Various bipedal robots have been developed to date, and in recent years, there has been a growing trend toward releasing these robots as open-source platforms. This shift is fostering an environment in which anyone can freely develop bipedal robots and share their knowledge, rather than relying solely on commercial products. However, most existing open-source bipedal robots are designed to be fabricated using 3D printers, which limits their scalability in size and often results in fragile structures. On the other hand, some metal-based bipedal robots have been developed, but they typically involve a large number of components, making assembly difficult, and in some cases, the parts themselves are not readily available through e-commerce platforms. To address these issues, we developed MEVITA, an open-source bipedal robot that can be built entirely from components available via e-commerce. Aiming for the minimal viable configuration for a bipedal robot, we utilized sheet metal welding to integrate complex geometries into single parts, thereby significantly reducing the number of components and enabling easy assembly for anyone. Through reinforcement learning in simulation and Sim-to-Real transfer, we demonstrated robust walking behaviors across various environments, confirming the effectiveness of our approach. All hardware, software, and training environments can be obtained from github.com/haraduka/mevita.
Design overview of MEVITA: excluding mirrored parts, the robot consists of 18 unique metal components, four of which are fabricated using sheet metal welding to achieve complex and large geometries as single integrated parts.
All metal components used in MEVITA.
Details of sheet metal welding for the Base-Link, Hip1-Link, and Hip2-Link. The welded sections are highlighted in red.
Comparison between existing bipedal robots and MEVITA
Circuit configuration of MEVITA: servo motors are connected to the PC via two CAN-USB interfaces, and the system is equipped with a wireless emergency stop, power relay, diode, and LiDAR/IMU.
The actual circuit layout inside MEVITA's Base-Link.
Control system of MEVITA: policies trained in IsaacGym are verified in MuJoCo and deployed to the actual hardware using common Python scripts.
Comparison of the tracking performance to the target commands when applying the trained policy to both MuJoCo simulation and the actual robot.
Walking experiments in various environments: uneven indoor terrain, grassy field, dirt surface, concrete tiles, and a gentle slope.
@inproceedings{kawaharazuka2025mevita, author={K. Kawaharazuka and S. Sawaguchi and A. Iwata and K. Yoneda and T. Suzuki and K. Okada}, title={{MEVITA: Open-Source Bipedal Robot Assembled from E-Commerce Components via Sheet Metal Welding}}, booktitle={2025 IEEE-RAS International Conference on Humanoid Robots}, year=2025, }
If you have any questions, please feel free to contact Kento Kawaharazuka.