Stanford Pupper: A Low-Cost Agile Quadruped Robot for Benchmarking and\n Education

摘要

We present Stanford Pupper, an easily-replicated open source quadruped robot\ndesigned specifically as a benchmark platform for legged robotics research. The\nrobot features torque-controllable brushless motors with high specific power\nthat enable testing of impedance and torque-based machine learning and\noptimization control approaches. Pupper can be built from the ground up in\nunder 8 hours for a total cost under $2000, with all components either easily\npurchased or 3D printed. To rigorously compare control approaches, we introduce\ntwo benchmarks, Sprint and Scramble with a leader board maintained by Stanford\nStudent Robotics. These benchmarks test high-speed dynamic locomotion\ncapability, and robustness to unstructured terrain. We provide a reference\ncontroller with dynamic, omnidirectional gaits that serves as a baseline for\ncomparison. Reproducibility is demonstrated across multiple institutions with\nrobots made independently. All material is available at\nhttps://stanfordstudentrobotics.org/quadruped-benchmark.\n