SCOPE enables 100% safe autonomous robot navigation
New algorithm guarantees zero unsafe movements in unknown 3D spaces
A new arXiv paper proposes SCOPE, a path-planning framework for navigating unknown 3D environments with a body-mounted limited-field-of-view sensor. SCOPE plans by certifying that the robot-inflated volume along a motion is observed and verified free before execution, building a graph of positions with fully known-free safety volumes. In 60 randomized tasks across three unknown 3D environments, the system reached every goal while maintaining near-zero entry into non-certified inflated space, and the preview mechanism reduced mean mission time by 27%. Real-robot demonstrations in two representative scenarios also validated the full system.
- SCOPE guarantees zero unsafe robot movements in unknown 3D environments through safety-volume certification
- Achieved 100% goal completion in 60 real-world tests with 27% faster mission times using preview optimization
- System decouples optimistic goal guidance from certified execution through recursive observation planning
Why It Matters
Critical breakthrough for autonomous robots in construction, search-and-rescue, and industrial inspection where safety failures are catastrophic.