Physical assets are tested rigorously and deployed at scale. Pgradient Labs builds the software layer to record those tests, compare vendors objectively, and assess real-time field capability.
| Parameter | Vendor A | Vendor B | Vendor C | Threshold |
|---|---|---|---|---|
| Alt. hold ±5m / 15 km/h | ✓ 3.1m | ✗ 7.8m | ✓ 4.2m | < 5m |
| Endurance full payload | ✓ 28 min | ✓ 31 min | ✗ 19 min | > 25 min |
| Failsafe RTH latency | ✓ 0.4s | ~ 0.9s | ✓ 0.5s | < 1.0s |
| EMI resistance (GPS lock) | ~ 3/5 | ✓ 5/5 | ✗ 1/5 | > 4/5 |
Three in testing and procurement. Two in live deployment. Each one limits what the next can solve which is why the platform addresses them together.
When a procurement authority opens a contract and ten vendors submit systems for field trial, each vendor is effectively tested on slightly different parameters different sorties, different weather, different evaluation forms. The committee receives ten separate PDF reports and compares them qualitatively. The winner is chosen on opinion, not data. The result is not retained for the next procurement cycle.
| Parameter | Vendor A | Vendor B | Vendor C | Threshold |
|---|---|---|---|---|
| Alt. hold ±5m | ✓ 3.1m | ✗ 7.8m | ✓ 4.2m | <5m |
| Endurance | ✓ 28m | ✓ 31m | ✗ 19m | >25m |
| Failsafe RTH | ✓ 0.4s | ~ 0.9s | ✓ 0.5s | <1.0s |
| EMI resistance | ~ 3/5 | ✓ 5/5 | ✗ 1/5 | >4/5 |
An OEM choosing between five ESCs runs bench tests step response, current draw, thermal rise at controlled throttle. Two units look identical. But above 3,500m altitude under sustained throttle in real flight, one overheats. The data for both tests exists in separate spreadsheets, different units, different engineers' machines. Nobody links them. The selection is made on incomplete evidence.
Six months into deployment, the Army observes ESC thermal runaway at altitude. The OEM needs telemetry to diagnose it: altitude, temperature, voltage, throttle. But the raw log contains GPS coordinates, waypoint paths, and sortie patterns. That is classified. The Army either sends everything a security violation or sends nothing. The bug doesn't get fixed.
Once a drone or robot is in the field, its real-time capability remaining battery, current payload, position, speed exists nowhere in software. Operators make decisions from static spec sheets written months ago. After two sorties, a battery cycle at altitude, and sustained throttle, that 28-minute drone has 11 minutes left. No system surfaces this gap at the moment of decision.
A command operator needs to know across drones from three manufacturers, ground robots from two which asset can respond to a detected threat right now. Not by the spec sheet. By current battery state, position, and mission load. Without live capability data per asset, the answer is a guess at exactly the moment when it cannot be. With ACE, it is computed in seconds.
ATP is deployed first it builds the structured test record that makes ACE accurate. ACE is the platform destination real-time capability estimation, calibrated by actual test data not manufacturer specs.
Structured, on-premise test data infrastructure for the procurement and engineering workflow. Defines what a field test means, captures what it produces, and makes results comparable and shareable with native controls for classified data.
Real-time capability scoring for every asset in a heterogeneous fleet. Reads live telemetry battery, payload, altitude, speed and produces an accurate current-state score. Accurate because it is calibrated against structured ATP test data, not nominal specs.
Built around specific decisions that real people in these roles need to make in testing, procurement, and live operations.
One shared test template across all eight vendors. Telemetry ingested per sortie. Side-by-side comparison dashboard. The procurement record stays queryable for the next cycle.
ATP · Problem AATP links bench telemetry and in-flight data under one test template and one timestamp axis. The thermal divergence invisible on the bench becomes a flagged anomaly in the combined view.
ATP · Problem BATP's selective share template defines which telemetry fields are exported. Performance data reaches the OEM. Operational data stays on-prem. The bug gets fixed.
ATP · Problem CACE reads live telemetry from every asset and returns a ranked capability score. The answer comes from current battery, position, and payload not a spec sheet written six months ago.
ACE · Problems D + EIf you procure, test, or operate drones, counter-UAS systems, or autonomous robots, we want to understand your specific challenge and build around it.