CANDIDATE COMPARISON
Explicit alternatives and rejection reasons
| Candidate | Feasible | Risk | Min clearance | Goal delay | Constraint | Status |
|---|---|---|---|---|---|---|
| Candidate generation has not yet occurred in this replay frame. | ||||||
MODENOMINAL
PRIMARY HAZARDNONE OBSERVED
TIME TO CONFLICTNOT DERIVED
REACTION MARGINNOT DERIVED
PLANFEASIBLE
NEXT REASSESSMENT50 ms
DERIVED SYNTHETIC METRICS
Reaction, separation, and latency
PHYSICAL SEPARATION9.000normalized units
MIN PREDICTED SEPARATIONNOT DERIVEDnormalized units
SAFETY MARGINNOT DERIVEDnormalized units
DECISION DEADLINENOT DERIVEDsynthetic time
GOAL DELAY0.00synthetic seconds
ROUTE DEVIATION0.000normalized units
SENSOR → TRACK200synthetic ms
TRACK → PLAN150synthetic ms
TOTAL REACTION950synthetic ms
COMMAND AGE0synthetic ms
CANDIDATES0generated
FEASIBLE / REJECTED0 / 0separate counts
APPEND-ORIENTED EVIDENCE
Canonical event timeline
| Time | Event code | Entities | Evidence | Detail | Digest |
|---|---|---|---|---|---|
| 0.000 | MISSION_STARTED | DRONE-01, GOAL-INSPECT-01 | CFG-BASELINE | Bounded autonomous inspection began. | 6c9eae771c34… |
| 3.000 | NOMINAL_ROUTE_ACTIVE | DRONE-01, ROUTE-NOMINAL | ROUTE-NOMINAL | Ownship followed the declared nominal route toward the fictional inspection objective. | 149c2f4e0511… |
| 4.600 | WORLD_OBJECT_ENTERED | WORLD-OBJ-017 | WORLD-OBJ-017 | An unexpected synthetic object entered world truth. | 08ca7f85cf57… |
| 5.000 | OBJECT_BECAME_OBSERVABLE | OBJ-017 | WORLD-OBJ-017 | The object entered the configured synthetic perception envelope. | e5f2f2ed74a5… |
| 5.100 | DETECTION_CREATED | OBJ-017 | OBS-017 | A synthetic observation created persistent identifier OBJ-017. | 5678e63d27fe… |
| 5.200 | TRACK_CONFIRMED | OBJ-017 | OBJ-017 | The persistent track reached the declared confirmation state. | f2a9314fe952… |
| 5.300 | NOMINAL_ROUTE_UNSAFE | ROUTE-NOMINAL, OBJ-017 | PRED-PRIMARY | Predicted occupancy intersected the nominal travel corridor. | 68970a8173ad… |
| 5.350 | CANDIDATES_GENERATED | PLANNER-01 | CANDIDATE-SET-001 | A bounded set of high-level synthetic alternatives was evaluated. | e53be83e5cae… |
| 5.350 | CONFLICT_PREDICTED | DRONE-01, OBJ-017 | CONFLICT-001 | A future safety-envelope conflict was derived. | 0be11bdd58d1… |
| 5.400 | TRAJECTORY_SELECTED | TRAJ-VERTICAL | TRAJ-VERTICAL | Selected VERTICAL DEVIATION using recorded criteria. | a28d844865bd… |
| 5.450 | DECISION_RECORDED | PLANNER-01 | DECISION-001 | Selection reason codes, candidate metrics, and rejected alternatives were appended. | a69646b04a3f… |
| 5.500 | SAFETY_FILTER_MODIFIED_REQUEST | SAFETY-FILTER-01 | SAFETY-FILTER-01 | Applied safe action: VERTICAL DEVIATION + SPEED REDUCTION. | f80d32e882ff… |
| 5.550 | SAFE_ACTION_APPLIED | DRONE-01, TRAJ-VERTICAL | TRAJ-VERTICAL | The synthetic action was applied to the browser-local normalized scene only. | 1ddf5e004159… |
| 7.000 | MINIMUM_SEPARATION_RECORDED | DRONE-01, OBJ-017 | METRIC-MIN-SEPARATION | The smallest predicted separation was recorded for review. | d4c76813222d… |
| 8.000 | HAZARD_CLEARED | OBJ-017 | OBJ-017 | The tracked hazard no longer intersected the active corridor. | ce99f1cfb02a… |
| 8.500 | ROUTE_REACQUIRED | DRONE-01, ROUTE-NOMINAL | ROUTE-RECOVERY | The active trajectory rejoined the declared mission route. | 0b7ebe8ffa11… |
| 12.500 | OBJECTIVE_REACHED | DRONE-01, GOAL-INSPECT-01 | GOAL-INSPECT-01 | The fictional inspection objective was reached after recovery. | d7f4eba32bf5… |
| 12.900 | DERIVED_RESULT_TERMINATED | SYNTHETIC_NULL_SINK | SYNTHETIC_NULL_SINK | The complete synthetic result terminated at the null sink; no external command was emitted. | be92f8a4853e… |
DETERMINISTIC RULES
INDEPENDENT SAFETY FILTER
Planner request versus applied safe action
NOMINAL PLANNER REQUESTNONE—
APPLIED SAFE ACTIONNONE—
The safety filter has not yet evaluated a planner request.
