Robotics Intelligence

We’ve designed our first robot, Proxie, from the ground up to incorporate AI. As AI and hardware evolve together, so will Proxie's ability to handle items in the world—freeing us to focus on what makes us uniquely human.

Onboard Intelligence Today

Proxie is equipped with advanced onboard intelligence, using lidar and visual perception to safely navigate environments and perform tasks with precision.

Key Features

Lidar-based world perception

Proxie uses lidar technology to map its surroundings in real-time.

Visual object recognition

With visual object recognition, Proxie can accurately detect and identify carts.

Autonomous obstacle avoidance

Proxie autonomously avoids obstacles, ensuring smooth navigation in dynamic environments.

Advancing Now

Cobot is actively developing technologies to redefine how people and robots collaborate to make interactions more intuitive, efficient, and adaptable in complex environments.

Vision-Language Models

We’re incorporating the ability for Proxies to perceive the visual scenes around them, improving awareness of their surroundings.

Voice Interaction

Speak commands naturally, and Proxie will understand, provide helpful information, and execute tasks.

Visual SLAM

Advances in Visual SLAM will enable Proxie to create semantically rich maps, enhancing its ability to navigate and adapt in challenging environments.

Fleet Optimization: Multitasking with Efficiency

AI-driven fleet optimization ensures tasks are completed efficiently, increasing throughput while reducing downtime.

Forward Research

Cobot is at the forefront of research, working with generative AI for task visualization and foundation models for advanced robot manipulation.

Bimanual Manipulation

Our robotics foundation models team, in partnership with the University of Washington, is advancing the state of the art in bimanual manipulation, investigating novel ways to teach Proxie to perform intricate, two-armed tasks. We’re pushing the limits on how foundation models in language, vision, and action domains can help Proxie plan longer-horizon tasks and generalize to new tasks.

Video Generation for Motion Planning

Improvements in video generation will allow Proxie to visualize and plan movements before execution, improving task accuracy and model interpretability.

 Careers

Explore the Future with Cobot

Cobot is shaping the future of robotics, tackling today’s challenges and ready for tomorrow’s opportunities. Join us in pushing the boundaries of AI and robotics.