Soft Mobility Theory
Christophe Eloy · 2026
Read Soft Mobility Theory by Christophe Eloy (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-23869.txt.
Introduction
Christophe Eloy offers a unified theory for how soft bodies move and deform in slow viscous flow. The problem spans swimming microbes to soft microrobots. Applying virtual power and a reciprocal theorem to a hyperelastic body yields one configuration dependent differential equation for its generalized coordinates. That equation extends rigid body mobility theory with deformation dependent tensors. Tests on spring linked sphere assemblies validate the approach. A differentiable JAX implementation enables inverse design. It recovers a three sphere swimmer optimum and designs a soft surfer that outclimbs its rigid twin.
Catalog record
| CREATOR | Christophe Eloy |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 77560 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.23869
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Soft Mobility Theory Christophe Eloy Abstract Predicting how a deformable body moves and deforms in a viscous flow underlies problems ranging from microorganism locomotion to soft microrobotics, yet existing frameworks are either problem-specific or ill-suited to inverse design. We propose the soft mobility theory: applying the principle of virtual power and the Lorentz reciprocal theorem to a hyperelastic body in a background Stokes flow yields a configuration-dependent ordinary differential equation for the generalized coordinates of the body. This soft mobility equation extends classical rigid-body mobility theory in that the mobility, elastic, body-force, and flow-coupling tensors all depend explicitly on the instantaneous deformation. We specialize the framework to assemblies of hydrodynamically interacting spheres connected by elastic springs, using the Rotne–Prager–Yamakawa approximation to compute the mobility, and validate it on canonical problems spanning rigid and flexible bodies in quiescent and shear flows. An open-source JAX implementation makes entire simulations end-to-end differentiable. This allows efficient gradient-based inverse design: as proofs of concept, we recover the asymptotic optimum of a three-sphere swimmer and design a soft gyrotactic “surfer” that exploits passive deformation to ascend faster than its rigid counterpart in a Taylor–Green flow. I Introduction Microorganisms rely on their deformability to interact with the surrounding fluid Visser 2001 ; Wheeler et al. 2019 . This deformation can be active when microswimmers propel themselves through their slender flagella and cilia Lauga and Powers 2009 ; Lauga 2016 ; Lauga 2020 ; Guasto et al. 2012 ; Wheeler et al. 2019 . It can also be passive, like red blood cells undergoing large shape changes as they flow through capillaries Barthès-Biesel 2016 ; Farutin et al. 2013 , or diatoms using flexible structures such as mucilage stalks and fine spines to adapt their sinking velocity Smayda 1974 ; Wiseman and Reynolds 1981 ; Padisák et al. 2003 . Across these systems, deformability is not incidental but functional, often conferring sensing, locomotion, or feeding advantages that a rigid body could not achieve. In artificial systems, flexible fibers have been shown to buckle and tumble in quiescent, shear, and turbulent flows Wiggins et al. 1998 ; Brouzet et al. 2014 ; Du Roure et al. 2019 ; Żuk et al. 2021 , and a growing class of soft microrobots is being engineered for biomedical applications, such as in-situ sensing, drug delivery, and microsurgery Sitti 2021 ; Yasa et al. 2023 ; Tsang et al. 2020 ; Elnaggar et al. 2024 ; Dabbagh et al. 2022 ; Mo et al. 2023 ; Jiang et al. 2022 . Even microplastics deform and disperse in the ocean in ways that depend on their shape and flexibility Sutherland et al. 2023 . These problems share a common physical structure: a deformable body whose shape evolves under the combined action of active, elastic, viscous, and body forces. This broad class of fluid–structure interactions is conventionally referred to as elastohydrodynamics . We adopt here the more specific term soft mobility , by which we mean the transport and deformation of a small deformable body immersed in a background flow, in direct analogy with the