Start here. This module builds your mental model of how UAVs are classified, how they are designed, and what military missions they are built to perform — from the smallest pocket drone to a 15-tonne surveillance aircraft. No prior aerospace knowledge assumed.
📚 REFERENCES: Fahlstrom & Gleason §2–5 · Sadraey §1–2
A UAV (Unmanned Aerial Vehicle) is simply an aircraft with no human pilot on board. Someone — or something — is still controlling it, either from the ground via radio link, or through onboard software that makes decisions autonomously. The key difference from a regular aircraft isn't that it flies differently; it's that removing the pilot changes everything about the design — you save weight, remove life-support needs, and can fly longer, higher, and into more dangerous places.
The broader term is UAS (Unmanned Aircraft System), which correctly captures that the flying vehicle is just one part of a larger system. You also need a ground station to send commands, a data link (radio) to carry those commands and send back video or sensor data, and a launch/recovery system. Fahlstrom & Gleason define UAS as "the integration of the air vehicle, payload, control element, and communications link into a functional system." Fahlstrom §2.1
Think of a UAS like a television news crew. The camera operator (UAV) goes out to collect footage. The director in the production truck (ground control station) calls the shots via radio headset (data link). Without the truck and the headset, the camera operator is just wandering around. The UAV alone is not the system — the whole chain matters.
Removing the pilot doesn't just mean removing a seat. It cascades through the entire design:
There are UAVs ranging from a 33-gram pocket drone to a 15-tonne aircraft with a 40-metre wingspan. You cannot use the same rulebook for all of them. Military organisations use classification systems to group UAVs by capability — roughly: how big are they, how high do they fly, and how far do they go? These groupings then determine what rules apply, what missions they can do, and what it costs to lose one.
NATO uses a class system based primarily on MTOW (Maximum Take-Off Weight) and operating altitude. This is the standard used across all NATO member militaries for interoperability. Fahlstrom §2.2
| Class | MTOW | Alt. Ceiling | Range | Real Example | Disposable? |
|---|---|---|---|---|---|
| I — Micro | <150 kg | <1,200 ft | <50 km | Black Hornet PRS (33 g, fits in a soldier's pocket, 25 min flight) | Yes |
| I — Mini | <150 kg | <3,000 ft | <50 km | RQ-11B Raven (1.9 kg, hand-launched, 60–90 min) | Yes |
| II — Tactical | 150–600 kg | <10,000 ft | <200 km | ScanEagle (22 kg, catapult-launched, 20+ hr endurance) | Partly |
| III — MALE | 600–5,670 kg | <FL450 | >500 km | MQ-9B SkyGuardian (5,670 kg MTOW, 40 m wingspan, 27+ hr) | No |
| III — HALE | >2,000 kg | >FL450 | Global | RQ-4B Global Hawk (14,628 kg, 35 km altitude, 32 hr) | No |
| III — Combat | Varies | Varies | Varies | MQ-28 Ghost Bat, Boeing UCAS-D | Partly |
The US military uses a five-group system that maps more directly to tactical employment — who carries it, who controls it, and from how far:
| Group | Max Weight | Altitude | Who Uses It | Typical Role |
|---|---|---|---|---|
| 1 | ≤20 lb (9 kg) | <1,200 ft AGL | Infantry squad | Look over the next hill |
| 2 | 21–55 lb | <3,500 ft AGL | Battalion HQ | Track vehicle movement |
| 3 | 55–1,320 lb | <18,000 ft | Brigade/Division | Persistent wide-area surveillance |
| 4 | >1,320 lb | <18,000 ft | Theater command | Armed ISR, strike support |
| 5 | >1,320 lb | >18,000 ft | Strategic command | Global ISR, SIGINT, UCAV |
Beyond weight, the physical shape of the aircraft determines what it can and cannot do. These are not interchangeable — each configuration is best suited to specific mission types: Sadraey §2.1
Think of it like a regular airplane. Wings generate lift as the aircraft moves forward. Highly efficient — can stay airborne for 27+ hours on a tank of fuel. But it cannot hover; it must always be moving forward or it falls. Used for long-range ISR and strike.