BOUNDED SCENARIO INPUTS
Change synthetic conditions—not a vehicle.
All inputs are scalar, allowlisted, bounded, same-origin, ephemeral, and validated by the Python domain engine. No public input can create proof claims, coordinates, routes, hardware output, payload functions, or external commands.
Scenario library
AUTHORITATIVE EQUIVALENT REPRESENTATION
Inspect every state without WebGL.
| Frame | FRAME-0000 |
|---|---|
| Mode | NOMINAL |
| Mission state | TRANSIT |
ENGINEER / REPLAY VIEW. World truth is not shown to the operator by default.
| Entity | Exists | Observable | Position | Velocity | Motion class |
|---|---|---|---|---|---|
| No world object exists at T+0.000. | |||||
| Track | Status | Observation refs | Detection confidence | Track confidence | Prediction uncertainty | Estimated state |
|---|---|---|---|---|---|---|
| No observation or track exists at T+0.000. | ||||||
| Active trajectory | ROUTE-NOMINAL |
|---|---|
| Applied action | NONE |
| Terminal | Not yet applied |
| Proof ID | Claim | Status | Evidence |
|---|---|---|---|
| TRUTH_NOT_IN_BELIEF_BEFORE_OBSERVABLE | World-truth objects are not present in autonomy belief before observability and detection | PASS | Belief activation follows each object's observable and latency boundary. |
| TRACKS_REFERENCE_OBSERVATIONS | Every active track references one or more synthetic observations | PASS | All active tracks contain observation_refs. |
| SELECTED_TRAJECTORY_FEASIBLE | The selected planner trajectory is dynamically feasible under declared normalized constraints | PASS | TRAJ-VERTICAL |
| APPLIED_ACTION_RESPECTS_SAFETY_VOLUME | The applied synthetic action is compliant or a minimum-risk hold/return | PASS | TRAJ-VERTICAL |
| INFEASIBLE_CANDIDATE_NOT_SELECTED | An infeasible candidate cannot become the selected planner candidate | PASS | TRAJ-VERTICAL |
| CRITICAL_GATES_BLOCK_INSPECTION | Existing critical gates cannot fail while the result remains an inspection state | PASS | COMPLETE |
| MINIMUM_RISK_WHEN_NO_FEASIBLE | Minimum-risk hold occurs when no compliant trajectory exists | PASS | feasible=3; applied=TRAJ-VERTICAL |
| HUMAN_OVERRIDE_PRECEDENCE | Human hold, return, land, or abort dominates bounded autonomy where specified | PASS | command=start-inspection; applied=TRAJ-VERTICAL |
| SAFETY_FILTER_SEPARATE_FROM_PLANNER | Planner intent and independent safety-filter output are recorded separately | PASS | requested=TRAJ-VERTICAL; applied=TRAJ-VERTICAL |
| APPLIED_ACTION_REFERENCES_ORIGIN | Every applied synthetic action references its originating frame and selected candidate | PASS | Applied action records include origin_frame_id, selected_candidate_id, and null-sink terminal. |
| APPEND_ORIENTED_AVOIDANCE_HISTORY | Every avoidance state transition is append-oriented and hash chained | PASS | be92f8a4853e953191a5d0702c660c0964e5a67e9435772c7deff78fe6568918 |
| NORMALIZED_COORDINATES_ONLY | Only bounded normalized scene coordinates are emitted | PASS | SYNTHETIC_NORMALIZED_SCENE_UNITS |
| NO_REAL_HARDWARE_PROTOCOL_OUTPUT | No real hardware, protocol, payload, targeting, weapon, or force output exists | PASS | hardware_interfaces=[]; flight_protocols=[]; external_command_channel=false |
| SYNTHETIC_NULL_SINK_TERMINATION | Every derived result terminates at the synthetic null sink | PASS | SYNTHETIC_NULL_SINK |
Text representation
Scenario Sudden crossing object. Outcome: SAFE AVOIDANCE — ROUTE REACQUIRED. Planner selection: VERTICAL DEVIATION. Applied safe action: VERTICAL DEVIATION + SPEED REDUCTION. Canonical events: 18. Proof checks: 14 of 14 passed.
