Self-Powered Vibration Control System

• Self-powered vibration control systems are integrated architectures that convert vibration-induced mechanical energy into electrical energy to autonomously drive damping, switching, and sensing functions.
• They employ varied transduction methods, including piezoelectric and electromagnetic techniques, to realize both passive and semi-active damping along with energy harvesting.
• Experimental validations demonstrate notable improvements in vibration attenuation and energy recovery while addressing challenges like parasitic losses and intermittent power supply.

A self-powered vibration control system is a control system in which the only energy required for operation is absorbed from the plant, typically by converting vibration-induced mechanical energy into electrical energy, storing or conditioning that energy, and using it to sustain damping, switching, sensing, or semi-active control functions. Within this broad category, the literature includes semi-passive piezoelectric shunt systems, passive electrically interconnected piezoelectric networks, regenerative electromagnetic absorbers for vehicle suspensions, and electromechanical metamaterial nodes that combine attenuation with autonomous sensing. A recurrent theme is that energy autonomy is not merely a power-supply issue: in strict formulations, feasibility depends on colocation, passivity, and explicit accommodation of parasitic losses in the transducers, storage elements, and power electronics (Ligeikis et al., 2022, Shen et al., 2010, Zhao et al., 3 Jun 2025, Chen et al., 2020, Hajidavalloo et al., 2021, Ligeikis et al., 14 Sep 2025).

1. Conceptual definition and scope

In the strict sense used in linear feedback theory, a control system is self-powered if the only energy it requires for operation is that which it absorbs from the plant. For a linear feedback law to be feasible, the feedback signal must be colocated with the control inputs, and its input-output mapping must satisfy an associated passivity constraint. The imposed feedback law can be viewed equivalently as the imposition of a linear passive shunt admittance at the actuation ports of the plant; however, when actively controlled electronics are used, passivity alone is insufficient because parasitic losses must also be overcome (Ligeikis et al., 2022).

This energetic definition covers both passive and semi-active realizations. In piezoelectric semi-passive damping, the structure’s vibration energy powers a switching circuit that intermittently changes the electrical boundary condition of a piezoelectric element, thereby generating dissipation through synchronized switching rather than through a continuously driven external actuator (Shen et al., 2010). In passive plate control, uniformly distributed PZT actuators interconnected through an electric network convert mechanical vibration energy into electrical energy and dissipate it via resistive elements without requiring external power supplies for actuation (Vidoli et al., 2010). In vehicle suspensions, self-powered operation is framed through simultaneous vibration suppression and energy harvesting, with harvested energy used to modulate damping characteristics or to power the controller and associated electronics (Chen et al., 2020, Hajidavalloo et al., 2021).

A persistent ambiguity in the literature concerns the term “self-powered.” The shape-memory two-bar truss controlled by fuzzy feedback linearization is explicitly presented as “self-adjustable” and “autonomous” in the sense of biological self-regulation, but the same source distinguishes this from self-powered vibration control in the strict energy sense, which involves energy harvesters and local actuation (Herôncio et al., 2022). This distinction is central when comparing smart-structure control, energy harvesting, and true energy-autonomous vibration mitigation.

2. Architectures and constituent subsystems

Across implementations, self-powered vibration control systems couple a vibratory plant to an energy transduction stage, a power-conditioning or storage stage, and a control or damping interface. The transduction stage is often piezoelectric or electromagnetic. Piezoelectric elements convert strain into charge; electromagnetic devices use relative motion to generate electrical power; some battery-free vibration-powered platforms also support triboelectric harvesters, although their primary focus is autonomous sensing rather than direct vibration attenuation (Shen et al., 2010, Li et al., 7 Jul 2025, Chen et al., 2020, Ligeikis et al., 14 Sep 2025).

