Friction with Bristle Dynamics: A Unified Model

Updated 27 January 2026

FrBD is a modeling paradigm that uses explicit bristle-like elements to represent frictional interfaces, linking microscale slip and macroscale behavior.
It unifies various friction models, from Dahl to LuGre, and is validated in applications such as robotics, tire–road interactions, and vibration-driven locomotion.
The framework ensures well-posedness and passivity through rigorous mathematical formulations, facilitating robust control design and observer synthesis.

Friction with Bristle Dynamics (FrBD) is a modeling paradigm that mechanistically describes frictional interfaces using explicit bristle-like rheological elements, capturing the emergent dynamics of stick–slip, directionality, and rate-dependent friction under both static and dynamic excitations. FrBD models unify and generalize a spectrum of classical and modern friction laws—from Prandtl-Tomlinson and Dahl to LuGre and full distributed-contact brush-type models—within a physically and mathematically rigorous framework. Bristle dynamics underpin the transition between micro-scale interfacial phenomena and macro-scale frictional behavior, making FrBD foundational in robotics, tire–road interaction, tribology, and the design of bidirectional vibration-driven robots, as well as in the control of mechatronic and rolling systems (Cicconofri et al., 2017, Cicconofri et al., 2014, Gidoni et al., 2016, Romano et al., 11 Jan 2026, Romano, 11 Jan 2026, Romano et al., 20 Jan 2026, Romano, 20 Jan 2026, Kim et al., 2020, Goicochea et al., 2016).

1. Foundational Principles and Model Hierarchy

FrBD models are built upon the explicit coupling of bristle deformation and interfacial slipping. A bristle, modeled as a microscopic spring-damper unit or a distributed field, develops internal restoring forces when subject to tangential displacement of the contacting body. The key variables are the bristle deflection $z(t)$ (scalar or vector) and the rigid-body relative slip $v(t)$ , with the local tip velocity $v_s = v - \dot{z}$ .

The bristle constitutive law can be realized as:

Linear elastic (Dahl model): $f = k z$ , with $F_\text{fric} = f$ .
Kelvin–Voigt (brush model): $f = k z + c \dot{z}$ , leading to viscoelastic dry friction.
Generalized Maxwell/Kelvin–Voigt: Combination of multiple relaxation branches for complex rheological behavior (Romano et al., 20 Jan 2026, Romano, 20 Jan 2026).
Nonlinear friction law: $F_\text{fric} = \mu(v_s) \frac{v_s}{|v_s|_\varepsilon}$ , with Stribeck, pre-sliding, and velocity-dependent effects.

This architecture generalizes single-DOF models to distributed and multi-dimensional variants, enabling the modeling of rolling/sliding contact, as in rolling tires and bristle arrays.

2. Mathematical Formulations: Lumped, Distributed, and Multi-Dimensional FrBD

Lumped (1D) ODE Models

Canonical FrBD systems couple the bristle rheology with the friction law through the force balance at the interface. For generalized Maxwell (GM) and Kelvin–Voigt (GKV) realizations (Romano et al., 20 Jan 2026):

GM-branch (with $n$ Maxwell arms):

$\begin{cases} f = \bar{k}_0 z + \sum_{i=1}^n f_i \ \dot{f}_i = -\tau_i^{-1} f_i + \bar{k}_i \dot{z} \ \dot{z} = -\frac{|v|_\varepsilon}{\mu(v)} f + v \end{cases}$

GKV-branch: Parallel arrangement of springs and dashpots with similar closure.

Distributed PDE Models

For spatially distributed contacts (e.g., tires), bristle deflection becomes a field $z(\xi,t)$ , and the governing evolution is a first-order PDE with transport and internal loss (Romano et al., 11 Jan 2026, Romano, 11 Jan 2026, Romano, 20 Jan 2026): $\frac{\partial z(\xi,t)}{\partial t} + V \frac{\partial z}{\partial \xi} = -\frac{\sigma_0 |v(t)|_\varepsilon}{g(v)} z(\xi,t) + \frac{\mu(v)}{g(v)} v(t)$ where $V$ is the rolling/transport velocity, $\sigma_0$ is micro-stiffness, and $g(v)$ encompasses both viscous and friction contributions.

Two-dimensional extensions represent the bristle state as $\bm{z}(\bm{x},s)$ over the contact patch. The local friction law and bristle evolution form a semilinear hyperbolic PDE system, incorporating longitudinal, lateral, and spin slip (Romano, 11 Jan 2026, Romano, 20 Jan 2026).

3. Physical Mechanisms: Stick–Slip, Directionality, and Inversion

Bristle dynamics explain essential frictional phenomena:

Stick–slip and pre-sliding: Bristle deflection accumulates elastic energy until a slip threshold is reached, followed by rapid release.
Friction inversion: In vibration-driven systems, such as bristle bots, the combined phase relation between normal force modulation and bristle slip reverses the direction of net locomotion when the actuation frequency crosses a critical value,

$\omega_c = \sqrt{ \frac{k}{M L^2 \cos^2 \alpha} }$

as experimentally confirmed in (Cicconofri et al., 2017, Cicconofri et al., 2014, Kim et al., 2020).