classical rigid-body mobility theory Brenner 1963 ; Happel and Brenner 1983 ; Kim and Karrila 2013 . Soft mobility encompasses both the forward problem of predicting how a given soft body moves and deforms in a prescribed flow, and the inverse problem of designing soft bodies to perform a target function: efficient locomotion, flow sensing, clustering, or autonomous navigation in complex environments. The latter is increasingly central to soft robotics Kim et al. 2013 ; Yasa et al. 2023 and embodied intelligence Pfeifer and Bongard 2006 , and underlies the related notions of morphological computation Pfeifer et al. 2014 ; Hauser et al. 2011 ; Hauser et al. 2014 ; Nakajima et al. 2015 and physical intelligence Sitti 2021 ; Hauser et al. 2023 , in which sensing, control, and decision-making are partly offloaded from a centralized processor to the mechanical response of the body itself. Solving this inverse problem at low Reynolds number requires a specific approach, which we aim to develop in the present paper. The natural starting point is the classical mobility theory developed in the 1960s for rigid bodies in Stokes flow Brenner 1963 ; Happel and Brenner 1983 ; Kim and Karrila 2013 . Owing to the linearity of the Stokes equations, the six-component velocity of a rigid body depends linearly on the applied force and torque and on the local background velocity, vorticity, and rate-of-strain. The mobility tensors entering this relation depend only on the body’s shape and incorporate all the hydrodynamic complexity of the Stokes equations into a finite set of geometric coefficients. These tensors can be computed with different methods: boundary integral methods Pozrikidis 1992 that exploit Lorentz reciprocal theorem Happel and Brenner 1983 , multiblob methods that discretize body surfaces with “blobs” Balboa Usabiaga et al. 2017 ; Balboa Usabiaga and Delmotte 2022 ; Broms et al. 2023 , or slender-body theory for fibers Tornberg and Shelley 2004 ; Nazockdast et al. 2017 ; Maxian et al. 2022 . This formulation has recently been revived for complex-shaped macromolecules García de la Torre and Bloomfield 1981 ; Zuk et al. 2014 ; Zuk et al. 2018 and reinterpreted as a low-dimensional dynamical system describing rigid-body sinking trajectories Gonzalez et al. 2004 ; Witten and Diamant 2020 . Its appeal, from the standpoint of the inverse problem, is the explicit, shape-dependent linear relation it provides between velocities and loads Moreau et al. 2022 . For N rigid bodies, the same linear relationship extends to a 6N\times 6N grand mobility tensor encoding both single-body mobility and many-body hydrodynamic interactions. This tensor is at the core of Stokesian dynamics, which aims to simulate a large ensemble of rigid bodies interacting in a flow Brady and Bossis 1988 ; Ichiki 2002 ; Fiore and Swan 2019 ; Swan and Wang 2016 . To compute the grand mobility tensor efficiently, multipole expansions are generally used, such as the Rotne–Prager–Yamakawa (RPY) approximation, which provides a closed-form mobility for assemblies of spheres of possibly different radii Rotne and Prager 1969 ; Yamakawa 1970 ; Wajnryb et al. 2013 ; Zuk et al. 2014 ; Cichocki et al. 2021 . This approach is now mature for both passive rigid suspensions and active suspensions of rigid swimmers, routinely tackling problems with millions of degrees of freedom Delmotte et al. 2015a . A natural next step has been to leverage these tools to simulate deformable bodies by treating them as assemblies of rigid sub-bodies linked by elastic or rigid constraints. For instance, bead models or multiblob methods approximate fibers as an ensemble of touching spheres connected by springs Delmotte et al. 2015b ; Żuk et al. 2021 ; Fuchter and Bloomfield-Gadêlha 2023 , and general articulated bodies can be simulated in a similar way Balboa Usabiaga and Delmotte 2022 . At each timestep, however, these formulations require enforcing the constraints, which is a saddle-point linear system, typically solved via preconditioned iterative solvers (GMRES, FMM-accelerated iterations Liang et al. 2013 ; Fiore and Swan 2019 ). This is efficient for forward simulation, but would be prohibitive for the inverse problem of shape optimization. Closest in spirit to the present paper is the framework