Open-Source 3D-Printed Humanoid Robot Berkeley Humanoid Lite Makes Debut

Despite the significant progress and widespread attention in the field of humanoid robotics, most commercial hardware remains expensive, closed-source, and lacks transparency within the robotics community. This lack of accessibility and customization is a major barrier to the field's growth and the broad advancement of humanoid technology. To address these issues and democratize the development of humanoid robots, we present the Berkeley Humanoid Lite, an open-source humanoid robot designed for easy access, customization, and community benefit. At the heart of the Berkeley Humanoid Lite is a modular 3D-printed gearbox that drives the robot’s various components. All parts can be sourced from common e-commerce platforms and manufactured using standard desktop 3D printers, keeping the total hardware cost below $5,000 (based on U.S. market prices). The design emphasizes modularity and ease of production. To overcome the inherent limitations of 3D-printed gearboxes, such as lower strength and durability compared to metal alternatives, we opted for involute gear designs, which offer the best performance under these constraints. Extensive testing was conducted on the 3D-printed actuators to ensure their reliability and address concerns about the robustness of plastic components. To showcase the capabilities of the Berkeley Humanoid Lite, we performed several experiments, including the development of a gait controller using reinforcement learning. These experiments successfully demonstrated zero-shot policy transfer from simulation to hardware, highlighting the platform's suitability for research validation. By fully open-sourcing the hardware design, embedded code, and training and deployment frameworks, our aim is to make the Berkeley Humanoid Lite a crucial step toward democratizing humanoid robot development. All resources, including design files, instructions, and code, are available at https://lite.berkeley-humanoid.org. This initiative not only lowers the barrier to entry for researchers and hobbyists but also fosters innovation and collaboration within the robotics community.