Lift-to-Drag ratio (L/D): 15–30. At L/D = 20, only 5% of lift force goes to fighting air resistance. Fahlstrom §3.1
Think of it like a consumer DJI drone, scaled up. Rotors push air downward to hover. Can take off and land vertically, hover in place, and fly in any direction. Highly manoeuvrable at low speed. Very power-hungry — most military multirotors have 20–60 minute endurance at most.
Hover power: 150–250 W/kg. At that rate, batteries drain fast. Sadraey §2.3
Tries to get the best of both worlds. Takes off like a multirotor, then transitions to fixed-wing cruise. Think of a drone that tilts its rotors forward after takeoff. The trade-off: mechanically complex, and the transition phase is tricky to control automatically. Increasingly used for ship-based and logistics missions. Sadraey §2.4
A drone designed to crash into its target. It circles overhead (loiters) watching for a target, then dives into it carrying an explosive warhead. Unlike a missile, it can abort, circle longer, and re-engage. Unlike a traditional armed drone, it is disposable — built to be lost. Fahlstrom §2.5
Examples: Switchblade 300 (2.7 kg, 15 min), Switchblade 600 (22.7 kg, 40 min), HERO-120 (12.5 kg, 60 min).
Every kilogram a UAV lifts off the ground has to "come from somewhere" — it's either payload, fuel, structure, electronics, or motors. The total weight has to equal the sum of all these parts. The engineering challenge is that all of these components are linked: add more payload, and you need a bigger motor, which needs more battery, which adds weight, which needs an even bigger motor. Engineers call this the "design spiral" — and it's how many UAV programs go over budget and over weight.
Sadraey formalises this dependency with the UAS weight equation: Sadraey §1.4
In plain English: Total weight ($W_0$) = payload weight + energy (fuel/battery) weight + structural weight + propulsion weight + avionics weight. The key insight is that structural, propulsion, and avionics weights are expressed as fractions of total weight ($f_s$, $f_p$, $f_{av}$) — meaning they grow automatically as the aircraft gets heavier.
Rearranging — collecting all $W_0$ terms on the left side:
$$\boxed{W_0 = \frac{W_{PL} + W_E}{1 - f_s - f_p - f_{av}}}$$Critical insight: If the denominator $(1 - f_s - f_p - f_{av})$ approaches zero, $W_0$ approaches infinity. A design where structure + propulsion + avionics fractions sum to 1.0 is physically impossible — there is no room for payload or fuel. This is the mathematical form of the design spiral.
Imagine packing for a hike. Your backpack frame weighs 20% of whatever you load in it (it's a heavy-duty frame). Your water purifier weighs 10% of total load. Your food is 15% of total load. That leaves only 55% for actual useful gear (tent, clothes, sleeping bag). If the frame + purifier + food together exceeded 100% of pack capacity, you simply cannot pack — the backpack is infeasible. This is exactly what happens when $f_s + f_p + f_{av} \geq 1$.
The Payload Mass Fraction (PMF) tells you what fraction of the aircraft's weight is actually useful. A higher PMF means a more efficient design — you're spending more of your weight budget on the mission. Fahlstrom §3.3
| Platform Class | Typical PMF | What This Means Practically |
|---|---|---|
| Group 1 Micro (<1 kg) | 0.05–0.10 | ~50–100 g payload — a tiny fixed camera, nothing more |
| Group 1 Mini (1–9 kg) | 0.10–0.18 | Up to 1.6 kg — a small EO camera, but not a stabilised thermal gimbal |
| Group 3 Tactical (25–600 kg) | 0.15–0.25 | Multi-axis EO/IR possible; no SAR radar |
| Group 4–5 MALE (MQ-9 class) | 0.25–0.35 | EO/IR + SAR + SIGINT + weapons simultaneously |
| Group 5 HALE (RQ-4 class) | 0.20–0.30 | Global persistent Multi-INT, massive SIGINT arrays |
Here is why this matters concretely: a Group 1 mini-UAV at 9 kg MTOW with PMF = 0.10 carries just 0.9 kg of payload. A stabilised thermal imaging gimbal — the sensor you'd actually want for night ISR — weighs 1.2–2.5 kg. It literally cannot fit. This is not a problem that better software solves. It is a hard physics constraint.
