X1 Pro Riteņu humanoīdais robots

Press Kit

Unique Properties

Driven by the FAM series of embodied models developed in house, the X1 PRO wheeled humanoid robot achieves fully autonomous recognition, grasping, and precise handling of boxes of various types, colors, and sizes in unstructured industrial warehousing and logistics scenarios.

Dimensions & Packaging

680 mm x 615.5 mm x 1772.6 mm

Tags

Dual Modality, Wheeled, Humanoid Robot, Efficient, Self Adaptive, Autonomous Learning and Evolutionary Capability

Production Technology

The X1 PRO features dual mode switching between locomotion and operation modes. In locomotion mode, the chassis adaptively lowers to enable high-speed movement of up to 10 meters per second. In operation mode, the chassis automatically raises, coordinating the MDOF system of the robotic arm, end effector, and chassis to efficiently perform a variety of tasks. This configuration fully covers the entire operational range from ground level to the top of shelving units.

Design Challenge

The X1 PRO operates in highly complex task scenarios. Enabling the model to extract core features from very few samples for generalized learning while maintaining accuracy and stability requires fundamental redesign of the underlying algorithmic architecture, realignment of multimodal data fusion, and logical reconstruction. The technical innovation and difficulty of this development are extremely high, with no established industry reference path available.

Research

Leveraging the physical interaction modeling capability of the FAM series, the X1 PRO generates optimal grasping and motion strategies in real time based on box weight, center of gravity, and material, autonomously adjusting force and posture. It can smoothly transport boxes exceeding 40 kilograms while maintaining strong robustness under complex load variations, achieving high load, high dynamic industrial grade full process operation.

Inspiration

Industrial robotics is evolving from rigid automation to flexible intelligence. Addressing the pain points of high deployment costs and insufficient task generalization in traditional robots, embodied intelligence technology equips robots with multimodal autonomous perception and learning capabilities, enabling them to achieve human like perception and execution, a pathway considered key to overcoming current technical bottlenecks.

Project Duration

The project began in Shanghai in March 2024, was completed in May 2025, and was exhibited at CeMAT Shanghai 2025.

Operation Flow

Built on a proprietary 3D thermal map alignment architecture, the FAM series models revolutionize the hourglass shaped information processing flow of traditional vision language action models. This allows the X1 PRO to learn and deploy each complex task with only three to five demonstration samples, reducing data requirements to one percent of conventional levels, meeting the industrial need for rapid deployment and flexible response.

Image Credits

Fan Wu