FlexiQuad: Flexible Quad Geometry & Mechanics

Updated 7 April 2026

FlexiQuad is a multidisciplinary framework combining isometric quad mesh deformations, adaptive quadrotors, and computational quad meshing to enhance design in geometry and mechanics.
It employs algebraic and geometric methodologies—such as Euler–Chasles correspondence and T-surface parametrizations—to model one-degree-of-freedom flex mechanisms and their kinematic behavior.
Its practical insights drive innovations in deployable architecture, soft robotics, and high-quality numerical simulations by optimizing structural parameters for performance and resilience.

FlexiQuad refers to a diverse set of mathematical, mechanical, and computational concepts centered around flexibility and quadrilateral geometry in discrete and continuum settings. The term is prominent in three principal research domains: (1) flexible and isometric quad-mesh and quad-surface geometries in discrete differential geometry and kinematics; (2) flexible, impact-resilient, or morphologically adaptive aerial vehicles (notably soft or morphing quadrotors); and (3) flexible quad meshing for numerical simulation and computer graphics. In all contexts, the unifying element is the design, classification, or control of systems—combinatorial, mechanical, or robotic—that exploit flexibility (in geometry or structure) subject to quadrilateral or quad-grid architectures.

1. Flexible Quad Meshes and Surfaces: The Geometric and Kinematic Framework

The core mathematical foundation of "FlexiQuad" is the study of flexible polyhedral meshes and surfaces composed of quadrilateral faces, their isometric deformations, and associated kinematic mechanisms.

Kokotsakis Polyhedra and 3×3 Quad Meshes

Flexible Kokotsakis polyhedra with a quadrangular base comprise a central quadrilateral surrounded by a belt of four side quadrilaterals and four corner triangles, all joined by revolute hinges. The space of isometric deformations for such polyhedra is generically trivial (rigid), but discrete flexible mechanisms exist and have been fully classified into eight algebraic families via the Euler–Chasles correspondence and the associated elliptic curve diagrams. These classes include orthodiagonal, isogonal (Miura/Miura-ori), equimodular, conjugate-modular, linear compound, linearly conjugate, and trivial flexes, each characterized by special constraints on the face angles and spherical quadrilateral parameters (Izmestiev, 2014).

Kinematic analysis leverages half-angle variables, Bricard biquadratic relations, and monodromy conditions around the mesh to determine flexion. Recent work has generalized this paradigm to flexible meshes with skew, nonplanar faces and to the reducible quadrilateral case, providing explicit algebraic constructions of mechanisms with up to eight degrees of freedom (Aikyn et al., 2023, Liu, 6 Mar 2026).

T-hedra, T-surfaces, and the Unified FlexiQuad Framework

The class of T-hedra (flexible quad-surfaces with planar strips—rows and columns—meeting orthogonally) and their smooth analogs, T-surfaces, admit a one-parameter family of isometric deformations. Explicit parametrizations exist both in the discrete (vertex grid, edge lengths, angles, and heights) and smooth (coordinate curves, conjugate nets, orthogonality of profile/trajectory planes) categories (Izmestiev et al., 2023). Flexibility arises from special planarity, trapezoidal, and orthogonality constraints; isometric deformation formulas maintain edge lengths and planarity while allowing global shape change.

These developments subsume classical examples such as Miura-ori and discrete surfaces of revolution and are fundamental to applications in deployable structures, architectural geometry, and origami-inspired design.

2. Flexible and Morphologically Adaptive Quadrotors

FlexiQuad also designates a class of bioinspired, highly compliant, and/or morphologically flexible quadrotors, including soft-frame drones and continuum morphing aerial robots.

Bioinspired Anisotropic and Distributed Mass Design

A FlexiQuad prototype realizes radical increases in frame compliance without sacrificing agility or collision resilience. By laser-cutting a high-aspect-ratio FR-4 strip into a closed frame with targeted in-plane stiffness ( $k$ ) and decoupling out-of-plane (lift-bearing) and in-plane (squeezing, impact absorption) stiffness, the design achieves an anisotropy ratio $k_\perp/k\approx2.23$ . Distributed mass–energy, notably allocating battery mass under each motor (“actuation unit”) rather than a central hub, further enhances modal response and impact buffering (Girardi et al., 7 Nov 2025).

Systematic mechanical analysis establishes a regime of optimal structural softness $0.006\lesssim k\lesssim0.77$ N/mm (for $M\in [0.02,3]$ kg), within which agility, squeezability (gap traversal down to 70% nominal width), and collision resilience (survives 5 m/s frontal impacts, 39× reduction in glancing collision force) are jointly achieved. Deviations outside this range yield either excessive internal collision under impact (too soft) or insufficient squeezing/collision buffering (too stiff). Peak accelerations ( $>3$ g, $>300$ rad/s $^2$ ) match rigid quadrotors for TWR $\leq 8$ .

Continuum Morphing and Thrust Vectoring

In tendon-driven continuum morphing quadrotors, each arm comprises a TPU/carbon composite yielding controlled curvature by antagonistic tendon actuation. Arm curvature $\beta$ is determined by the servomotor input with a cubic mapping, and the thrust vector of each propeller tilts with $\beta$ , enabling thrust-vector control for translation without pitch/roll excursions (Verdin et al., 1 Apr 2026). Morphology-driven translation demonstrably produces 0.7 m/s horizontal velocities at constant altitude, with baseline rigidity and stability preserved under standard autopilot stacks.