Let's verify the weight equation against published MQ-9B specs. We know the actual MTOW from GA-ASI's datasheet. Can the equation predict it?
| Parameter | Value | Source / How Derived |
|---|---|---|
| MTOW $W_0$ | 5,670 kg | GA-ASI published specification |
| Max payload $W_{PL}$ | 1,746 kg | GA-ASI internal + external stores spec |
| Fuel mass $W_E$ | ~1,800 kg | Estimated from fuel capacity spec |
| Structural fraction $f_s$ | ≈ 0.28 | Airframe + landing gear (estimated from similar aircraft) |
| Propulsion fraction $f_p$ | ≈ 0.085 | TPE331-10 turboprop engine ~481 kg estimated |
| Avionics fraction $f_{av}$ | ≈ 0.045 | FCC, comms, EW suite estimated |
Step 1: Calculate the denominator: $1 - 0.28 - 0.085 - 0.045 = \mathbf{0.59}$
Step 2: Apply the equation: $W_0 = (1746 + 1800)/0.59 \approx \mathbf{6{,}010\text{ kg}}$
Result: Our prediction is 6,010 kg vs. the actual 5,670 kg — an error of about 6%. For conceptual design stage (where we're estimating fractions, not measuring real hardware), that's excellent accuracy. The 6% gap comes mostly from underestimating avionics weight (satellite communications hardware, electronic warfare suite, de-icing systems).
PMF: $1746/5670 = 0.308$ — about 31% of MQ-9B's weight is usable payload. That's why it can simultaneously carry a multi-sensor EO/IR turret, SAR radar pod, and air-to-ground missiles.
ISR stands for Intelligence, Surveillance, and Reconnaissance. In plain terms: watching, listening, and gathering information about the enemy — from an aircraft that can stay up for 24+ hours without anyone getting tired. For ISR UAVs, the primary design goal is endurance (staying up as long as possible) and sensor coverage (seeing as much ground as possible). Speed matters much less.
For turboprop-powered fixed-wing UAVs like the MQ-9, endurance is calculated using the Breguet endurance equation. Before the math, here's the intuition: endurance improves when you (1) burn fuel slowly, (2) carry a lot of fuel, and (3) fly at the aerodynamically optimal speed — not too fast, not too slow. Fahlstrom §4.1
What each piece means:
$\eta_p$ = propeller efficiency — how much of engine power becomes thrust (0.82 for MQ-9).
$c_p$ = fuel burn rate per unit of power. Lower = more efficient engine.
$C_L^{3/2}/C_D$ = a measure of aerodynamic efficiency specifically for endurance
(different from the efficiency that maximises range).
$W_i, W_f$ = weight at start and end of mission. The $\sqrt{}$ terms reflect diminishing returns
as you burn fuel — the lighter aircraft becomes, the less fuel it needs to stay up.
Here is a non-obvious result: the speed that maximises how long you stay up is NOT the same as the speed that maximises how far you go. For a propeller-driven UAV:
$$V_{\text{max endurance}}^* \approx 0.76 \times V_{\text{max range}}^*$$Why? To maximise endurance, you want to minimise power consumption (power = drag × speed). To maximise range, you want to minimise fuel per kilometre (drag alone). These are different optimisation problems, solved at different speeds.
Mathematical derivation: Using the parabolic drag polar model $C_D = C_{D0} + K C_L^2$, the endurance-optimal lift coefficient is: $$C_L^{E*} = \sqrt{\frac{3C_{D0}}{K}} = \sqrt{3} \cdot C_L^{R*}$$ Since airspeed $V \propto 1/\sqrt{C_L}$, we get $V_E^* = V_R^* / 3^{1/4} \approx 0.76\, V_R^*$. Fahlstrom §4.2
Operational consequence: A ground controller who commands an ISR UAV at cruise speed instead of loiter-optimal speed is wasting roughly 4–5 hours of mission time on a 24-hour flight. That's the difference between catching an event and missing it.
