Plasma-Propelled Ultra-Silence Blimp

• PUB is a low-altitude aerial robot featuring a helium envelope and a single EHD plasma thruster with a two-degree-of-freedom gimbal for precise, ultra-quiet maneuvering.
• The design combines buoyant lift with an ionic wind propulsion system to compensate drag and enable smooth indoor flight with controlled yaw and pitch.
• The system notably reduces acoustic signatures to 55–65 dB, leveraging simplified hardware and closed-loop slip control for stability in noise-sensitive environments.

Searching arXiv for the specified PUB paper and closely related plasma propulsion context. Plasma-propelled Ultra-silence Blimp (PUB) denotes a helium-lift aerial robot whose propulsion is provided solely by electrohydrodynamic (EHD) plasma thrusters, with thrust direction controlled by a two-degree-of-freedom vectoring mechanism rather than by mechanical propellers. The system reported in “PUB: A Plasma-Propelled Ultra-Quiet Blimp with Two-DOF Vector Thrusting” combines buoyant lift, a ring-type asymmetric-capacitor ionic-wind thruster, a 2‑DOF gimbal “head,” and a closed-loop slip-control architecture to realize take-off, climb, hover, descent, and landing in indoor flight experiments (Wang, 17 Aug 2025). Within aerial robotics, PUB is significant because it frames ultra-quiet blimp flight as a coupled problem in EHD plasma generation, low-speed airship dynamics, and vector-thrust control, while situating plasma propulsion as an alternative to motor–propeller actuation in noise-sensitive environments (Wang, 17 Aug 2025).

1. Definition, motivation, and system concept

PUB is a low-altitude aerial robot organized around three primary subsystems: a buoyant helium envelope, a ring-type plasma thruster that generates ionic wind, and a two-degree-of-freedom gimbal that vectors the thrust in pitch and yaw (Wang, 17 Aug 2025). In this configuration, the helium envelope supplies the majority of the lifting force, while the plasma thruster is used principally to overcome drag and to maneuver the vehicle.

The motivation stated for plasma propulsion is tied to the acoustic limitations of conventional electric UAVs. Motor–propeller architectures are dominated acoustically by rotating blades, and that noise is difficult to mitigate in urban, indoor, or reconnaissance settings. By contrast, EHD propulsion has no moving mechanical parts in the thrust-generation stage; thrust is created by an electric field in air, yielding inherently low acoustic emissions (Wang, 17 Aug 2025). The reported objective is not to maximize thrust density, but to demonstrate that ionic-wind propulsion can generate sufficient controllable thrust for stable maneuvering of a small blimp, and that a single vectoring plasma unit can replace multiple fixed thrusters while simplifying the control problem (Wang, 17 Aug 2025).

The reported in-flight sound level is 55–65 dB, compared in the paper with typical multirotor drones of similar size that often exceed 75–85 dB (Wang, 17 Aug 2025). This acoustic result is central to the designation “ultra-silence,” although the same source also attributes residual noise to corona discharge and servo motion rather than to aerodynamic blade noise (Wang, 17 Aug 2025).

2. Physical architecture and buoyancy platform

The PUB airframe is based on an ellipsoidal aluminum-film helium balloon of length $1.97\,\text{m}$ and maximum diameter $0.51\,\text{m}$ (Wang, 17 Aug 2025). Using the ellipsoid approximation,

$V = \frac{4}{3}\pi a b c \approx 268.29\,\text{L},$

the reported helium lift is about $1.11\,\text{g}/\text{L}$, giving a total buoyant lift of approximately $297.8\,\text{g}$ (Wang, 17 Aug 2025). With an envelope mass of $79.36\,\text{g}$, the net available lift for propulsion, electronics, battery, and payload is approximately $218.44\,\text{g}$ (Wang, 17 Aug 2025).

A lightweight 3D-printed gondola is rigidly attached beneath the envelope and houses the ESP32-S3 controller, high-voltage stage, IMU, RC receiver, and a 2S 800 mAh Li-ion battery of mass $42.5\,\text{g}$ (Wang, 17 Aug 2025). This architecture exploits the conventional airship principle that static buoyancy carries most of the weight, allowing the propulsion subsystem to be sized around drag compensation and control authority rather than full weight support.

