# Gap Audit

> Generated from repository source files. Do not edit this page directly; edit the source file listed below and rerun `vp run docs:generate`.

> Source: `library/GAP_AUDIT_AND_VERIFICATION_PLAN.md`

## Gap Audit And Verification Plan

## Blocking Gaps

| Gap | Affected Output | Evidence Needed | Next Action |
| --- | --- | --- | --- |
| Hardware constraints are unknown | experiment protocols, timing analysis | EEG/display/camera/robot specs | Fill `experiments/hardware-inventory.md` and session records before data collection. |
| Project hand-eye calibration protocol needs hardware binding | robot calibration claims, system design, Exp3 | local camera, robot, calibration target, solver, station count, and acceptance metric | Protocol skeleton exists in `experiments/exp03-closed-loop-grasping/hand-eye-calibration.md`; bind it to `experiments/hardware-inventory.md` and `experiments/templates/threshold-freeze-record.md` before claiming execution accuracy. |
| Runtime interfaces are not bound to middleware | method implementation, logging, trial evidence | selected middleware, message serialization, clock synchronization, log storage | Interface skeletons exist in `system/runtime-interfaces.md` and `system/state-machine.md`; bind them to implementation details before claiming an operational system. |

## Non-Blocking Gaps

| Gap | Affected Output | Evidence Needed | Next Action |
| --- | --- | --- | --- |
| SSVEP dynamic target-box evidence is missing | SSVEP track, Exp1 | detector-bound object-level SSVEP evidence and pilot data | New AR/VR and deep SSVEP papers strengthen the surrounding literature, but object boxes generated from YOLO are still unverified; run Exp1 pilot against YOLO-generated/frozen target boxes. |
| MI no-control and false-activation evidence is missing | MI track, Exp2 | no-control/asynchronous MI evidence and pilot data | New CSP/Riemannian/transfer/deep/wearable papers deepen the lineage, but they do not replace project no-control tests; run Exp2 with rest/no-control state. |
| YOLO-to-grasp boundary needs implementation choice | vision track, system design | RGB-D, grasp pose, calibration procedure | Treat YOLO as object-candidate generator; `GRASP-005` supports closed-loop depth grasping and `GRASP-006` supports hand-eye calibration rationale, but the project implementation remains open. |
| Shared autonomy baseline needs task-level binding | Exp3 | fixed command-space, scene-aware command-space, hybrid/single-modality, and low-level/shared-autonomy task instances | Baseline families are defined in `experiments/analysis-plan.md`; bind concrete task scenes, trial counts, and numeric success/safety thresholds in the pre-data freeze checklist before data collection. |
| Low-channel feasibility is untested | Exp4 | offline ablation results | Analysis plan is defined in `experiments/analysis-plan.md`; fill exact channel sets in `experiments/hardware-inventory.md`, collect 16ch data first, then replay 12/8/6/3 channel subsets before any online low-channel claim. |

## Conflicts

| Conflict | Sources | Current Handling | Next Action |
| --- | --- | --- | --- |
| SSVEP and MI roles are now literature-supported but still not tested in dynamic scene-aware grasping. | README-001; SSVEP-001; SSVEP-003; SSVEP-004; SSVEP-005; SSVEP-011; SSVEP-016; SSVEP-019; SSVEP-020; SSVEP-021; SSVEP-022; SSVEP-023; MI-004; MI-005; MI-009; MI-014; MI-016; MI-018; MI-023; MI-025; BRI-001; BRI-004 | Mark as `inferred`, not final verified system performance. | Run Exp1 and Exp2 pilots. |
| Rehabilitation literature is stronger than the current project evidence. | MI-030; MI-031; MI-032 | Treat rehabilitation as a possible product direction, not a verified SAH-BRI-Grasp clinical claim. | Keep clinical claims out of the first system paper unless a separate clinical protocol is created. |
| YOLO can identify objects, but grasp pose estimation requires separate RGB-D/geometric/grasp-network method. | README-001; YOLO-001; YOLO-004; YOLO-005; GRASP-001; GRASP-005; GRASP-006 | Treat YOLO as object-candidate generator; cite `GRASP-005` for closed-loop depth grasping and `GRASP-006` for calibration rationale, while stating that project calibration is not yet validated. | Define the project grasp-pose implementation and bind the hand-eye protocol to hardware. |
| Shared autonomy improves manipulation in related systems, but not yet with noninvasive SSVEP-MI. | MI-009; BRI-002; BRI-003; BRI-004; BRI-005; BRI-006; SA-001; SA-002 | Mark system transfer as `inferred`; treat `BRI-005` and `BRI-006` as invasive upper-bound comparators only. | Run Exp3 closed-loop comparison. |