Let's compute MQ-9 endurance at its operating altitude using ISA (International Standard Atmosphere) conditions. Every number here is either from the published spec sheet or standard engineering tables.
| Parameter | Value | Where It Comes From |
|---|---|---|
| Altitude | 7,620 m (25,000 ft) | GA-ASI published cruise altitude |
| Air density $\rho$ | 0.549 kg/m³ | ISA table (thinner air up high — about 45% of sea-level density) |
| Wing area $S$ | 16.84 m² | GA-ASI spec (about the size of a large living room rug) |
| Start weight $W_i$ | 4,760 kg | MTOW minus max payload (sensors stay, fuel burns) |
| End weight $W_f$ | 2,960 kg | $W_i$ minus usable fuel (1,800 kg burned) |
| Aero efficiency $(C_L^{3/2}/C_D)_{\max}$ | ≈ 12.5 | Estimated from wing aspect ratio 20, Oswald efficiency 0.85 |
| Propeller efficiency $\eta_p$ | 0.82 | TPE331-10 turboprop at cruise |
| Fuel consumption $c_p$ | $9.0\times10^{-8}$ kg/W·s | Typical turboprop BSFC |
Plugging into the Breguet equation: $$E = \frac{0.82}{9.0\times10^{-8} \times 9.807} \times 12.5 \times \sqrt{2 \times 0.549 \times 16.84} \times (\sqrt{4760} - \sqrt{2960})$$ $$= 930{,}000 \times 12.5 \times 4.30 \times (69.0 - 54.4) = 97{,}800\text{ s} \approx \mathbf{27.2\text{ hours}}$$ This matches GA-ASI's published >27-hour endurance. Our equation works.
Sensitivity check: If engine fuel consumption ($c_p$) degrades by 5% (engines get less efficient as they age), endurance drops by ~5%: $27.2 \times 0.95 \approx 25.8$ hours — losing 1.4 hours of ISR coverage per mission. This is why engine maintenance directly impacts ISR capability.
Staying airborne for 27 hours is only useful if the sensor is actually collecting intelligence. The size of the "patch" of ground the camera can see at any moment depends on how high you are, the camera's angle, and its field of view. Fahlstrom §5.1
$h$: altitude above ground; $\delta$: sensor tilt angle from horizontal; $\theta_H, \theta_V$: horizontal/vertical camera field of view; $p_{\mu m}$: camera pixel size; $f_{mm}$: focal length; $f_{ov}$: image overlap (for stitching multiple images); $V_{GS}$: ground speed. GSD tells you the resolution — at GSD = 0.25 m/pixel you can identify a car type, but not read a number plate (need ~0.05 m/pixel for that).
UAVs are not only used for watching — they are increasingly used for attacking. Two categories dominate current military thinking: loitering munitions (one drone, one precise strike) and swarms (hundreds of cheap drones acting together). Both represent fundamentally different approaches to the same problem: how do you destroy a target without risking a human life?