The self-powered SSDV beam system provides a canonical piezoelectric architecture. It contains a piezoelectric energy harvesting subsystem and a switching control circuit. One piezo patch harvests vibrational energy, a rectification and storage stage accumulates it, and a threshold controller ensures that the control circuit activates only when sufficient charge is available. A second piezo patch acts as actuator, while another piezo patch is used as a sensor for extremum detection. At each displacement extremum, a low-loss switch briefly connects the actuator piezo, an inductor, and voltage sources to form a resonant LC circuit with voltage boost; the entire control electronics, including positive and negative voltage rails, are powered by the harvested energy (Shen et al., 2010).

Distributed plate control uses a different architecture. Vidoli and dell’Isola describe uniformly distributed PZT actuators embedded in a plate and electrically interconnected through a passive two-dimensional network. Each actuator plays a two-fold role: capacitive element in the electric network and mechanical couple supplier. The network can include inductances, resistances, and the actuators’ own capacitances, enabling coupled electro-mechanical waves and passive dissipation over a broad area of the structure (Vidoli et al., 2010).

Recent systems broaden the architectural scope. The EMetaNode integrates a host beam with periodically attached local resonators, piezoelectric patches, synchronized switching interface circuits such as SECE, SSHI, and S3BF, an energy management circuit, and a compact IoT board with Bluetooth and sensors. The result is simultaneous broadband vibration attenuation and self-powered sensing in a single electromechanical metamaterial node (Zhao et al., 3 Jun 2025). At larger scale, a prototype self-powered system uses a linear ballscrew coupled with a permanent-magnet synchronous machine, a custom three-phase inverter, and a custom half-bridge DC-DC power converter connected to a storage capacitor, with hardware-in-the-loop validation for a stochastically-excited tuned vibration absorber (Ligeikis et al., 14 Sep 2025).

System class	Core hardware	Reported function
Piezoelectric synchronized switching	Piezo patches, rectifier, storage capacitor, threshold controller, inductor, voltage source	Semi-passive damping and self-powered switching
Distributed PZT network	Uniformly distributed PZT actuators and passive 2D electric network	Passive plate vibration control through electro-mechanical energy exchange
Electromagnetic regenerative absorber	Ball-screw, rotary generator, variable electrical load	Simultaneous vibration suppression and energy harvesting
Electromechanical metamaterial node	Local resonators, piezo patches, synchronized switching circuits, IoT board	Broadband attenuation and self-powered sensing

Battery-free vibration-powered platforms highlight the importance of power-management logic under intermittent excitation. ViPSN 2.0 formalizes capacitor-voltage-driven operation into Cold Start, Energy Build-Up, Task Operation, Checkpoint, and Shutdown phases, with UVLO, PID, and APC power-management solutions (Li et al., 7 Jul 2025). A plausible implication is that similar energy-state partitioning is increasingly relevant for self-powered control loops whenever harvested power is intermittent or strongly excitation-dependent.

3. Attenuation mechanisms and electromechanical models

The defining mechanism in piezoelectric synchronized switching is phase manipulation between mechanical strain and piezoelectric voltage. In SSD and SSDV, the piezoelectric element is intermittently switched from open-circuit to a prescribed impedance synchronously with structural vibration. Because switching is timed to displacement extrema, a phase difference appears between the strain induced by vibration and the resulting voltage, which creates energy dissipation. In SSDV, a voltage source is added to the resonant inversion path, boosting the inverted voltage and increasing the dissipated mechanical energy relative to SSDI, particularly in weakly coupled structures (Shen et al., 2010).

In the resonant switching interval, the piezoelectric branch obeys the LC relation

$Lq_0'' + rq_0' + \frac{1}{C_0}q_0 = 0,$

and the piezoelectric voltage evolution is written as

$V(t) = \alpha \left( \dot{u}(t) + h(t) \right).$

The energy dissipated per vibration cycle is

$$E_S = \frac{\alpha^2}{C_0} \int_0^T h(t)\dot{u}(t)\,dt,$$

while the power consumed in SSDV per period is

$P_{VS} = \frac{2 V_S C_0 (1 + \alpha_i) V_M}{T}.$

These expressions formalize the central trade-off: the added voltage source increases damping but also introduces an additional power requirement (Shen et al., 2010).

