Ingenuity

NASA/JPL-Caltech/ASU. Public domain (NASA / US government work).

Overview

Ingenuity performed the first powered, controlled flight on another planet on 19 April 2021 [3]. Planned for five flights over thirty sols, it completed 72 flights over nearly three years before rotor damage ended operations in January 2024.

Specifications

Parameter	Value	Source
Mass	1.8 kg	[2], [11], [15]
Weight on Mars at g = 3.71 m/s2	6.75 N	[2]
Height	approximately 0.8 m	[9]
Rotor diameter / radius	1.21 m / 0.605 m	[2], [4]
Blades	two per rotor, four total, coaxial counter-rotating	[3]
Blade mass	approximately 28 g each	[2]
Design rotor speed	2600 rpm, operating to 2800 rpm	[4], [3]
Design flight endurance	90 s per sol on a single solar recharge	[11], [6], [15]

Mission profile

Parameter	Value	Source
Deployment	from the belly of Perseverance in Jezero crater	[15]
First flight	19 April 2021	[3], [15]
Planned campaign	five flights over thirty sols	[3]
Flights completed	72	[3], [15]
Total flight time	128.8 minutes	[15]
Total distance flown	17.0 km
Maximum altitude	24.0 m
Maximum groundspeed	10 m/s
Final flight	18 January 2024
Mission declared complete	25 January 2024

The program page carries the cumulative flight record the papers do not: 128.8 minutes aloft over 17.0 km, a maximum altitude of 24.0 m and a maximum groundspeed of 10 m/s across the 72 flights, with the mission declared complete on 25 January 2024, a week after the final flight on 18 January [15].

Flight in a thin atmosphere

Quantity	Earth	Mars
Density	1.225 kg/m³	0.017 kg/m³
Reference temperature	15 C	-50 C
Dynamic viscosity	1.75e-5 Ns/m²	1.13e-5 Ns/m²
Speed of sound	340.3 m/s	233.1 m/s
Tip speed at M = 0.7	238 m/s	163 m/s

Values from [11].

Density varies over 0.010 to 0.020 kg/m³ with ground elevation and with the yearly and daily cycle [11]. Two consequences follow, and they act against each other. Low density reduces the lift per unit blade area, which forces high tip speed and low vehicle mass. The low speed of sound, itself a consequence of a CO2 atmosphere at low temperature, caps the tip speed achievable at any given design Mach number.

The resulting blade sections operate at chord Reynolds numbers of 10,000 to 25,000, against order 10⁶ for terrestrial rotorcraft [11], [14]. Ingenuity’s own blade Reynolds number at 75 percent radius is 11,412. Below approximately Re = 100,000 the boundary layer can be subcritical [11], remaining laminar across the whole angle of attack range, which reduces maximum lift coefficient and raises drag coefficient by four to five times Earth values. The optimum airfoil shape in that regime differs from high-Reynolds practice: the flown CLF5605 section is 5 percent thick, and the ROAMX successor section is 1 percent thick [13].

Hover figure of merit against thrust coefficient for the Ingenuity rotor, measured and predicted by CAMRAD II free wake (dashed), against a circular-arc plate rotor (solid), with the two sections at right. Peak figure of merit is approximately 0.63 inside a design thrust coefficient band of 0.015 to 0.020, against roughly 0.8 for a well-designed terrestrial rotor [11]. Public domain (NASA).

Rotor parameter	Value	Source
Airfoil, outboard	CLF5605, t/c 0.05, designed by AeroVironment	[2][13]
Thrust-weighted solidity	0.07391 per rotor, 0.148 for the coaxial pair	[4][12]
Tip speed at 2600 rpm	165 m/s	[2]
Tip Mach number	0.62 at the reference condition, held below 0.75	[2][3]
Design blade loading C_T/σ	0.10	[11]
Design thrust coefficient range	C_T 0.015 to 0.020	[2][11]
Peak hover figure of merit	approx 0.63, measured and calculated	[11]

Rotor performance was computed with CAMRAD II using a free wake for the coaxial pair and CFD-derived airfoil tables at 0.017 kg/m³ and -50 C, and agreed with measured performance [11]. Coaxial counter-rotation cancels torque without a tail rotor, which saves mass, at the cost of a trim problem: the two rotors must run at different collective settings, and the required difference is not monotonic in airspeed [2].

Differential collective required to trim the coaxial rotor against forward flight speed, from two independent solvers. The required difference peaks near 5 to 10 m/s [2] where the upper rotor wake sweeps across the lower disk. Including the airframe changes the answer by roughly 0.2 degrees below 15 m/s. Source: [2]. Public domain (NASA).

Parametric work since the flight indicates the flown rotor is not near the optimum for the regime. Optimal solidity for a Mars rotor is approximately 0.3 in a six-bladed configuration, with peak figure of merit at blade loadings C_T/σ of 0.1 to 0.2 [12]. Wind tunnel measurement of the ROAMX successor airfoil at M = 0.60 and Re = 20,000, over angles of attack from -2.0 to 6.0 degrees, gives 25 percent higher lift-to-drag ratio at 75 percent radius than CLF5605, and a 29 percent higher rotor figure of merit [13], [14].