A loitering munition is best understood as a patient missile. A regular missile is fired and cannot be recalled. A loitering munition can patrol an area for 15–60 minutes, wait until the right target appears, and then dive into it. If the target disappears, it can abort. This patience is what makes them so operationally powerful — and legally complex. Fahlstrom §2.5
| Guidance Mode | Human Role | Latency Needed | Legal Status | Example |
|---|---|---|---|---|
| Man-in-the-loop | Human presses fire button for every shot | <500 ms link required | Compliant with IHL | Switchblade 300, 600 |
| Man-on-the-loop | System can auto-engage; human can abort | 1–5 s tolerable | Contested under IHL | Harop (IAI) |
| Fully autonomous | Human authorises the mission, not each shot | No link needed | Active legal debate | Harpy NG (IAI) |
A single sophisticated UAV can be shot down. But what if you sent 300 cheap ones simultaneously? Even if a defender destroys 100 of them, 200 still reach the target. This is the core idea behind drone swarm doctrine — saturation attack. Swarms are designed to be decentralised: no single command node controls all drones. Instead, each drone follows local rules that produce intelligent collective behaviour. Fahlstrom §2.7
Have you seen a flock of thousands of starlings moving as one fluid shape in the sky? No single bird is the leader. Each bird just follows three rules: don't crash into your neighbours, match their direction, and stay close to the group. The result looks like intelligent choreography, but it emerges entirely from simple local rules. Military drone swarms use the same principle — decentralised coordination through local rules, not a central commander. Fahlstrom §2.7
The formal version of the three flocking rules, defining forces on drone $i$ from its neighbours $\mathcal{N}$ within communication range $r$:
$$\mathbf{f}_{\text{sep}} = -\sum_{j \in \mathcal{N}_{r_s}} \frac{\mathbf{p}_j - \mathbf{p}_i}{\|\mathbf{p}_j - \mathbf{p}_i\|^2} \qquad \text{(push away from nearby drones)}$$ $$\mathbf{f}_{\text{align}} = \frac{1}{|\mathcal{N}|}\sum_{j \in \mathcal{N}} \mathbf{v}_j - \mathbf{v}_i \qquad \text{(match their velocity)}$$ $$\mathbf{f}_{\text{cohesion}} = \frac{1}{|\mathcal{N}|}\sum_{j \in \mathcal{N}} \mathbf{p}_j - \mathbf{p}_i \qquad \text{(move toward the group center)}$$ $$\mathbf{u}_i = w_s\mathbf{f}_{\text{sep}} + w_a\mathbf{f}_{\text{align}} + w_c\mathbf{f}_{\text{cohesion}} + w_m\mathbf{f}_{\text{mission}}$$Military drones add a fourth term $\mathbf{f}_{\text{mission}}$ — an attraction toward the target and repulsion from threat zones (radar, missile batteries). The weights $w_s, w_a, w_c, w_m$ are tuned during testing. Get them wrong and the swarm either collides with itself ($w_s$ too small) or fragments and loses coordination ($w_c$ too small).
Military UAVs don't fly in safe airspace. They operate where adversaries actively try to destroy them, jam their communications, or feed them false position information. Understanding these threats is not optional — they define the entire design philosophy of military autonomous systems. This section covers the three most impactful threats: GPS jamming, GPS spoofing, and radar detection.
GPS signals arrive at Earth's surface incredibly weakly — about $-130$ dBm, which is roughly 10 billion times weaker than a WiFi signal. This makes GPS trivially easy to jam. A jammer just broadcasts loud radio noise on the same frequency, drowning out the satellite signals. The UAV's GPS receiver can't hear the satellites anymore and loses its position. Fahlstrom §3.5
What it means: $J/S$ is the ratio of jammer signal to GPS satellite signal at the UAV receiver. If $J/S \geq +25$ dB, the GPS is effectively denied. $P_j, G_j$: jammer power and antenna directivity. $P_s, G_s$: GPS satellite signal power and receive gain. $R_s \approx 20{,}200$ km: how far away GPS satellites are. $R_j$: distance to the jammer.
A basic 10 W jammer ($+40$ dBm) with an omnidirectional antenna at $R_j = 30$ km achieves $J/S \approx +55$ dB — far above the denial threshold. This means a $\$50$ handheld jammer bought online can deny GPS to a military UAV from 30 km away. This is why the entire Module 04 of this course exists.
Jamming is obvious — the GPS receiver knows it's lost signal and raises an alarm. Spoofing is worse: the attacker broadcasts fake GPS signals that look real. The receiver thinks it knows exactly where it is — but the position is wrong. The UAV navigates confidently to the wrong location without any warning. This is what allegedly caused an Iranian capture of a US RQ-170 Sentinel in 2011 (though the specific method remains disputed). The defence requires cross-checking GPS against other sensors — barometers, terrain matching, inertial navigation — all topics covered in Module 04.