In distributed PZT plate systems, the essential mechanism is electro-mechanical wave coupling under a self-resonance criterion. The homogenized continuum model, valid when the wave lengths are large relative to the actuator dimension, uses the Kirchhoff-Love theory for plate bending and a scalar electric potential. In reduced dimensionless form, the coupled equations are

$$\begin{cases} a_4 \Delta^2 v + \ddot{v} - \gamma \Delta \phi = 0 \ - \beta \Delta \phi + \ddot{\phi} + \delta \dot{\phi} + \gamma \Delta \dot{v} = 0 \end{cases}$$

with self-resonance achieved when the natural frequency of a mechanical mode coincides with that of an electrical mode it couples to. For simply supported plates, mode coupling is one-to-one; for clamped plates, the coupling matrix is quasi-diagonal (Vidoli et al., 2010).

In electromechanical metamaterials, synchronized switching is interpreted as nonlinear electromechanical friction. For the $j$-th local resonator,

$$m_r \ddot{u}_r + c_r \dot{u}_r + k_r u_r + 0.5 \cdot f_n(u_r, \dot{u}_r) = -m_r (\ddot{w} + \ddot{w}_b),$$

with reaction force

$f_n(t) = \alpha_p v_p(t),$

and a generalized form

$$f_n(t) = Z_u (1-V_M) \cdot \text{sign}(\dot{u}_r) + K u_r.$$

For harmonic-balance computation, a continuously differentiable approximation is introduced:

$$f_n(t) = Z_u (1 - V_M) \tanh(\beta \dot{u}_r) + K u_r.$$

This Coulomb-like friction term broadens the attenuation bandgap and enables attenuation associated with higher harmonics (Zhao et al., 3 Jun 2025).

Vehicle suspension absorbers exploit a different mechanism: mechanical vibration is converted to rotary motion by a ball-screw, nonlinear inertia is introduced either by an adjustable rotational inertia or by a pendulum absorber, and an electromagnetic generator extracts power while providing tunable electrical damping. In the IPVA, harvested electrical power is

$V(t) = \alpha \left( \dot{u}(t) + h(t) \right).$0

and the control objective combines ride comfort and power recovery through

$V(t) = \alpha \left( \dot{u}(t) + h(t) \right).$1

The ERVA uses an analogous economic cost balancing chassis acceleration and harvested power under preview-based NMPC (Hajidavalloo et al., 2021, Chen et al., 2020).

4. Feasibility, passivity, and control synthesis

Theoretical work on self-powered linear feedback control places energetic feasibility at the center of controller design. For colocated LTI systems with input $V(t) = \alpha \left( \dot{u}(t) + h(t) \right).$2 and output $V(t) = \alpha \left( \dot{u}(t) + h(t) \right).$3, a desired feedback law can be written as $V(t) = \alpha \left( \dot{u}(t) + h(t) \right).$4, where $V(t) = \alpha \left( \dot{u}(t) + h(t) \right).$5 is an admittance. In ideal hardware, the self-powered requirement reduces to passivity. With losses, the constraint becomes

$V(t) = \alpha \left( \dot{u}(t) + h(t) \right).$6

which explicitly accounts for actuation losses. In the finite-dimensional, time-invariant case, sufficient feasibility conditions reduce to a more conservative version of the Positive Real Lemma parameterized by loss parameters such as actuator resistance and storage leakage or charge-rate constants (Ligeikis et al., 2022).

