Commercial Drone Acceptance Testing: From Factory Spec to Site Handover

On this page

• Separate factory checks from site acceptance
• Freeze the inputs before testing
• Turn specifications into measurable tests
• Run a complete site workflow
• Read the results within their limits
• Resolve failures and repeat affected tests
• Make handover a usable transfer

Separate factory checks from site acceptance

Use factory acceptance testing (FAT) for agreed checks before shipment or release from the supplier. Use site acceptance testing (SAT) for checks after installation and integration at the receiving location. Define both terms in the purchase documents: a boxed aircraft and a permanently installed drone system will need different scopes.

NASA's systems engineering guidance distinguishes verification against requirements, environmental qualification of a design, acceptance of individual units, and certification through an appropriate authority. Those distinctions are useful here, but NASA's process is not a mandatory commercial drone acceptance program. Passing the buyer's test does not itself confer certification. NASA, Product Verification.

Ask the supplier which records apply to the design and which identify the delivered serial number. A design-level environmental report and an individual unit's inspection record answer different questions. Reserve site-dependent checks for SAT, and name any supplier-facility checks that must be repeated after transport or installation.

If the purchase is an inspection service rather than an aircraft system, the drone inspection scope-of-work guide covers the separate task of defining commissioned coverage and deliverables.

Freeze the inputs before testing

Start with a signed test schedule linked to the purchase requirements. Identify the aircraft and payload serial numbers, batteries, controller, firmware, flight application, processing software, accessories, and service subscriptions. Record configuration changes during the trial rather than treating them as invisible adjustments.

The buyer should provide a representative task, permitted test area, intended crew, reference targets or survey information where needed, and an actual receiving system for the files. The supplier should provide configuration records, relevant manuals, factory results, and the equipment needed to run the agreed procedure.

Assign a test lead, a person authorized to stop the activity, a technical reviewer for each output, and the buyer's acceptance signatory. Record prerequisites, weather limits, stop conditions, required repeats, and how a postponed test will be rescheduled. NASA treats an agreed requirements baseline, prepared procedures, suitable instruments, and readiness of people and facilities as verification inputs. NASA, Product Verification.

Give every run a test ID and retain the procedure revision, date, operator, configuration, instruments or reference data used, observed conditions, result, and evidence filenames. Agree the repeat count and treatment of aborted runs in advance so the final report cannot select only the best attempt.

For an industrial location, settle entry, operating-state, and work-area constraints before arrival. The refinery inspection planning guide develops those site inputs for a demanding application.

Turn specifications into measurable tests

A published maximum is useful only with its conditions. For example, DJI lists a 55-minute maximum flight time for the Matrice 350 RTK, measured at approximately 8 m/s without a payload, in windless conditions, until the battery reaches 0%. That is a manufacturer specification, not a working inspection mission or a suitable battery target for a site test. DJI, Matrice 350 RTK specifications.

Translate the mission into completion requirements using the actual payload and an agreed operational battery reserve consistent with the aircraft instructions. Record useful task completion and landing reserve alongside elapsed flight time.

Example acceptance-test schedule

This schedule is an editorial starting point. Choose the actual values, sample sizes, repeat counts, and allowed exceptions for the intended work before the trial; no row supplies a universal pass value.

Source basis: adapted purchasing questions informed by NASA's product validation guidance, NIST's drone test explanation, and PIX4D's accuracy guidance. The commercial decisions in the final column are recommendations for agreement, not requirements imposed by those sources.

For mapping, specify whether the concern is geometry within the model or position in an external reference frame. PIX4D distinguishes relative from absolute accuracy and recommends checkpoints to assess absolute accuracy. A sharp-looking orthomosaic does not settle either requirement. State the reference system, units, checkpoint provenance, and error measure in the test schedule. The responsible survey reviewer should also agree how reference-measurement uncertainty and borderline results will be handled before seeing the trial output. PIX4D, relative and absolute accuracy.

Run a complete site workflow

Start with the buyer's task request and end with the recipient opening the deliverable. Include setup, collection, processing, review, export, and corrections. Keep their durations separate so an aircraft's flight time cannot stand in for total delivery time.

Use the intended operating crew where possible. NASA's validation guidance evaluates the product against user expectations in its intended environment and includes the people who operate it. A supplier specialist's demonstration can establish a useful reference, but record any assistance needed when the buyer repeats the task. NASA, Product Validation.