Power and energy

More than half of battery capacity is consumed by overnight survival heating rather than by flight [6]. Surface temperature at the vehicle ranges from -140 C to +30 C, with equatorial night reaching -80 C. A solar panel above the rotors, with approximately 0.04 m² of exposed area, recharges between sorties [9]. The energy scaling calibrated to the flight vehicle is E_sleep = 15 W^(1/3) Wh/sol, equivalent to P_sleep = 0.518 W^(1/3) W over a 1477 minute sol at 85 percent battery efficiency, where W is gross weight [11]. Sleep energy therefore grows as the cube root of vehicle mass while flight energy grows faster, which is why survival heating dominates at the 1.8 kg scale and not at the 18 kg scale of the proposed successors.

Motor efficiency, measured in flight

The two rotor motors are the same design and were matched on a dynamometer before integration. In forward flight on flights 14 to 16 at 2700 rpm the lower motor drew 117 W from the battery against 105 W for the upper at the same shaft power, which is 74 percent total electrical efficiency including commutation electronics and harness against 83 percent [20]. A commutation offset retune on sol 297 changed power by less than 1 W, which rules out commutation and leaves the cause unresolved. Nine points of efficiency between two nominally identical motors did not appear on the dynamometer and was found only by flying them, so a power budget that assumes paired-motor symmetry is assuming something this vehicle did not have.

Surviving the load shed

Below 15 V the power electronics shut down nearly all avionics including the battery heater and the FPGA, losing the clock and every wakeup alarm [20]. That happened. Over sols 426 to 429 [20] the aircraft’s components dropped to ambient Mars nighttime ground temperature, well below their designed operational range, for three consecutive nights; the radio link was re-established at 11:44 LMST on sol 429 and the downlinked data showed no evidence of hardware degradation. It is a single unplanned survival demonstration, not a qualification. The survival heater setpoint was walked from the default -15 C down to -20, then -25, then -28 C, saving roughly 4 Wh overnight per 5 C of setpoint reduction [20]. That saving is an energy balance model prediction used for operations planning rather than a measurement, and the floor is set by battery impedance, which rises sharply below -28 C.

Compute and avionics

Commercial smartphone-class processors were used rather than radiation-hardened parts, accepting higher single-event upset rates for greater processing capability. That capability is required because the one-way Earth-Mars light time is at least five minutes, so real-time piloting is impossible and each flight executes autonomously from an uploaded plan [8].

Element	Rate or specification	Source
Inertial measurement units	two, redundant	[8]
IMU raw acceleration / angular rate	1600 Hz / 3200 Hz
IMU rate into the filter after downsampling	500 Hz
Altimeter	50 Hz
Downward navigation camera	30 Hz
State estimator	Kalman filter over IMU, altimeter and camera
Motors	eight, four per rotor
Motor allocation per rotor	one holds rotation speed, three control blade pitch
Collective pitch range	-4.5 to +17.5 deg	[3]
Cyclic pitch range	±10 deg
Servo actuator bandwidth	12 Hz
Inner-loop crossover frequency	2.5 to 3 Hz

An inclinometer calibrates IMU biases before each flight [8]. The IMU chain filters, integrates, downsamples, applies bias corrections and buffers measurements for a state propagation process that runs between filter updates.

Autonomy

Navigation is dead reckoning on the inertial measurement units, corrected by tracking ground features. The camera designates certain frames as base frames, whose detected features become pseudo-landmarks; subsequent search frames are matched against those landmarks and the resulting visual shift corrects the accumulating dead reckoning drift [8]. A new base frame is generated when the maximum number of search frames since the last base frame is reached, or when the count of successfully tracked features drops. Over 200 s of flight the estimator drifted by under 1 m in position and under 40 cm/s in velocity [7].

Flight over featureless sand deprives the navigation camera of trackable features, which degrades the velocity estimate and therefore landing accuracy [8]. A hard landing on such terrain damaged the rotor blades in January 2024 and ended the mission after 72 flights [3].

Flight speed on Ingenuity was limited by the visual navigation system rather than by aerodynamics; successor design studies assume a navigation system supporting up to 50 m/s [11].

The laser altimeter has a known failure mode that is a sensor property rather than a flight rule: it does not return a valid range over certain metallic materials, and a bad value can trip fault protection into an immediate landing directly on the object [20]. It was handled operationally, by standoff distance and route planning, not by a fix.

Operations and results

Flights progressively expanded range and altitude, and the vehicle transitioned from technology demonstration to scout, imaging terrain ahead of Perseverance. Ambient winds during the first five flights ran 2 to 6 m/s at 1.5 m altitude as measured by MEDA on the rover, which sat within about 100 m during those flights [7].