This loss-aware perspective directly informs hardware design. The large-scale self-powered prototype models parasitic dissipation as

$V(t) = \alpha \left( \dot{u}(t) + h(t) \right).$7

with stored capacitor energy

$V(t) = \alpha \left( \dot{u}(t) + h(t) \right).$8

The energy dynamic equation is

$V(t) = \alpha \left( \dot{u}(t) + h(t) \right).$9

subject to the feasibility condition

$$E_S = \frac{\alpha^2}{C_0} \int_0^T h(t)\dot{u}(t)\,dt,$$0

The same work emphasizes that feasibility is tighter than passivity because converter resistance, capacitor leakage, and related parasitics must be covered before any net actuation energy is available (Ligeikis et al., 14 Sep 2025).

Semi-active suspension systems demonstrate how advanced control synthesis is adapted to this energetic setting. The ERVA uses NMPC with road-profile preview to optimize either ride comfort, energy harvesting, or both (Chen et al., 2020). The IPVA extends this line by introducing a stochastic linearization method with guaranteed stabilizability and embedding it within SL-MPC, producing a controller that is much more computationally efficient than the nonlinear MPC counterpart with no major performance degradation (Hajidavalloo et al., 2021). These approaches are not framed through passive shunt synthesis, but they preserve the central idea that electrical damping is modulated by energy extracted from vibration.

5. Experimental realizations and quantitative performance

Experimental results show that self-powered vibration control has been demonstrated across scales and application classes, but with markedly different performance envelopes.

The self-powered SSDV beam system validates the semi-passive piezoelectric concept in single-mode resonant damping of a composite beam. With one actuator, SSDI produces a damping effect of approximately $$E_S = \frac{\alpha^2}{C_0} \int_0^T h(t)\dot{u}(t)\,dt,$$1 dB. SSDV with $$E_S = \frac{\alpha^2}{C_0} \int_0^T h(t)\dot{u}(t)\,dt,$$2 V increases the damping effect to approximately $$E_S = \frac{\alpha^2}{C_0} \int_0^T h(t)\dot{u}(t)\,dt,$$3 dB, and SSDV with higher $$E_S = \frac{\alpha^2}{C_0} \int_0^T h(t)\dot{u}(t)\,dt,$$4 up to $$E_S = \frac{\alpha^2}{C_0} \int_0^T h(t)\dot{u}(t)\,dt,$$5 V reaches approximately $$E_S = \frac{\alpha^2}{C_0} \int_0^T h(t)\dot{u}(t)\,dt,$$6 dB. The pulse-charger circuit harvested $$E_S = \frac{\alpha^2}{C_0} \int_0^T h(t)\dot{u}(t)\,dt,$$7–$$E_S = \frac{\alpha^2}{C_0} \int_0^T h(t)\dot{u}(t)\,dt,$$8 of the available piezoelectric energy depending on excitation level. At $$E_S = \frac{\alpha^2}{C_0} \int_0^T h(t)\dot{u}(t)\,dt,$$9 V, the control circuit consumed about $P_{VS} = \frac{2 V_S C_0 (1 + \alpha_i) V_M}{T}.$0W, which could be supported continuously with four piezo harvesters (Shen et al., 2010).

The EMetaNode establishes a distinct operating regime: broadband attenuation coupled to self-powered wireless sensing. Experiments show widened bandgaps and suppressed resonance peaks when SECE, SSHI, or S3BF circuits are added, and confirm third-harmonic generation with its associated attenuation band around $P_{VS} = \frac{2 V_S C_0 (1 + \alpha_i) V_M}{T}.$1. The harvested energy powers a compact IoT board that supports Bluetooth-based reporting of temperature and acceleration, with the energy management circuit accumulating and releasing approximately $P_{VS} = \frac{2 V_S C_0 (1 + \alpha_i) V_M}{T}.$2 V and a sensing/broadcast sequence consuming approximately $P_{VS} = \frac{2 V_S C_0 (1 + \alpha_i) V_M}{T}.$3C (Zhao et al., 3 Jun 2025).

