AI Skill Report Card

Engineering UAV Systems

A87·Sep 24, 2026·Source: Web

UAV Development Engineering

14 / 15

When given a UAV design request, immediately establish the mission profile before proposing solutions:

Mission Profile Checklist:
- Payload mass & type (EO/IR camera, radar, sensor package, cargo)
- Required endurance (minutes/hours) and range (km)
- Cruise/dash speed requirements
- Environment (wind, temperature, altitude, GPS-denied?)
- Configuration constraint (fixed-wing, multi-rotor, VTOL hybrid, HALE/MALE)
- Regulatory/certification context

Example immediate response to "Design a surveillance drone for 3 hours endurance carrying a 2kg EO/IR gimbal":

  1. Flag the core trade-off: fixed-wing (efficient cruise, no hover) vs VTOL hybrid (hover capability, ~30-40% energy penalty)
  2. Size battery/fuel by required endurance × power draw, then back-calculate airframe weight budget
  3. Recommend configuration: for 3hr + 2kg payload, propose fixed-wing or VTOL-hybrid, not multirotor (poor endurance-to-weight)
Recommendation▾
Add 1-2 more concrete examples with fuller numeric input/output (e.g., actual battery mass/power calculation worked through end-to-end) to strengthen the examples section further.
15 / 15

Progress checklist for full UAV design tasks:

  • Step 1: Characterize the mission (payload, endurance, range, environment, autonomy needs)
  • Step 2: Select airframe configuration and flag primary trade-offs
  • Step 3: Size propulsion/energy system against the weight/endurance budget
  • Step 4: Design/verify structure (materials, load cases, weight targets)
  • Step 5: Define payload integration (power budget, data link, gimbal/stabilization)
  • Step 6: Specify GNC/autonomy stack (control laws, navigation, path planning)
  • Step 7: Define test protocol (bench, wind tunnel/CFD, flight test stages)
  • Step 8: Address certification/safety requirements

Step Details

Step 1 — Mission Characterization Extract or ask for (but don't over-ask — infer reasonable defaults and state assumptions): payload mass/type, required loiter/endurance, range, cruise speed, operational environment (wind, GPS-denied, temperature), and any regulatory class.

Step 2 — Configuration Selection Match configuration to mission using this decision logic:

  • Need hover/VTOL + short endurance → multirotor
  • Need long endurance + no hover requirement → fixed-wing / flying wing
  • Need hover + long endurance/range → VTOL hybrid (accept ~20-40% energy penalty for transition + hover systems)
  • HALE/MALE → flying wing or high-aspect-ratio fixed-wing, turbo-prop or electric-solar hybrid propulsion

Always state the trade-off explicitly, don't just recommend silently.

Step 3 — Propulsion & Energy Sizing

  • Compute required power: P = (payload+airframe+propulsion weight) × factor from L/D or hover disk loading
  • Battery: Energy(Wh) = Power(W) × endurance(hr) / efficiency; then Battery mass = Energy / specific energy (~150-250 Wh/kg for current LiPo, higher for Li-ion/fuel cells)
  • Iterate: battery mass affects total weight affects required power — do 2-3 iteration passes and show convergence
  • For >2hr endurance or heavy payload, evaluate fuel cells or small piston engines vs batteries

Step 4 — Structures

  • Set weight budget top-down from Step 3, allocate: structure ~25-35%, propulsion ~15-25%, energy ~30-45%, payload per mission requirement
  • Recommend material by load case: carbon fiber for primary structure (stiffness-to-weight), Kevlar for impact zones, 3D-printed composites for low-load/complex geometry brackets
  • Flag fatigue/vibration isolation needs when payload includes optics or sensitive sensors

Step 5 — Payload Integration

  • Define power budget separately for flight-critical systems vs payload (never let payload starve flight controller/actuators)
  • Specify data link bandwidth needs (EO/IR video vs telemetry-only)
  • Gimbal stabilization requirements tied to sensor resolution/zoom

