From the underlying architecture of the Covariant Brain to the design of the end-to-end robotic solutions, every aspect is optimized for speed, without compromising on accuracy.

For example, when a robot needs grasp points to pick up an object, the Covariant Brain responds in under 350 milliseconds — that translates to over 10,000 PPH or picks per hour.

Other key capabilities, such as intelligent motion planning, further optimize the path and speed of transporting objects based on the size, orientation, etc. of the object and the destination.

When designing end-to-end solutions, each action or step is designed for speed (e.g. running multiple steps in parallel to reduce wait times). All of these, in combination, result in Covariant-powered robots achieving high throughput that increases efficiency.

Speed without accuracy would hinder operational and economic efficiency by requiring too much human intervention. So, just as important as speed, Covariant-powered robots can operate without human intervention over 99.9% of the time.

The Covariant Brain is trained on millions of picks, across a wide variety of SKUs and product types from a range of verticals such as health & beauty, apparel, mail & parcel, general ecommerce, grocery, and pharmaceutical.

This means that Covariant-powered robots can successfully handle virtually any type of object one might encounter in a warehouse setting. Furthermore, capabilities such as double-pick detection, transport-drop detection, perturbation, etc. help further minimize any errors.

Robots that are limited to handling a defined set of items might be able to achieve high speed and accuracy due to the repeatable and predictable nature of the application. However, adaptability is key in modern warehouses where new SKUs are constantly introduced with ever-changing conditions—adaptability that only an AI-based solution can provide.

So, instead of being limited to a specific set of SKUs, the Covariant Brain is trained on a virtually infinite number of scenarios with hundreds of thousands of SKUs with variations in packaging, lighting, positioning, etc.

This allows Covariant-powered robots to easily adapt to practically any type of SKU or object, from a tiny cosmetics product or apparel wrapped in a transparent polybag to a large cardboard parcel or cases for a pallet. And what if a robot experiences a new challenge? The Covariant Brain will try different methods to overcome the challengeuntil it is successful.

The Brain continuously learns fromthese experiences and keeps getting smarter