In suspension systems, regenerative absorbers show substantially larger power levels. For the ERVA, NMPC with road preview achieves a $P_{VS} = \frac{2 V_S C_0 (1 + \alpha_i) V_M}{T}.$4 reduction in the 2-norm of chassis acceleration relative to NMPC without preview in the ride-comfort case. In the energy-harvesting case, average harvested power is $P_{VS} = \frac{2 V_S C_0 (1 + \alpha_i) V_M}{T}.$5 W with preview, $P_{VS} = \frac{2 V_S C_0 (1 + \alpha_i) V_M}{T}.$6 W without preview, and $P_{VS} = \frac{2 V_S C_0 (1 + \alpha_i) V_M}{T}.$7 W for passive ERVA. In the dual-objective case, preview provides a $P_{VS} = \frac{2 V_S C_0 (1 + \alpha_i) V_M}{T}.$8 increase in harvested power and a $P_{VS} = \frac{2 V_S C_0 (1 + \alpha_i) V_M}{T}.$9 decrease in vibration compared to the no-preview case; the average harvested power levels are stated to be sufficient to power the suspension controller and associated electronics (Chen et al., 2020).

The IPVA further reports up to $$\begin{cases} a_4 \Delta^2 v + \ddot{v} - \gamma \Delta \phi = 0 \ - \beta \Delta \phi + \ddot{\phi} + \delta \dot{\phi} + \gamma \Delta \dot{v} = 0 \end{cases}$$0 increase in harvested power and $$\begin{cases} a_4 \Delta^2 v + \ddot{v} - \gamma \Delta \phi = 0 \ - \beta \Delta \phi + \ddot{\phi} + \delta \dot{\phi} + \gamma \Delta \dot{v} = 0 \end{cases}$$1 reduction in RMS acceleration compared to optimal linear designs in passive comparisons. With perfect road preview, NMPC and SL-MPC increase harvested power by $$\begin{cases} a_4 \Delta^2 v + \ddot{v} - \gamma \Delta \phi = 0 \ - \beta \Delta \phi + \ddot{\phi} + \delta \dot{\phi} + \gamma \Delta \dot{v} = 0 \end{cases}$$2 and $$\begin{cases} a_4 \Delta^2 v + \ddot{v} - \gamma \Delta \phi = 0 \ - \beta \Delta \phi + \ddot{\phi} + \delta \dot{\phi} + \gamma \Delta \dot{v} = 0 \end{cases}$$3, respectively, over passive operation, while RMS acceleration is reduced by up to $$\begin{cases} a_4 \Delta^2 v + \ddot{v} - \gamma \Delta \phi = 0 \ - \beta \Delta \phi + \ddot{\phi} + \delta \dot{\phi} + \gamma \Delta \dot{v} = 0 \end{cases}$$4 for NMPC and $$\begin{cases} a_4 \Delta^2 v + \ddot{v} - \gamma \Delta \phi = 0 \ - \beta \Delta \phi + \ddot{\phi} + \delta \dot{\phi} + \gamma \Delta \dot{v} = 0 \end{cases}$$5 for SL-MPC. For a mixed objective, the reported improvement is a $$\begin{cases} a_4 \Delta^2 v + \ddot{v} - \gamma \Delta \phi = 0 \ - \beta \Delta \phi + \ddot{\phi} + \delta \dot{\phi} + \gamma \Delta \dot{v} = 0 \end{cases}$$6 increase in harvested power and a $$\begin{cases} a_4 \Delta^2 v + \ddot{v} - \gamma \Delta \phi = 0 \ - \beta \Delta \phi + \ddot{\phi} + \delta \dot{\phi} + \gamma \Delta \dot{v} = 0 \end{cases}$$7 reduction in acceleration. SL-MPC is over $$\begin{cases} a_4 \Delta^2 v + \ddot{v} - \gamma \Delta \phi = 0 \ - \beta \Delta \phi + \ddot{\phi} + \delta \dot{\phi} + \gamma \Delta \dot{v} = 0 \end{cases}$$8 faster than NMPC, and the IPVA is reported to deliver $$\begin{cases} a_4 \Delta^2 v + \ddot{v} - \gamma \Delta \phi = 0 \ - \beta \Delta \phi + \ddot{\phi} + \delta \dot{\phi} + \gamma \Delta \dot{v} = 0 \end{cases}$$9–$j$0 W per damper under Class C road excitation at $j$1 km/h, with up to $j$2 mechanical efficiency (Hajidavalloo et al., 2021).