Ground-based radars illuminate UAVs with radio waves and detect the reflected signal. How well a UAV reflects radar is measured by its Radar Cross Section (RCS) in square metres. A conventional MALE UAV might have $\sigma \approx 1$ m² — detectable at ~200 km by a modern surveillance radar. Stealth shaping and radar-absorbent materials can reduce RCS to $0.01$ m², cutting detection range to ~63 km. Sadraey §2.6
Key insight from the 4th-root dependence: cutting RCS by 100× (20 dB) only reduces detection range by $100^{1/4} = 3.16\times$. You need enormous RCS reduction for modest range benefit — which is why stealth is difficult and expensive, and why it must be combined with other tactics (flying low, avoiding radar line-of-sight, limiting emissions).
The visualisation below shows a UAV executing a waypoint navigation mission. Yellow spheres are waypoints — pre-planned geographic coordinates the UAV must fly through, corresponding to mission phase transitions (launch, climb, ISR loiter orbit, egress, recovery). The path is a smooth curve (Catmull-Rom spline) through the waypoints, ensuring the velocity vector is always continuous — abrupt turns waste energy and stress the airframe. Vertical drop lines show the altitude profile above the ground track.
Drag with mouse to orbit the view. Scroll to zoom in/out. Use the buttons below to control speed.
The code below implements everything from sections 1.3 and 1.4 in a single runnable Python file. It's structured as a class so you can easily swap in different platform parameters (try changing the MQ-9 specs to model a ScanEagle or a Predator). All equations include comments referencing the Fahlstrom and Sadraey sections they come from.
isr_analysis.py and run
python isr_analysis.py. No special libraries needed beyond NumPy
(install with pip install numpy).
# isr_mission_analysis.py # ───────────────────────────────────────────────────────────────── # Implements: Fahlstrom & Gleason §4.1–5.3, Sadraey §1.4 # Platform modelled: MQ-9B SkyGuardian (GA-ASI) # Sensor modelled: MTS-C representative EO/IR parameters # Data link: C-band LOS ground control station # ───────────────────────────────────────────────────────────────── import numpy as np from dataclasses import dataclass # ── Platform Parameters (change these to model other UAVs) ─────── @dataclass class Platform: """ Physical and performance parameters for the UAV. All values are for MQ-9B SkyGuardian unless noted. """ name: str = "MQ-9B SkyGuardian" mtow_kg: float = 5670.0 # Maximum Take-Off Weight [kg] w_fuel_kg: float = 1800.0 # Usable fuel mass [kg] w_payload_kg: float = 1746.0 # Maximum payload mass [kg] wing_area_m2: float = 16.84 # Reference wing area [m²] cruise_alt_m: float = 7620.0 # Cruise/loiter altitude [m] (25,000 ft) ground_speed_ms: float = 55.0 # Loiter ground speed [m/s] (~107 kt) # Aerodynamic efficiency at endurance-optimal condition (Fahlstrom §4.2) cl32_cd_max: float = 12.5 # (CL^1.5 / CD)_max eta_prop: float = 0.82 # Propeller efficiency (0–1) bsfc: float = 9.0e-8 # Brake-specific fuel consumption [kg/W·s] # Weight fractions for weight budget (Sadraey §1.4) f_struct: float = 0.280 # Structural mass fraction f_prop: float = 0.085 # Propulsion system mass fraction f_avionics: float = 0.045 # Avionics mass fraction @dataclass class EO_Sensor: """EO/IR sensor parameters — representative of MTS-C turret.""" hfov_deg: float = 35.0 # Horizontal field of view [degrees] vfov_deg: float = 25.0 # Vertical field of view [degrees] depression_deg: float = 60.0 # Sensor depression from horizontal [degrees] pixel_pitch_um: float = 5.5 # Detector pixel pitch [micrometres] focal_length_mm: float = 200.0 # Focal length [mm] overlap_frac: float = 0.60 # Image overlap fraction for mosaic (0–1) @dataclass class DataLink: """C-band LOS data link parameters.""" freq_ghz: float = 5.8 # Carrier frequency [GHz] p_tx_dbm: float = 43.0 # Transmit power [dBm] → 20 W g_tx_dbi: float = 30.0 # GCS antenna gain [dBi] → 1.2 m dish at C-band g_rx_dbi: float = 6.0 # UAV antenna gain [dBi] → patch array misc_loss_db: float = 3.0 # Cable and connector losses [dB] rx_sens_dbm: float = -95.0 # Receiver sensitivity [dBm] req_margin_db: float = 10.0 # Required link margin [dB] class ISRAnalyser: """ Full ISR mission analysis. Covers: weight budget, endurance, sensor geometry, GSD, PACR, link budget. """ G = 9.80665 # gravitational acceleration [m/s²] C_LIGHT = 299_792_458.0 # speed of light [m/s] R_EARTH = 6_371_000.0 # mean Earth radius [m] def __init__(self, platform: Platform, sensor: EO_Sensor, link: DataLink): self.p = platform self.s = sensor self.l = link # Compute air density at cruise altitude using ISA troposphere model self.rho = self._isa_density(platform.cruise_alt_m) def _isa_density(self, h_m: float) -> float: """ International Standard Atmosphere density at altitude h [m]. Valid in troposphere (h < 11,000 m). Lapse rate = 6.5 K/km. """ T = 288.15 - 0.0065 * h_m # temperature at altitude [K] # density ratio from barometric formula rho = 1.225 * (T / 288.15) ** 4.256 return rho def weight_budget(self) -> dict: """ Sadraey §1.4 — closed-form MTOW from component fractions. Predicts gross take-off weight and validates against actual spec. """ p = self.p denominator = 1 - p.f_struct - p.f_prop - p.f_avionics W0_predicted = (p.w_payload_kg + p.w_fuel_kg) / denominator error_pct = abs(W0_predicted - p.mtow_kg) / p.mtow_kg * 100 PMF = p.w_payload_kg / p.mtow_kg return { 'denominator': denominator, 'W0_predicted': W0_predicted, 'W0_actual': p.mtow_kg, 'error_pct': error_pct, 'PMF': PMF, } def endurance_hours(self) -> float: """ Breguet endurance — Fahlstrom §4.1, Eq. 4.7. Maximum endurance condition: (CL^1.5 / CD)_max. Note: this differs from max-range condition (CL/CD)_max. """ p = self.p W_i = p.mtow_kg # initial weight (full fuel + full payload) W_f = p.mtow_kg - p.w_fuel_kg # final weight (fuel burned, payload stays) E_seconds = ( (p.eta_prop / (p.bsfc * self.G)) # propulsion efficiency term * p.cl32_cd_max # aerodynamic efficiency term * np.sqrt(2 * self.rho * p.wing_area_m2) # altitude term * (np.sqrt(W_i) - np.sqrt(W_f)) # fuel weight change term ) return E_seconds / 3600.0 def sensor_geometry(self) -> dict: """ Fahlstrom §5.1 — Sensor footprint, slant range, GSD, coverage rate. Assumes flat Earth (acceptable for typical UAV sensor geometry). """ s = self.s h = self.p.cruise_alt_m dep_rad = np.radians(s.depression_deg) hfov_rad = np.radians(s.hfov_deg) vfov_rad = np.radians(s.vfov_deg) R_slant = h / np.sin(dep_rad) # slant range [m] W_fp = 2 * R_slant * np.tan(hfov_rad / 2) # footprint width [m] L_fp = 2 * R_slant * np.tan(vfov_rad / 2) # footprint length [m] # Ground Sample Distance — image resolution at this range GSD_m = (s.pixel_pitch_um * 1e-6 * R_slant) / (s.focal_length_mm * 1e-3) # Persistent Area Coverage Rate (accounting for image overlap) PACR_km2ph = (W_fp * (1 - s.overlap_frac) * self.p.ground_speed_ms * 3600) / 1e6 return { 'R_slant_km': R_slant / 1000, 'W_fp_m': W_fp, 'L_fp_m': L_fp, 'GSD_cm': GSD_m * 100, 'PACR_km2ph': PACR_km2ph, } def link_budget(self) -> dict: """ Fahlstrom §5.3 — Friis transmission equation for LOS link budget. Computes maximum RF range and compares to geometric horizon. """ lk = self.l # Total available link budget [dB] allowed_path_loss = (lk.p_tx_dbm + lk.g_tx_dbi + lk.g_rx_dbi - lk.misc_loss_db - lk.rx_sens_dbm - lk.req_margin_db) # Convert path loss budget to maximum range using Friis equation wavelength = self.C_LIGHT / (lk.freq_ghz * 1e9) R_rf_m = (wavelength / (4 * np.pi)) * 10 ** (allowed_path_loss / 20) # Geometric LOS horizon at cruise altitude R_horizon_m = np.sqrt(2 * self.R_EARTH * self.p.cruise_alt_m) limiting = 'horizon-limited' if R_horizon_m < R_rf_m else 'power-limited' return { 'allowed_path_loss_dB': allowed_path_loss, 'R_rf_km': R_rf_m / 1000, 'R_horizon_km': R_horizon_m / 1000, 'limiting_factor': limiting, } def report(self): """Print a formatted analysis report.""" p = self.p wb = self.weight_budget() E = self.endurance_hours() sg = self.sensor_geometry() lb = self.link_budget() print(f"\n{'='*56}") print(f" ISR MISSION ANALYSIS — {p.name}") print(f"{'='*56}") print(f"\n Altitude: {p.cruise_alt_m/0.3048:.0f} ft | Air density: {self.rho:.4f} kg/m³") print(f"\n ── WEIGHT BUDGET (Sadraey §1.4) ──────────────────") print(f" Denominator: {wb['denominator']:.3f} (must be > 0)") print(f" Predicted W0: {wb['W0_predicted']:.0f} kg") print(f" Actual W0: {wb['W0_actual']:.0f} kg (error: {wb['error_pct']:.1f}%)") print(f" Payload frac: {wb['PMF']:.3f} ({wb['PMF']*100:.1f}% of MTOW is useful payload)") print(f"\n ── ENDURANCE (Fahlstrom §4.1) ─────────────────────") print(f" Breguet E: {E:.2f} hours on station") print(f"\n ── SENSOR GEOMETRY (Fahlstrom §5.1) ───────────────") print(f" Slant range: {sg['R_slant_km']:.2f} km") print(f" Footprint: {sg['W_fp_m']:.0f} m × {sg['L_fp_m']:.0f} m") print(f" GSD: {sg['GSD_cm']:.1f} cm/pixel (recognition needs < 45 cm)") print(f" PACR: {sg['PACR_km2ph']:.0f} km²/hour") print(f" Mission total: {sg['PACR_km2ph']*E:.0f} km² over {E:.1f} hr mission") print(f"\n ── DATA LINK (Fahlstrom §5.3) ─────────────────────") print(f" Path loss bdgt: {lb['allowed_path_loss_dB']:.0f} dB") print(f" Max RF range: {lb['R_rf_km']:.0f} km") print(f" Geom. horizon: {lb['R_horizon_km']:.0f} km") print(f" Conclusion: {lb['limiting_factor']}") print(f"\n (Both ranges ≈ equal → extending range needs relay or SATCOM)") print(f"{'='*56}\n") # ── Run the analysis ───────────────────────────────────────────── if __name__ == "__main__": analyser = ISRAnalyser(Platform(), EO_Sensor(), DataLink()) analyser.report() # ── Try it with a ScanEagle (Group 2 tactical UAV) ─────────────── print("\n--- Comparing to ScanEagle (Group 2 Tactical UAV) ---") scaneagle = Platform( name = "ScanEagle", mtow_kg = 22.0, w_fuel_kg = 3.8, w_payload_kg = 3.4, wing_area_m2 = 0.49, cruise_alt_m = 4570, # 15,000 ft ground_speed_ms = 28.0, cl32_cd_max = 9.5, eta_prop = 0.78, bsfc = 1.1e-7, # piston engine, less efficient f_struct = 0.32, f_prop = 0.09, f_avionics = 0.08, ) ISRAnalyser(scaneagle, EO_Sensor(hfov_deg=20, depression_deg=70), DataLink()).report()