Rotor speed was raised from 2537 to 2800 rpm for flights 14 to 24, sols 241 to 398, to carry the vehicle through the low atmospheric density of southern summer, and returned to 2537 rpm as density recovered [20]. 2800 rpm was above anything tested on Earth before flight, so it was qualified on Mars: a stationary high-speed spin on sol 204 followed by a two meter checkout flight.

Dust cost the vehicle both power and mechanism margin. During the sols 310 to 320 dust storm the Perseverance SkyCam measured a peak optical depth of 1.9, and Ingenuity’s solar array charge current at about 11:30 LTST fell from 296 mA on sol 302 to 221 mA on sol 316 before recovering to 302 mA, a 25 percent loss of charging current [20]. On sol 335 all six swashplate actuators failed their pre-flight self-test on an unallowable actuator load and flight 19 did not take off. A successful flight appeared to clear the mechanism, and the same failure recurred 25 sols later on sol 370 with no preceding dust event [20]. The diagnosis is from integrated motor currents, not from direct observation of the actuators.

Attitude telemetry from the flights was subsequently used as a wind instrument. A steady-state model converting yaw, pitch and roll into a wind vector, on the basis that the rotorcraft must tilt further into a stronger headwind, produced the first Martian wind measurements between 3 m and 24 m altitude [7]. For comparison, MEDA measured 3.2 ± 2.3 m/s at 1.5 m in northern spring and summer with 99 percent of values below 10 m/s, an afternoon peak of 6.1 ± 2.2 m/s, and ten-minute standard deviations of 0.57 ± 0.29 m/s at night against 1.85 ± 0.57 m/s by day. MEDA wind accuracy is 1 m/s below 10 m/s, 10 percent above, and 15 degrees in direction [7].

Electrostatic environment

Rotor blades charge triboelectrically against suspended grains, and the CO2 atmosphere at approximately 5 Torr breaks down at approximately 30 kV/m against approximately 3 MV/m at Earth sea level [9]. Townsend dark discharge begins above 20 kV/m and spark discharge near 50 kV/m, with a primary ionization coefficient of order 1e4 /m, a Townsend parameter of approximately 285 kV/m, and a secondary-to-primary ionization ratio of 0.01 to 0.02. At 2800 rpm the blades sweep approximately 290 revolutions per second through the near field [9]. In an ambient dust load, current balance and charge dissipation back into the atmosphere proceed without breakdown; inside a rotor-created particulate cloud the tribocharging current can be large enough to stimulate local breakdown near the vehicle.

Successor designs

Concept	Gross weight	Rotor radius	Solidity	Rotor speed	Disk loading	FM	Source
Ingenuity	1.8 kg	0.605 m	0.148 (coaxial pair)	2600 rpm		approx 0.63	[11][4]
MSH coaxial	18.032 kg	1.25 m	0.310	1247 rpm	3.67 kg/m²	0.675	[11]
MSH hexacopter	17.662 kg	0.64 m (six rotors)	0.193	2435 rpm	2.29 kg/m²	0.615
Sample Recovery Helicopter		0.700 m	0.06412 per rotor	2350 rpm			[4]

The Mars Science Helicopter design mission is 30 s of takeoff at hover power, a climb to 200 m, a 1 km cruise to the science site, a 2 minute hover, landing, one sol of sleep and recharge, with the battery sized to that energy plus a 20 percent reserve [11]. Payload is 2.02 kg at 35 W of equipment power, with 1.2 kg of avionics including 0.1 kg of telecom. Design blade loading is C_T/σ = 0.11 [11], 10 percent above Ingenuity’s, justified by the measured rotor thrust capability and the expected gain from a circular-arc airfoil. Hover tip Mach number is 0.7 for a tip speed of 163.2 m/s, and cruise is 30 m/s, the minimum power speed, giving an advance ratio of 0.18 and an advancing tip Mach number of 0.83 [11]. Blade Reynolds numbers at 75 percent radius rise to 24,752 for the coaxial and 15,784 for the hexacopter, against 11,412 for Ingenuity. Aeroelastic stability for these rotors has been swept across densities of 0.013 and 0.020 kg/m³ and tip Mach numbers of 0.4 to 0.9 [4].

Technologies developed

Ingenuity established that a rotor sized to a 0.017 kg/m³ atmosphere and a Reynolds number of order 10⁴ can be flown, controlled and navigated without ground intervention, and produced flight data against which CAMRAD II free-wake predictions and airfoil tables were validated [11], [2]. In-flight system identification over 0.15 to 8 Hz gave a control model derived from the vehicle itself rather than from ground testing [3].

Its use of commercial smartphone-class processors moved a class of avionics from prohibited to demonstrated. Its attitude telemetry became an atmospheric instrument, delivering the first wind profiles between 3 m and 24 m on Mars [7]. And its measured rotor performance is the baseline against which the ROAMX airfoil and blade work claims a 29 percent figure of merit improvement, which is the difference between a technology demonstration and a vehicle with useful range [14].