All public controls are bounded high-level synthetic conditions. No manual flying, real route entry, hardware command, target selection, payload, weapon, or force-authority path exists.
POST-RUN ASSURANCE SUMMARY
SAFE AVOIDANCE
SAFETY
- collision
- False
- minimum separation
- 0.628
- near miss
- False
- safety filter interventions
- 1
- unsafe state duration seconds
- 0.0
AUTONOMY
- objects detected
- 1
- track confirmation latency ms
- 100
- replans
- 1
- candidate count
- 6
- selected maneuver
- VERTICAL DEVIATION + SPEED REDUCTION
- recovery time seconds
- 2.9
MISSION
- objective reached
- True
- arrival delay seconds
- 1.05
- path length penalty
- 1.904
- final mission progress percent
- 100.0
REAL TIME
- sensor to track latency ms
- 100.0
- track to plan latency ms
- 150.0
- total reaction latency ms
- 450.0
- deadline misses
- 0
ASSURANCE
- proof checks passed
- 14
- proof checks total
- 14
- authority state
- CURRENT
- navigation integrity
- True
- software configuration attested
- True
- separation assurance
- True
- human supervision
- True
- record digest
- dad037e36f9a3d052b2c180d91315e892fc36749afb67f26868c8dd815bc465f
CROSS-RUN ASSURANCE REVIEW
Compare evidence, not just outcomes.
Hold the current deterministic run as the baseline, derive a second bounded run, and identify which evidence, decisions, and assurance claims changed. This is a canonical run-to-run comparison, not a counterfactual branch.
No comparison has been derived.
RUN DELTA
BASELINE ONLY
The current run is ready to serve as a deterministic baseline. Select a bounded variation and compare.
| Measure | Baseline | Candidate | Delta / state |
|---|---|---|---|
| Outcome | SAFE AVOIDANCE — ROUTE REACQUIRED | NOT COMPARED | — |
| Selected trajectory | TRAJ-VERTICAL | NOT COMPARED | — |
| Applied action | VERTICAL DEVIATION + SPEED REDUCTION | NOT COMPARED | — |
CHANGE IMPACT
Affected claims and evidence
NO CHANGES DERIVED
| Claim | State | Triggered by | Qualification |
|---|---|---|---|
| Run a comparison to populate the change-impact review. | |||
PORTABLE, EPHEMERAL EVIDENCE.Exports contain normalized synthetic inputs, event and proof hashes, accessible equivalents, and the controlled-release boundary. They are generated in memory and are not stored by the public endpoint.Open the full comparison workbench →Run accepted-baseline regression →Read implementation record →Open evidence contract →Open comparison/export proof →
NON-CANONICAL INSPECTION
COUNTERFACTUAL — NOT EXECUTED
Compare another bounded choice without rewriting the canonical event history. The branch is derived separately and is never applied to the replay above.
NO BRANCH RUN
The canonical history digest remains dad037e36f9a3d05…
RESEARCH → SOFTWARE → PROOF
Source reports and implementation record
DOC Autonomous C2 Simulation Architecture →DOC Autonomous Delegation Architecture Research →DOC Autonomous System Assurance Platform →DOC Autonomous System Defensive Architecture →DOC Autonomous Systems Uncertainty Architecture →DOC Decision Provenance System Architecture →DOC Human Judgment Interface Design →DOC Synthetic Kill Web Simulator Architecture →DOC Designing An Ai Simulation For Real Time Evasion And Goal Directed Navigation →DOC Autonomous Simulation Mission Control →DOC Machine Sovereign Assurance Research →DOC Evulgare Simulation Workbench Research →