en.Wedoany.com Reported - Atom Robot has officially launched the lightweight bipedal intelligent robot Tianzu AR-L1, targeting embodied intelligence algorithm R&D for universities and enterprises, positioned as an open platform that covers both laboratory debugging and industrial real-world testing.
Currently, embodied intelligence research in China is iterating rapidly, and lightweight bipedal robots have become an important hardware carrier for gait training and visual-motor fusion research. However, most research-oriented bipedal platforms are designed for steady-state indoor experiments, with weak environmental adaptability and difficulty migrating to real-world conditions. Meanwhile, the deep coupling between motion systems and visual perception prevents developers from conveniently deploying self-developed algorithms, resulting in high research trial-and-error costs. Leveraging its accumulated expertise in industrial robot body development and mass production, Atom Robot has launched this product with the core concept of using a single hardware platform to accommodate both algorithm iteration and industrial scenario testing.
In terms of key specifications, the Tianzu AR-L1 weighs 20 kg, with body dimensions not exceeding 392mm × 380mm × 845mm, a maximum obstacle-crossing height of 15 cm, a maximum climbing slope of 20°, joint peak torque of 60 N·m, peak speed of 5 rad/s, a rated top payload of 5 kg, and a standard operating endurance of 2 hours with support for rapid battery swapping.
The body adopts a symmetrical bionic bipedal layout with multi-degree-of-freedom serial joints, capable of climbing slopes of ≤20° and crossing obstacles of ≤15cm in height, smoothly navigating steps, trenches, and gravel surfaces. With a total weight of 20 kg, it can easily pass through standard doors, corridors, and elevators, accommodating both fine indoor debugging and complex outdoor industrial site experiments. The frame is constructed from aviation-grade aluminum alloy combined with high-strength engineering plastics, and the top features a standardized expansion platform with a rated payload of 5 kg, allowing direct mounting of LiDAR, depth cameras, dual-light gimbals, computing modules, and other peripherals—enabling the construction of multi-sensor fusion experimental environments without secondary mechanical processing.
At the architecture level, the Tianzu AR-L1 adopts a decoupled design between high-dynamic joints and the vision system, with fully open low-level control interfaces. The motion system and perception system operate independently without locking each other, allowing R&D teams to debug gait separately, iterate vision strategies independently, and flexibly replace self-developed perception solutions. In terms of hardware power, the device is equipped with a 46.8V high-voltage power platform, delivering joint peak torque of 60 N·m and peak speed of 5 rad/s. The body integrates humanoid intelligent gait algorithms that can adaptively level and restore steady states under external force disturbances, reducing manual parameter tuning workload.
In terms of structural rigidity, overall tolerance control, material fatigue resistance, and environmental adaptability, the product is manufactured to industrial robot mass-production standards, capable of withstanding vibration, temperature variations, and unstructured terrain in industrial settings. The device is paired with a 4.5Ah industrial-grade lithium battery, offering a standard operating endurance of 2 hours and supporting rapid battery swapping to meet long-duration continuous experiment requirements. Indoor simulation training, algorithm iteration, outdoor real-world validation, and industrial scenario pre-research can all be completed on the same device, reducing the need for research teams to purchase multiple redundant systems.
In terms of application directions, the Tianzu AR-L1 primarily targets two types of scenarios: first, core research scenarios in universities and enterprises, suitable for laboratories in automation, artificial intelligence, robotics engineering, and related fields, supporting algorithm validation and research in legged gait control, reinforcement learning, visual embodied intelligence, multi-sensor fusion, autonomous navigation, and mobile manipulation; second, technology pre-research in industrial scenarios, directly applicable to real-world testing such as factory inspection, rail transit environmental perception, complex terrain navigation in industrial parks, and obstacle avoidance in infrastructure scenarios.
The Tianzu AR-L1 is now officially available, with prototype testing appointments and technical collaboration open to university laboratories and enterprise R&D teams, accompanied by development documentation, SDK resources, and localized technical support.