Directional friction: Bristle interaction with microscale surface asperities gives rise to an emergent “with-the-nap/against-the-nap” asymmetry. The macroscopic friction is the product of a geometric factor (dependent on bristle angle and surface profile slope) and an energetic factor (proportional to normal force) (Gidoni et al., 2016):

$F_\text{fric} = -\mu_\text{dir} N \, \mathrm{sign}(\dot{u}), \quad \mu_\text{dir} = G(\theta, w) \cdot E(N)$

4. Well-Posedness, Passivity, and Computational Structure

FrBD models, both ODE and PDE, exhibit robust mathematical properties:

Well-posedness: For bounded and Lipschitz continuous parameters, solutions exist, are unique, and remain bounded globally in time (Romano et al., 20 Jan 2026, Romano, 20 Jan 2026, Romano et al., 11 Jan 2026, Romano, 11 Jan 2026).
Passivity: There exists a Lyapunov (storage) functional $V$ such that input–output pairs (velocity, force) satisfy

$\dot{V} \leq F_\text{fric}(t) \, v(t)$

ensuring energy cannot be generated by friction and facilitating feedback-stable control design.

In distributed and higher-order models, passivity holds when the pressure profile non-increases along the transport direction and the rheological/damping and friction matrices commute.

FrBD models admit linearization for spectral analysis and transfer function computation, making them amenable to observer and controller synthesis for high-performance tribological and robotic systems (Romano et al., 11 Jan 2026, Romano et al., 20 Jan 2026).

5. Applications: Robotics, Rolling Contact, Biotribology, and Material Design

FrBD’s reach encompasses diverse application domains:

Vibration-driven reversible locomotion: Analytical and experimental studies confirm that locomotion direction in bristle bots can be inverted by varying actuation frequency or bristle compliance (Cicconofri et al., 2017, Cicconofri et al., 2014, Kim et al., 2020). Average horizontal speed, net inversion conditions, and the role of parametric tuning have been explicitly derived and experimentally validated.
Vehicle dynamics and tires: Distributed FrBD models within single-track vehicle frameworks reproduce transient, lateral, and micro-shimmy phenomena under steering, accommodating tire carcass flexibility and pressure distributions. They rigorously capture build-up/phase lag and modal instabilities absent from static friction models (Romano et al., 11 Jan 2026).
Polymer brush tribology: Particle-based simulations of erukamide brushes demonstrate that dynamic bristles, especially when coupled with free chains, reduce friction, stabilize thin films, and agree quantitatively with scaling predictions for osmotic pressure and viscosity under shear (Goicochea et al., 2016).
Directional adhesives and crawlers: The geometric–energetic decomposition of macroscale friction informs the design of devices exploiting asymmetric friction for unidirectional or controlled bidirectional motion, using bristle geometry and pre-load (Gidoni et al., 2016).

6. Extensions: Viscoelasticity, Multidimensionality, and Model Identification

Recent research integrates linear viscoelasticity into the FrBD framework:

Generalized Maxwell/Kelvin–Voigt extensions: FrBD $_{n+1}$ –GM/GKV models incorporate multiple relaxation spectra, replicating frequency-dependent hysteresis, internal damping, and complex rate-dependent effects. The distributed counterparts, governed by $2(n+1)$–component PDEs, admit spectral fitting to experimental modulus data and can approximate fractional-derivative behavior (Romano et al., 20 Jan 2026, Romano, 20 Jan 2026).
Two-dimensional and rolling contact: Theories for simultaneous longitudinal/lateral slip, arbitrary spin, and finite patch geometry are established, extending FrBD to state-of-the-art tire–road and rail–wheel modeling (Romano, 11 Jan 2026, Romano, 20 Jan 2026).
Control and observer design: Passivity-based controllers and friction observers exploiting the well-posedness of FrBD models are illustrated for precision robotics (Romano et al., 20 Jan 2026, Romano et al., 11 Jan 2026).
Parameter identification: Stepwise protocols using small-slip and large-slope experiments facilitate empirical fitting of bristle stiffness, damping, and friction law parameters to experimental friction data (Romano, 11 Jan 2026).

7. Limitations, Implementation, and Future Directions

FrBD models abstract the microstructure of the interfacial contact into effective bristle ensembles and do not directly resolve all microscopic slip zones or bulk elasticity; however, coupling with broader elastic models is feasible (Romano, 11 Jan 2026, Romano et al., 20 Jan 2026). Stochastic roughness, temperature effects, nonlinear or history-dependent rheology, and multidimensional bristle networks remain active research frontiers.

Numerical implementation relies on upwind finite volume formulations and method-of-characteristics solvers for distributed systems, with order-reduced variants enabling real-time control applications in vehicles and mechatronics (Romano et al., 11 Jan 2026, Romano, 11 Jan 2026, Romano, 20 Jan 2026).

FrBD thus serves as a rigorously-founded and physically-grounded modeling paradigm, unifying friction phenomena across scales and systems, and providing direct design and control insights for engineered frictional interfaces.