Step 6 — GNC/Autonomy

  • Default control approach: PID cascaded loops for standard platforms; recommend adaptive/state-space control only when envelope is highly nonlinear or variable (VTOL transition, heavy payload variation)
  • Navigation: GPS+INS fusion baseline; add visual-inertial odometry or optical flow explicitly for GPS-denied requirements
  • Path planning: state algorithm class (A*/RRT* for offline, potential fields/MPC for real-time obstacle avoidance) matched to compute budget

Step 7 — Testing Sequence: bench/HIL (hardware-in-loop) → tethered hover/taxi test → incremental flight envelope expansion (low altitude/speed first) → full mission profile test. Always specify go/no-go criteria per stage.

Step 8 — Certification Flag applicable weight-class regulations (e.g., recreational vs commercial thresholds) and note that specific national regulations must be verified by the user — do not assert compliance without local regulatory confirmation.

Recommendation▾
Name could be more gerund-consistent with domain convention but 'engineering-uav-systems' is acceptable; consider 'designing-uav-systems' for closer alignment with action-oriented skill naming.
16 / 20

Example 1: Input: "I need a drone that can hover to inspect towers and also fly 20km to get there efficiently." Output: Recommend VTOL hybrid tail-sitter or quad-plane configuration. State trade-off: pure multirotor achieves the hover but burns range fast (~15-20 min/20km round trip is marginal); pure fixed-wing has range but can't hover for inspection. Propose quad-plane: fixed-wing cruise for transit (efficient L/D ~12-15), quad rotors for hover during inspection, accepting ~25% weight penalty from dual propulsion systems. Follow with power budget calculation for both flight phases.

Example 2: Input: "My multirotor's flight time dropped by 30% after adding a new gimbal payload." Output: Walk through weight-vs-power relationship: hover power scales roughly with weight^1.5 (disk loading), so added payload mass has outsized effect vs added weight in cruise flight. Ask for payload mass added and total AUW (all-up-weight) before/after. Recommend checking: (1) whether battery C-rate is now undersized for peak current draw, (2) whether prop selection still matches new disk loading, (3) whether gimbal draws continuous power reducing propulsion budget. Suggest re-sizing props or battery capacity rather than accepting the loss.

Recommendation▾
Consider adding a brief troubleshooting/pitfall table format for faster scanning of common failure modes vs prose paragraphs.
  • Always state trade-offs explicitly — never present a single option as if there were no alternative.
  • Do iterative weight/power convergence for any sizing calculation (2-3 passes minimum) rather than a single-pass estimate.
  • Separate flight-critical power budget from payload power budget — never let payload load risk flight safety margins.
  • Recommend the simplest control/navigation architecture that meets mission needs; reserve adaptive control, VIO, or complex path planning for cases that actually require them (compute and complexity cost).
  • Tie every structural/material recommendation to a specific load case or requirement, not generically.
  • Stage flight testing incrementally — never recommend jumping to full mission profile testing without incremental envelope expansion.
  • Don't recommend multirotor configurations for long-endurance/long-range missions — disk loading physics make this inefficient regardless of battery improvements.
  • Don't quote flight time/range figures without stating the weight and power assumptions behind them.
  • Don't ignore the battery-mass-vs-endurance feedback loop — adding battery capacity adds weight which increases power draw which reduces the endurance gain (diminishing, sometimes negative, returns past a certain point).
  • Don't assert regulatory/certification compliance — always flag that local aviation authority rules must be independently verified.
  • Don't propose complex GNC algorithms (adaptive control, SLAM, etc.) as defaults when the mission profile doesn't require them — this adds unnecessary computational and integration risk.
0
Grade AAI Skill Framework
Scorecard
Criteria Breakdown
Quick Start
14/15
Workflow
15/15
Examples
16/20
Completeness
18/20
Format
14/15
Conciseness
13/15