The paper distinguishes two propulsion configurations. The first generation used six identical fixed plasma thrusters distributed around the envelope, each aligned with a fixed direction; full 6‑DOF motion required simultaneous modulation of all six units, which the authors describe as heavy and complex because it needed six high-voltage stages and six controllers (Wang, 17 Aug 2025). The second-generation system, which is the focus of the paper, replaces that arrangement with a single thruster mounted on a compact 2‑DOF gimbal directly under the envelope, thereby reducing mass and simplifying the control architecture (Wang, 17 Aug 2025). Future work proposes a second symmetrically mounted thruster to further decouple pitch and yaw effects (Wang, 17 Aug 2025).

3. Plasma propulsion mechanism and thruster design

The PUB propulsion unit is a four-layer concentric ring asymmetric capacitor that generates thrust through EHD ionic wind (Wang, 17 Aug 2025). The mechanical ring spacing is 9.44 mm between adjacent rings; the positive electrode is a 0.1 mm copper wire, selected to intensify the electric field and induce corona discharge, while the negative electrode is a 40 mm wide aluminum foil (Wang, 17 Aug 2025). The final wire-to-foil electrode separation is reported as 25–30 mm, with experiments indicating that approximately 30 mm is near optimal (Wang, 17 Aug 2025). The net thruster mass is $19.64\,\text{g}$ (Wang, 17 Aug 2025).

The high-voltage chain is described as DC source $\rightarrow$ pre-boost inverter $0.51\,\text{m}$0 transformer $0.51\,\text{m}$1 multi-stage voltage doubler, producing approximately 27 kV at the thruster (Wang, 17 Aug 2025). Thrust is modulated by a PWM-driven regulator acting on the low-voltage side in the range 0–6 V (Wang, 17 Aug 2025). In this architecture, control over EHD thrust is therefore indirect, via the high-voltage discharge characteristics.

The thrust model begins from a plasma–neutral continuum description in which ions are sparse, ion inertia is neglected, and momentum exchange between ions and neutrals is balanced against electric-field acceleration (Wang, 17 Aug 2025). The momentum exchange force density is written as

$0.51\,\text{m}$2

and equilibrium between electric forcing and ion–neutral collision drag is imposed through

$0.51\,\text{m}$3

Ion mobility and diffusivity are then defined as

$0.51\,\text{m}$4

The coupled field–flow model is given as a Maxwell–Navier–Stokes system: $0.51\,\text{m}$5 The thrust is represented as the volume integral

$0.51\,\text{m}$6

Under the simplifying assumptions that the neutral macroscopic velocity is negligible relative to ion drift, diffusion and ionization/recombination in the main drift region are neglected, and the system is steady-state, the current density reduces to

$0.51\,\text{m}$7

leading to the compact design relation

$0.51\,\text{m}$8

where $0.51\,\text{m}$9 points from the positive to the negative electrode (Wang, 17 Aug 2025). The reported implication is that thrust magnitude is proportional to discharge current $V = \frac{4}{3}\pi a b c \approx 268.29\,\text{L},$0, directed along the electrode separation vector, and determined in efficiency by ion mobility and geometry (Wang, 17 Aug 2025).

Experimentally, with an electrode spacing of approximately 3.0 cm, the four-ring thruster reaches a maximum thrust of $V = \frac{4}{3}\pi a b c \approx 268.29\,\text{L},$1 or $V = \frac{4}{3}\pi a b c \approx 268.29\,\text{L},$2, corresponding to a thrust-to-weight ratio of $V = \frac{4}{3}\pi a b c \approx 268.29\,\text{L},$3 for the $V = \frac{4}{3}\pi a b c \approx 268.29\,\text{L},$4 device (Wang, 17 Aug 2025). A spacing sweep using a dual-ring thruster showed thrust increasing as spacing decreased from 5.0 cm to 3.0 cm, while at 2.5 cm the discharge became intense enough to puncture the aluminum foil, establishing dielectric breakdown as a lower design bound (Wang, 17 Aug 2025). The illustrative values reported are 0.80 g at 5.0 cm, 1.16 g at 3.0 cm, and failure at 2.5 cm (Wang, 17 Aug 2025).