At larger scale, hardware-in-the-loop experiments on a tuned vibration absorber validate a prototype with power flows above $j$3 W and forces on the order of $j$4 kN. The LTI self-powered synthetic admittance outperformed optimal static damping by approximately $j$5–$j$6 in the mean-square vibration measure $j$7, while the nonlinear performance-guaranteed controller improved further by more than $j$8. The nonlinear controller also reduced mean-square absolute acceleration of both main and absorber masses by nearly $j$9 (Ligeikis et al., 14 Sep 2025).

6. Applications, limitations, and research directions

The application space spans structural vibration mitigation, smart infrastructure, autonomous sensing, and vehicle suspensions. The EMetaNode is explicitly positioned for structural health monitoring and Internet of Things applications, with self-powered sensing of temperature and acceleration and a stated practical path to digitalizing structures and systems for autonomous sensing and vibration control (Zhao et al., 3 Jun 2025). The self-powered weigh-in-motion concept combines vibration energy harvesting with self-sensing composite pavements and a low-power data acquisition system for traffic load identification, showing laboratory feasibility for self-sustainable infrastructure sensing (Birgin et al., 2023). Vehicle-oriented systems target simultaneous ride comfort improvement and energy recovery in suspensions (Chen et al., 2020, Hajidavalloo et al., 2021).

Several limitations recur. First, increased damping performance often requires increased power consumption. In SSDV, the addition of the voltage source improves damping relative to SSDI but raises circuit energy consumption, requiring careful scaling of the harvester and threshold-based energy management (Shen et al., 2010). Second, passive distributed PZT networks are strongly mode-dependent: maximal effectiveness requires tuning to the mode or narrow band of interest, and the homogenized model is valid only for long wavelengths relative to actuator size (Vidoli et al., 2010). Third, intermittent and unstable ambient vibrations complicate battery-free operation, motivating explicit energy-aware state management as in ViPSN 2.0 (Li et al., 7 Jul 2025).

A further limitation is feasibility under realistic losses. Self-powered linear feedback cannot be established from passivity alone once parasitic losses are included; colocation and stricter, loss-dependent feasibility conditions are necessary (Ligeikis et al., 2022). The large-scale prototype reinforces the same point experimentally, showing that hardware parasitics must be built into the control synthesis rather than treated as secondary implementation details (Ligeikis et al., 14 Sep 2025).

Infrastructure-scale harvesting also remains challenging. In the self-powered weigh-in-motion system, the harvested power of up to $$m_r \ddot{u}_r + c_r \dot{u}_r + k_r u_r + 0.5 \cdot f_n(u_r, \dot{u}_r) = -m_r (\ddot{w} + \ddot{w}_b),$$0 mW per event or vehicle passage is explicitly stated to be insufficient for full autonomy, leading to the conclusion that dense EH networks or combinations of different EH technologies are needed for complete self-sustainability (Birgin et al., 2023). This suggests that, for many deployments, the central research problem is no longer the isolated design of a damper or harvester, but the co-design of transduction, energy management, control law, and application duty cycle.

Taken together, the literature presents self-powered vibration control not as a single method but as a systems paradigm. Its most mature formulations combine electromechanical transduction, loss-aware power electronics, storage-aware control synthesis, and, increasingly, joint vibration attenuation and autonomous sensing. The main research trajectory indicated by recent work is toward broader bandwidth, higher power scale, and tighter integration between control and energy-management layers (Zhao et al., 3 Jun 2025, Hajidavalloo et al., 2021, Ligeikis et al., 14 Sep 2025).