed the technology.
Directly the project has contributed to the labour of 3 people and through other activities it will indirectly help other manufacturing and industry companies and enable them to thrive.
Website & more info
More info: http://caint.io.