Skip to content
SAH-BRI

SAH-BRI

Scene-Aware Hybrid Brain-Robot Interface

SAH-BRI is the broader framework: scene understanding, hybrid non-invasive BCI, and robot shared autonomy for turning high-level human intent into physical robot action.

SAH-BRI-Grasp is the current system instance for robotic grasping.

This site is the working research wiki for the project. Use it to move from questions to evidence, experiments, manuscript sections, and product directions.

中文入口：简体中文

Research Target

The project explores whether non-invasive EEG should serve as a high-level intention interface inside a scene-aware robot system, instead of being forced into low-level continuous arm control.

Start with Research Compass, then use Reading Paths or the Research Map.

  1. Research Compass
  2. Research Goal
  3. SAH-BRI-Grasp System Overview
  4. Research Map
  5. Open Questions

This route is designed to fit a 20-minute orientation. Use Reading Paths for evidence review, technical deep dives, experiments, manuscript work, product directions, or the dexterous arm-hand extension.

Current State

  • Corpus: 73 local user-downloaded PDFs.
  • Paper library: 73 card pages across BCI, SSVEP, MI, YOLO, grasping/calibration, hybrid BRI, and shared autonomy, with internal PDF download links.
  • Paper cards: 73 usable (72 extracted + 1 extracted from OCR).
  • Evidence policy: claims are labeled verified, inferred, or needs confirmation.
  • Runtime binding: JSONL event logs, stable record IDs, and monotonic_ns timing are drafted.
  • Experiment binding: Exp1-Exp4 are pre-data drafts; no project performance results exist yet.

Site Map

  • Research Compass: question-first guide to the research wiki.
  • Open Questions: unresolved questions that drive reading and experiments.
  • Claim Map: route from claims to evidence, experiments, and writing boundaries.
  • Evidence Index: claim IDs linked to paper-card packages, experiment routes, and writing routes.
  • Decision Log: why the framework assigns roles to SSVEP, MI, YOLO, and shared autonomy.
  • Paper Library: classified view of the 73-paper corpus and individual paper cards.
  • Literature Map: narrative map of the seven technical evidence tracks.
  • Evidence Matrix: claim-level evidence gate for manuscript writing.
  • Experiment Roadmap: Exp1-Exp4 path from component validation to closed-loop grasping.
  • Paper Kit: draft manuscript skeleton and sections backed by repository evidence.
  • Product Directions: research tooling, embodied AI + BCI, and rehabilitation/assistive positioning.
  • Dexterous Arm-Hand Extension: staged future direction beyond the current grasping baseline.

Boundaries

This documentation site is generated from local research source files. For private/internal deployments, paper-card pages can link to user-provided local PDFs prepared under /papers/. Extracted full-paper text is not published.