4. Two-DOF vector thrusting and flight dynamics

The vector-thrust mechanism is implemented by a two-stage gimbal driven by two micro RC servos, each of mass 9 g and rated torque 1.6 kg·cm; the total gimbal mass is approximately 33 g (Wang, 17 Aug 2025). The upper stage rotates in the blimp’s $V = \frac{4}{3}\pi a b c \approx 268.29\,\text{L},$5 symmetry plane and provides pitch control over $V = \frac{4}{3}\pi a b c \approx 268.29\,\text{L},$6, while the lower stage rotates about the vertical body axis $V = \frac{4}{3}\pi a b c \approx 268.29\,\text{L},$7 and provides yaw control over $V = \frac{4}{3}\pi a b c \approx 268.29\,\text{L},$8 (Wang, 17 Aug 2025).

Let the thruster be mounted at the end of a link of length $V = \frac{4}{3}\pi a b c \approx 268.29\,\text{L},$9, with installation offset $1.11\,\text{g}/\text{L}$0 from the blimp’s center of mass. The thrust vector in body coordinates is

$1.11\,\text{g}/\text{L}$1

and the corresponding moment about the center of mass is

$1.11\,\text{g}/\text{L}$2

(Wang, 17 Aug 2025). This is the formal basis for the paper’s claim that a single thruster can generate both translational forces and yaw/pitch moments.

The vehicle model employs inertial, body, and airflow frames. The body-to-ground transformation is

$1.11\,\text{g}/\text{L}$3

but for PUB the authors simplify to $1.11\,\text{g}/\text{L}$4 because the buoyancy–weight balance greatly exceeds the maximum thruster force, giving an almost self-stable envelope in pitch and roll (Wang, 17 Aug 2025). Under that simplification,

$1.11\,\text{g}/\text{L}$5

Aerodynamic forces for the symmetric ellipsoidal envelope in airflow coordinates are modeled as

$1.11\,\text{g}/\text{L}$6

with $1.11\,\text{g}/\text{L}$7, $1.11\,\text{g}/\text{L}$8, and $1.11\,\text{g}/\text{L}$9 (Wang, 17 Aug 2025). The sideslip angle $297.8\,\text{g}$0 enters the yaw dynamics explicitly through

$297.8\,\text{g}$1

which makes sideslip regulation central to lateral-directional stability (Wang, 17 Aug 2025).

The translational dynamics in the body frame are written as

$297.8\,\text{g}$2

and the rotational dynamics as

$297.8\,\text{g}$3

(Wang, 17 Aug 2025). With $297.8\,\text{g}$4, the angular-rate kinematics reduce to

$297.8\,\text{g}$5

and position dynamics become

$297.8\,\text{g}$6

(Wang, 17 Aug 2025).

5. Control architecture and closed-loop stability

The control architecture reported for PUB has two layers: an inner-loop linear state-feedback stabilizer around a low-speed trim condition, and a sliding-mode controller (SMC) for robust tracking in the presence of disturbances and nonlinearities (Wang, 17 Aug 2025). The primary control inputs are the thruster magnitude $297.8\,\text{g}$7, yaw gimbal angle $297.8\,\text{g}$8, and pitch gimbal angle $297.8\,\text{g}$9 (Wang, 17 Aug 2025).

For the linearized inner loop, the states are $79.36\,\text{g}$0, the inputs are $79.36\,\text{g}$1, and the reference trim is low-speed level flight with $79.36\,\text{g}$2 and $79.36\,\text{g}$3 (Wang, 17 Aug 2025). The linear state-space model is

$79.36\,\text{g}$4

with

$79.36\,\text{g}$5

(Wang, 17 Aug 2025).