Control and Planning of Flexible Multi-Body Aerial Systems

The broader class of FlexiQuad systems with suspended payloads and flexible tethers can be rigorously modeled via coordinate-free, geometric Newton–Euler equations for a system comprised of quadrotor(s), multi-link flexible cables ( $k_\perp/k\approx2.23$ 0 links per cable), and possibly rigid-body payloads. Differential flatness holds for all configurations (single/multiple quads, point/ridid payload), allowing regulator and nonlinear planning in the high-dimensional underactuated configuration space (Kotaru et al., 2017). Flat outputs (typically payload position, cable tensions, and quad yaw) parametrize all states/inputs; finite-horizon LQR control via geometric variation-based linearization on $k_\perp/k\approx2.23$ 1 manifolds provides locally exponentially stabilizing trajectories.

3. Flexible Quad Layouts for Meshing and Geometric Modeling

The FlexiQuad pipeline in numerical meshing solves the problem of partitioning a planar or surface domain into a layout of quadrilaterals that accommodate an arbitrary (possibly high valence) singularity pattern, thus enabling high-quality block-structured quad meshing (Jezdimirović et al., 2021).

The workflow proceeds as:

Nonlinear cross-field computation by minimizing the Dirichlet or Ginzburg–Landau energy, extracting singularities from cross-winding patterns.
Mesh refinement near singularities, particularly around high valence points by introducing “bicycle-spokes” local refinements to accommodate $k_\perp/k\approx2.23$ 2 angular variation.
Resolution of a linear system for the holomorphic cross-field via Poisson (for the log-norm scalar field $k_\perp/k\approx2.23$ 3) and Cauchy–Riemann (for direction field $k_\perp/k\approx2.23$ 4) equations with Dirac-delta sources at prescribed singularities.
Extraction and correction of separatrices (limit cycles, T-junctions, non-quad patches) via deterministic algorithms to obtain a pure quadrilateral partition.
Per-patch block-structured quad mesh generation using bilinear transfinite interpolation (TFI) in a harmonic parameterization space.

Mesh quality is assessed via the Remacle ArkMesh metric, with typical results reporting average quality $k_\perp/k\approx2.23$ 5 up to $k_\perp/k\approx2.23$ 6, and $k_\perp/k\approx2.23$ 7 of quads exceeding $k_\perp/k\approx2.23$ 8 in complex CAD domains. The entire algorithm executes in seconds on input meshes with $k_\perp/k\approx2.23$ 9 triangles.

4. Bistable and Mechanically Flexible Quad Net Structures

Within discrete differential geometry, FlexiQuad identifies large bistable networks composed of spatial four-bar linkages, realized as quad-nets in the Study quadric (Szewieczek et al., 1 Apr 2026). Starting from an infinitesimally flexible Koenigs net $0.006\lesssim k\lesssim0.77$ 0, Whiteley de-averaging produces two isometric, generally non-congruent configurations $0.006\lesssim k\lesssim0.77$ 1 for each $0.006\lesssim k\lesssim0.77$ 2. Each corresponds to a valid spatial linkage, and their difference yields a structure with two distinct equilibrium states (bistability). The geometry of the underlying Study quadric and the associated dual quaternions parameterize axis positions and snap angles of the revolute hinges, enabling the geometric prescription of snap-through motions and multi-cell bistable tilings.

5. Orthogonal Graph Drawing and Algorithmic Flexibility

In algorithmic graph drawing, the FlexiDraw problem concerns finding plane orthogonal drawings of 4-planar graphs where each edge $0.006\lesssim k\lesssim0.77$ 3 may have at most $0.006\lesssim k\lesssim0.77$ 4 bends. The computational complexity is sensitive to the distribution of "inflexible" edges (edges with $0.006\lesssim k\lesssim0.77$ 5). Notably, the problem is:

NP-complete when all edges are inflexible, and remains NP-complete with $0.006\lesssim k\lesssim0.77$ 6 inflexible edges at pairwise $0.006\lesssim k\lesssim0.77$ 7 distance for any $0.006\lesssim k\lesssim0.77$ 8.
Fixed-parameter tractable (FPT) in $0.006\lesssim k\lesssim0.77$ 9 (number of "critical" edges) via a dynamic programming SPQR-tree decomposition coupled with planar flow subproblems. The time bound is $M\in [0.02,3]$ 0, where $M\in [0.02,3]$ 1 is the flow computation time (Bläsius et al., 2014).

6. Structural, Mechanical, and Optimization Insights

A recurring theme in FlexiQuad research is the identification of parameter regimes—be it stiffness windows for mechanical drones or algebraic angle/factor constraints for geometric flexors—where desirable combinations of performance, stability, or flexibility are achieved. In quadrotors, structural anisotropy and distributed actuation enable resilience and agility. In geometric and kinematic frameworks, monodromy and factorization conditions unify all 1-DOF flexors of the quadrilateral mesh class. The explicit parametrizations, closure, and compatibility equations provide a foundation for synthesis, analysis, and fabrication of flexible/quadrilateral structures across length scales.

Key references: (Izmestiev, 2014, Izmestiev et al., 2023, Aikyn et al., 2023, Liu, 6 Mar 2026, Verdin et al., 1 Apr 2026, Girardi et al., 7 Nov 2025, Jezdimirović et al., 2021, Bläsius et al., 2014, Szewieczek et al., 1 Apr 2026, Kotaru et al., 2017).