An Open-Source, Reproducible Tensegrity Robot that can Navigate Among Obstacles

The project addresses the complex modeling and control challenges of tensegrity robots—structures made of rigid rods and elastic cables. While tensegrity is a known field (pioneered by NASA's Super Ball Bot and academic labs like Berkeley's BEST Lab), this project focuses on 'reproducibility' and 'navigation,' providing a full-stack open-source reference for a 3-bar configuration. The 0-star count combined with 8 forks is a classic signature of an academic project: it has no general developer traction, but is being used or scrutinized by a small peer group of researchers. The defensibility is low (3) because it is essentially a reference implementation for a niche academic interest; there is no 'moat' other than the specific physical-math mapping for this 3-bar geometry. Frontier labs like OpenAI or Google DeepMind have largely exited the niche hardware space to focus on general-purpose foundation models for robotics (e.g., RT-2), making the frontier risk 'low.' The primary 'competitors' are other specialized labs (e.g., Squishy Robotics) who have commercialized similar concepts for emergency response. Displacement is likely within 1-2 years as more sophisticated soft-robotics simulators (like MuJoCo's recent updates or NVIDIA Isaac) make these custom modeling implementations obsolete.