AICA robotic software enables manufacturers to quickly automate and enhance complex parts of their production process. By using dynamic force-compliant control and machine learning algorithms, we enable robots to safely perform and improve assembly, finishing or machine tending tasks. Schaeffler, our main client in the automotive industry, used to spend 200 hours programming a robot for gearbox insertion. Instead, with our solution, they are training a robot in one hour, and the approach is more robust to external perturbations. This “smart assembly” application is now part of our library of software modules that can be combined into specific applications to simplify robotic programming.
We operate both as a full-stack robotics software provider (offering complete solutions to our clients to develop their industrial applications), or intervene at specific/desired steps of our customer’s manufacturing process. Our approach includes multi-hardware control, perception and sensor fusion (force & vision) and Artificial Intelligence (AI) components. Lastly, our learning approach requires only a few iterations of autonomous self-learning trials, e.g. about 100 assembly and disassembly steps in the gearbox insertion use case. This is a unique feature from our solution, not available on the market otherwise.
The software solution we have developed is the outcome of research, led at the Learning Algorithm and System Laboratory (LASA) from EPFL, in control and machine learning for collaborative robotics. Our solution is not restricted to those advanced robotic hardwares, but can also enhance production of industrial robotic systems.
Throughout the Innobooster project, AICA has successfully leveraged its solution to acquire several Proof of Concept (POC) projects, 2 with Schaeffler, an international manufacturer for the automotive industry, and 1 with SunMiL, a chemistry laboratory from EPFL, focusing on lab automation.. Those projects are now in their final phase and will be included in a new production line, where our robotic task learning approach and dynamic programming will be a central part of the manufacturing process.
To support its growth, and move from the Proof of Concept phase to industrial grade application, AICA is closing a CHF 1.2M Seed round with private investors from Switzerland and Germany.
Am Projekt beteiligte Personen
Letzte Aktualisierung dieser Projektdarstellung 10.05.2023