For a hypothetical visual-inspection purchase, require the trial report to link every selected observation to an original image and the correct asset identifier. Have the maintenance recipient open the exported package without relying on the supplier to navigate it. If that fails, record the handoff as incomplete even if the aircraft portion succeeded.

Keep each payload's trial tied to the purchased output. For a pipeline system, use the visual, thermal, and methane workflow explanation to distinguish the outputs before writing the test schedule. One sensor demonstration should not silently stand for every proposed service.

Read the results within their limits

NIST's aerial test methods provide repeatable measures of system capability and pilot proficiency. The organization using them decides what performance is acceptable for its mission. NIST also states that it does not train or certify pilots. Ask for the method, configuration, conditions, and scores behind a proficiency claim. NIST, aerial drone test FAQ.

In the acceptance report, separate demonstrated results from remaining questions:

• A successful representative task covers that configuration and those conditions; list the untested operating conditions explicitly.
• A short series of successful flights should not be presented as a quantified long-term reliability result without a separate reliability study.
• A simulated interruption demonstrates the modeled response; identify what still depends on the real aircraft, communications, or installation.
• A passed software export proves the tested transfer; record the application versions and receiving-system settings needed to repeat it.

These are limits on how the buyer should interpret the trial, not reasons to discard it. Where an untested condition is essential to the purchase, schedule further proof or narrow the accepted use in writing.

Resolve failures and repeat affected tests

Mark each requirement as passed, failed, or not tested, and retain failed and aborted runs with the successful ones. For each unresolved item, record the unmet requirement, observed result, file or log reference, likely cause if known, corrective owner, and proposed retest. Distinguish an equipment defect from a test that could not proceed because its prerequisites were absent. A postponed weather-dependent test remains not tested.

Agree commercial treatment before mobilization: who pays for supplier-caused rework, how site delays are handled, and which incomplete items prevent acceptance or an invoice milestone. Those are negotiated purchase terms, not automatic consequences of a test score.

After a change, identify affected functions and repeat the relevant tests. If a processing update corrects an export, check that the corrected output still preserves the required coordinates and asset references. Do not relabel the earlier result as a pass.

Use documented manufacturer procedures and a qualified test plan for contingency behavior. Do not improvise airborne power removal, radio interference, or navigation disruption to create a failure. If safe demonstration is unavailable, keep that requirement open or agree a suitable alternative method with its limits recorded.

Make handover a usable transfer

NASA's transition guidance includes configuration identification, operating documentation, receiving-site readiness, training, and support. Apply that principle to a package the buyer can actually use. NASA, Product Transition.

Hand over the tested configuration list, signed results with supporting files, resolved and outstanding discrepancies, operating and maintenance instructions, training records, service entitlements, and support contacts. Verify buyer account access and a usable data export. Record restrictions that remain in force and who owns each follow-up action.

Before signing, resolve these buyer questions:

• Are these the delivered units and software versions, or a demonstration configuration?
• Which requirements passed at the supplier, and which passed at this site?
• Can our crew and data recipient repeat the task without unrecorded supplier help?
• What remains untested, who will close it, and by when?
• What changes require another check, and what support applies after handover?
• Does this signature acknowledge receipt, technical acceptance, or a payment milestone?

Sign against the named requirements, configuration, and retained results. If the parties agree conditional acceptance, list the permitted use, outstanding items, closure dates, and commercial consequences explicitly. Do not use that label to imply an unresolved safety-critical function passed. If a necessary capability remains unproven, settle the next test or scope change before calling the system accepted for that work.

Source notes

• NASA, Product Verification: systems engineering guidance on requirements, test preparation, and the distinction between unit acceptance and design qualification.
• NASA, Product Validation: guidance on intended environments, users, and operational expectations.
• NASA, Product Transition: guidance on configuration records, documentation, training, and transfer to the recipient.
• NIST, aerial drone test FAQ: government explanation of performance measurement, organization-set pass criteria, and certification limits.
• DJI, Matrice 350 RTK specifications: manufacturer source for the stated maximum-flight-time example and its test conditions.
• PIX4D, relative and absolute accuracy of drone mapping: manufacturer technical guidance on mapping accuracy and independent checkpoint assessment.

Last checked: September 9, 2026.