IDEA-002 scope v002 — project boundaries and assessment
Proposed project boundaries and concrete idea assessment scope.
Version and approval boundary
Proposed v002, replacing v001 for future approval. Change: explicitly separates overall DroneRadar scope, completion of the existing radio-coverage project, and IDEA-002's current assessment deliverables. Both user approval and manager start authorization remain pending. Existing pre-gate evidence can be reused.12
Overall DroneRadar project — proposed boundary
Build an evidence-based plan for a distributed drone-observation network: what detectors observe, what data they can contribute, how observations can be combined, and whether the resulting service and participation model are useful and viable. Current phase: research, requirements and business/technical planning. A deployed detector network, payment scheme, data-sharing entitlement and commercial service are future decisions. The proposed $10 monthly benefit is an unvalidated proposal. This paragraph is a project-scope recommendation for manager reconciliation, not an edit to the shared project charter.3
Radio-coverage project — proposed completion target
A reproducible tool that predicts radio propagation for a stated environment, frequency, transmitter/receiver configuration and height, with explicit units, baselines, domain limits and measured prediction errors. Learned models should be retained only where evidence shows a useful accuracy/speed benefit. Completion requires reproducible evaluation and usable outputs; a plausible-looking map alone is insufficient. This is a proposed end state to assess, not authorization to implement it now.
IDEA-002 — current authorized-work proposal
Determine how to complete that existing tool and whether it can support DroneRadar measurement planning and receiver placement. Establish the evidence required before any propagation output can be called passive detector coverage.
Deliver:
- Reconciled inventory of the main repository, three named worktrees and four related Claude sessions. Distinguish implemented code, recorded results, independently checked findings and unexecuted proposals.
- Readiness assessment and recommended completion path, including alternatives, dependencies, measurable acceptance criteria and unresolved evidence gaps.
- Proposed DroneRadar data/interface contract: inputs, outputs, units, geographic/height conventions, model versions, uncertainty and missing-data behavior. Include the gap between propagation, signal reception, packet decoding/classification and operational detection.
- One bounded next-experiment plan: hypothesis, existing inputs, baselines, held-out evaluation, metrics, pass/fail or stop conditions, and compute/cost assumptions. Execution is a later decision.
- Curated local evidence archive, idea data manifest and session inventory; independent evidence review and knowledge validation. Final concise handoff to manager.
Methods, ownership and boundaries
Read-only inspection and lightweight arithmetic on existing results. Preserve relevant local documents with SHA-256, absolute source path, checkout/commit and dirty-file context. Use existing landscape evidence/canonical profiles; targeted primary-source research only where necessary, with archived captures and limitations. Keep original user inputs and published scope versions intact.
Own IDEA-002 files and relevant reusable evidence. Manager owns shared project scope, indexes, log and Git commits. Any helper dispatch follows current manager routing policy; one independent Sol review is required. Preserve all source checkout and untracked outputs.
Excluded: source changes, branch merges, new training/inference experiments, dataset/model downloads, paid APIs, deployment, live hardware integration, private raw flight-log imports, vendor contact and recurring monitoring. No approved compute budget for experiment execution is implied. Current assessment uses existing local records and bounded document research only.
Acceptance criteria
The manager receives a traceable current-state assessment, actionable completion milestones, a concrete proposed interface and one separately authorizable experiment. Material factual claims cite preserved evidence and its version. Simulator results are distinguished from measured field validation; missing detector parameters stay unknown. Review findings and capture/validation gaps are resolved or disclosed. No user decision or performance claim is invented.
Dependencies and decisions
No launch geography, detector export schema or sensitivity is assumed. Their absence must be recorded in the contract/evidence gaps and does not prevent the assessment. Recommended first target is reliable propagation maps and reproducible evaluation; receiver-placement and field-calibrated detection remain subsequent validation/application gates. User may amend this priority before signing off.
Next action
User approves or revises this exact v002 in the IDEA-002 chat. Record approval separately with this file's SHA-256 and notify manager. Substantive continuation requires the manager's separate start authorization for that version. The shared project boundary remains a proposal until reconciled by the manager.