CuePilot

CuePilot is a web + desktop app for teams running human-robot interaction (HRI) studies in labs, hospitals, warehouses, and schools. It ingests synchronized video, robot logs (ROS/ROS2 bag files), and simple observer notes, then helps you label key interaction moments (handoffs, hesitations, overrides, near-misses) with consistent taxonomies. An AI assistant proposes timestamps, event labels, and short summaries, but keeps humans in control with review queues and audit trails. The app outputs clean datasets, highlight reels, and metrics dashboards (task success, time-to-trust, intervention rate, proxemics violations) that are actually usable in papers, safety reviews, and product decisions. It also generates “study-ready” artifacts: consent checklists, anonymization workflows (face blurring), and reproducible exports for R/Python. This is not a robot controller; it’s the boring, high-leverage glue between experiments and decisions.

← Back to idea list