In recent years, deep RL has emerged as a game-changer in robotic control. Instead of programming a robot arm’s motions line by line, we define a goal (through a reward function) and allow the robot to figure out how to achieve it through trial and error. It’s akin to how one might teach a child a new skill: provide feedback on good vs. bad outcomes and let them improve over many attempts. By running countless simulated trials, a deep neural network controller can gradually discover an effective strategy, whether it’s balancing a two-legged robot, picking up a delicate object, or precisely inserting a peg in a hole, all without a human writing those instructions in code directly.

One recurring challenge is the notorious sim-to-real gap, the gulf between the perfect, convenient world of a simulator and the messy complexities of reality. A policy that performs flawlessly in simulation can still fail in the real world... To close it, engineers enhance simulator realism by tuning physics parameters and use techniques like domain randomization and actuator dynamics modelling to train policies that can handle real-world unpredictability.

AICA is a prime example of this much-needed new generation of robotic platforms. Our innovative framework seamlessly integrates deep reinforcement learning (RL), classical control, and learning from demonstration within a plug-and-play architecture that empowers engineers to build smarter, more adaptive systems. With these capabilities, AICA transforms RL policy deployment into a seamless experience, combining event-driven execution with valuable tools that bridge the sim-to-real gap.

The latest version of the AICA Core introduces a new level of clarity, control, and adaptability for robotics development. Version 5.0, named Enduring Eiger, focuses on performance, usability, and scalability - making it easier to design, test, and deploy intelligent robotic applications.

Disassembling high-voltage Electric Vehicle (EV) batteries is a dangerous, manual process that creates a bottleneck for the circular economy. To solve this challenge, the Eureka-funded Project LAMBDA brought together three partners—AICA, Circu Li-ion, and the Swiss Battery Technology Center (SBTC)—to teach an industrial robot how to perform this complex task with a human-like sense of touch.

AICA, a Swiss innovator transforming industrial automation with AI-powered force adaptation technology, has secured a strategic investment led by Momenta, the leading venture capital firm specializing in Industrial Impact®. This investment, made through Momenta’s Industry 5.0 Fund, will propel AICA’s global expansion and enhance its AI-driven automation capabilities. Existing investors Spicehaus, HTGF, and Schaeffler have reaffirmed their confidence in AICA by participating in this round, which was also joined by Zürcher Kantonalbank.

AICA presented at the NVIDIA GTC AI Conference in San Jose together with partners (Universal Robots, Schaeffler, SCHUNK).

AICA at PolyScope X Festival hosted by Universal Robots

Since 2019, AICA has been focused on developing intuitive robotics software that enables engineers to automate complex robotics tasks. On October 10th 2024, we reach a major milestone: the first public release of the AICA System!

In the quest for reliability in robotics, the limitations of vision systems are becoming more evident. Despite significant progress, some challenges remain, casting a shadow over the effectiveness of visual perception alone. However, a reliable alternative or an additional solution is emerging, which offers robots the ability to reliably navigate and interact with the world even in the absence of clear visual cues.

In the rapidly evolving realm of manufacturing, innovation isn't merely a trendy term; it's a necessity. To remain competitive in today's landscape requires not only embracing change but also rethinking conventional approaches to automation.

The Swiss startup AICA closes CHF 1.2 million financing round, to give robots a sense of touch and enable autonomous learning.