Covariant announced RFM-1, a robotics foundation model that allows robots use AI reasoning abilities. Covariant has been able to collect the necessary data for model training from its fleet of warehouse robots deployed across 15 countries, and the model can be further expanded by supplying it with additional varied real-world data. RFM-1 functions as a prediction engine for suction-based object manipulation robots in warehouse settings, and its capabilities are not restricted to only the situations or robots it has worked with before. It can generalize responses within its domain of training data, making it more adaptable to new situations. Additionally, RFM-1 serves as an effective simulation tool that can accurately model challenging objects and scenarios that are difficult to simulate otherwise.
Covariant’s long-term vision involves powering billions of robots worldwide with its foundation model. The company hopes to collect a wider variety of data to expand its model’s capabilities. The benefit to other robotics companies is that Covariant can help them enter production with a robust and powerful off-the-shelf AI that is cost-effective. Covariant’s goal is to build a single universal AI platform for warehouse robots by incorporating all of the manipulation data the company has within its RFM-1 foundation model.
The whytry.ai article you just read is a brief synopsis; the original article can be found here: Read the Full Article…