The longitudinal speed channel is controlled by

$79.36\,\text{g}$6

and the vertical speed channel by

$79.36\,\text{g}$7

(Wang, 17 Aug 2025). The lateral/yaw-coupled channel is

$79.36\,\text{g}$8

(Wang, 17 Aug 2025). Lyapunov’s second method is then used with a positive-definite symmetric matrix $79.36\,\text{g}$9 satisfying

$218.44\,\text{g}$0

yielding

$218.44\,\text{g}$1

and therefore asymptotic stability (Wang, 17 Aug 2025). The paper interprets this as a “slip control” scheme that regulates sideslip and transverse velocity through yaw vectoring (Wang, 17 Aug 2025).

For the $218.44\,\text{g}$2 channel, the reported parameter example uses $218.44\,\text{g}$3, $218.44\,\text{g}$4, $218.44\,\text{g}$5, and $218.44\,\text{g}$6, yielding the closed-loop transfer function

$218.44\,\text{g}$7

The paper states that choosing $218.44\,\text{g}$8 eliminates steady-state error to a step input and gives good tracking in simulation (Wang, 17 Aug 2025).

The robust outer-layer controller is an SMC applied to lateral-directional dynamics: $218.44\,\text{g}$9 where $42.5\,\text{g}$0, $42.5\,\text{g}$1, $42.5\,\text{g}$2 contains added-mass terms, and $42.5\,\text{g}$3 includes aerodynamic derivatives and wind effects (Wang, 17 Aug 2025). The sliding surface is

$42.5\,\text{g}$4

and the reaching law is

$42.5\,\text{g}$5

(Wang, 17 Aug 2025). The resulting control input is

$42.5\,\text{g}$6

with Lyapunov function

$42.5\,\text{g}$7

which the paper uses to establish global stability for $42.5\,\text{g}$8 (Wang, 17 Aug 2025).

6. Experimental performance, acoustics, and operational characteristics

Flight experiments were conducted in a near-ground indoor space using marker-based video analysis. A fixed side-view camera was used for speed and altitude tracking, and a front-view camera for yaw response (Wang, 17 Aug 2025). The reported maximum speed is $42.5\,\text{g}$9, estimated from side-view frames at 3 s intervals (Wang, 17 Aug 2025). Altitude begins at approximately $19.64\,\text{g}$0 and decreases gradually to about $19.64\,\text{g}$1 after 12 s; despite this drift, the motion is described as smooth and well controlled (Wang, 17 Aug 2025). Front-view imagery shows rapid response of heading to yaw commands transmitted through RC rudder inputs and executed by the gimbal servos (Wang, 17 Aug 2025).

The paper characterizes the platform as having “full-envelope capability,” specifically take-off, climb, hover, descent, and smooth landing (Wang, 17 Aug 2025). The take-off and climb phases use buoyancy plus an upward component of vectored plasma thrust. Hover and low-speed cruise align the thrust to balance drag and preserve altitude, while descent and landing are achieved by reducing or redirecting thrust so that buoyancy and drag yield slow sinking (Wang, 17 Aug 2025). Although detailed RMS tracking errors are not reported, the authors state that the trajectories are smooth and non-oscillatory at the demonstrated scale (Wang, 17 Aug 2025).

Acoustically, the measured sound pressure level during typical indoor flight is 55–65 dB, recorded with a sound level meter placed near the vehicle during speed and yaw experiments (Wang, 17 Aug 2025). The reported sources are corona discharge, servo motion, and minimal aerodynamic noise owing to low speed and streamlined geometry (Wang, 17 Aug 2025). The absence of rotating blades eliminates blade–vortex interaction noise and tonal propeller components, which is the principal basis for PUB’s acoustic differentiation from propeller-driven UAVs (Wang, 17 Aug 2025).

The platform also has several operational attributes emphasized in the source. Structural simplicity follows from the fact that the thruster itself has no moving parts, consisting only of electrodes, supports, and HV wiring, while the gimbal uses two off-the-shelf micro servos (Wang, 17 Aug 2025). Safety derives partly from helium’s non-flammability and passive lift: if power fails, the blimp descends slowly rather than crashing violently (Wang, 17 Aug 2025). At the same time, high-voltage operation at approximately 27 kV requires careful insulation; the thruster frame is printed from insulating resin, the HV section is self-contained, and the selected electrode spacing avoids the foil puncture observed near 2.5 cm (Wang, 17 Aug 2025).

7. Applications, limitations, and broader plasma-propulsion context

The application domains identified for PUB are noise-sensitive environments, enclosed spaces, and near-space or stratospheric airship missions (Wang, 17 Aug 2025). The specific examples given are urban low-altitude corridors with strict noise constraints, silent reconnaissance and surveillance, indoor inspection, environmental monitoring, entertainment or advertising, and high-altitude buoyant platforms using ionic wind for fine maneuvering (Wang, 17 Aug 2025). A plausible implication is that PUB should be interpreted less as a high-speed transport platform than as a low-disturbance robotic sensor carrier.

Several limitations are explicitly acknowledged. The first is thrust magnitude: the current maximum ionic-wind thrust is $19.64\,\text{g}$2, so the envelope must provide nearly all lift and the achievable speed and aggressiveness are limited (Wang, 17 Aug 2025). The second is high-voltage complexity, since HV electronics add design burden and corona discharge behavior may differ in low-pressure near-space environments (Wang, 17 Aug 2025). The third is structural stiffness: the paper reports insufficient stiffness at the 2‑DOF gimbal base, introducing small yaw oscillations and drift over time (Wang, 17 Aug 2025). Proposed future improvements include adding a second symmetric thruster, optimizing ring geometry and materials to increase thrust without raising breakdown risk, improving gimbal stiffness, extending the sliding-mode control framework, and exploring cooperative multi-blimp control (Wang, 17 Aug 2025).

In broader context, PUB is situated within a lineage of EHD and plasma propulsion research rather than within conventional airship engineering alone. The source contrasts PUB with earlier ionic-wind UAVs, including a fixed-wing aircraft by Xu et al. in Nature in 2018 and a micro ionic-wind UAV by Zhang et al. in 2023, arguing that PUB extends this line by combining buoyant lift, vectored thrust with a single ring thruster, and a complete dynamics-and-control treatment (Wang, 17 Aug 2025). The comparison highlights a design shift: instead of requiring EHD thrust to sustain the aircraft aerodynamically, buoyancy reduces the propulsion requirement enough for ionic wind to become practical.

A distinct but related plasma-propulsion trajectory appears in “First Breakthrough for Future Air-Breathing Magneto-Plasma Propulsion Systems” (Goksel et al., 2016). That work reports a pulsed air-breathing magneto-plasma compressor operating at one atmosphere, with capacitor voltages of 200–600 V, energy inputs of 52–320 J/pulse, and impulse bits of 1.2–8.0 mNs, and states that at one thousand pulses per second such a system would yield thrust-to-area ratios of 50–150 kN/m$19.64\,\text{g}$3 (Goksel et al., 2016). Unlike PUB’s EHD ionic-wind mechanism, this magneto-plasma concept is based on pulsed dense plasma focus discharge rather than steady asymmetric-capacitor corona flow. The commonality is the replacement of propellers by electrically driven plasma actuation, but the operating regimes, power scales, and intended performance envelopes differ substantially.

This comparison clarifies a frequent misconception. “Plasma propulsion” in atmospheric flight does not denote a single technology class. PUB specifically uses electrohydrodynamic ionic wind in air, with no moving propulsive parts and modest thrust suited to buoyant low-speed vehicles (Wang, 17 Aug 2025). The magneto-plasma compressor literature instead addresses pulsed, high-power plasma jets, which the cited paper itself presents as a proof of concept for future air-breathing propulsion and not as an ultra-quiet blimp solution (Goksel et al., 2016). This suggests that PUB’s distinctive contribution lies not merely in using plasma, but in coupling a low-thrust EHD regime with helium buoyancy and vector-thrust control to make quiet, mechanically simple blimp flight experimentally viable (Wang, 17 Aug 2025).