Dynamic Coupling

Definition

Dynamic coupling is the bidirectional inertial interaction by which manipulator
motion disturbs the base spacecraft’s position and attitude — and, conversely, base
motion perturbs the arm — in a spacecraft-manipulator system. It is the structural
consequence of the off-diagonal base–arm block of the generalized inertia matrix
being non-zero, so the base and arm equations of motion cannot be solved
independently (das2025understanding,
wilde2018equations). Its manifestation is
regime-dependent: in the free-floating regime (uncontrolled base, conserved
momentum) coupling appears as the reactive base drift that arm motion induces,
which corrupts end-effector accuracy (das2025understanding).
In our free-flying regime (fully-actuated 6-DOF base) the same coupling block is
present, but the base actuators can actively counteract the reaction rather than
absorb it — so coupling becomes a control term to be compensated or coordinated,
not an unavoidable drift (zhang2020adaptive,
rutkovskii2023control).

Key Equations

Symbols per notation.md.

Coupling enters through the off-diagonal block of the full coupled inertia matrix
. With the base twist and joint rates
, the kinetic-energy / inertia partition is
(wilde2018equations Eq. 27,
virgili-llop2016spacecraft Eq. 44):

The block is exactly the canonical base–arm coupling inertia
( in notation.md; written in the sources).
Coupling vanishes iff ; its kernel is the
reaction null space. In the free-floating regime momentum conservation ties the
two blocks directly (das2025understanding,
wilde2018equations Eq. 42):

i.e. every joint motion forces a reactive base twist. The coupling can be reduced to a
scalar strength metric from a
coupling factor (ellery2004engineering),
and das2025 instead takes the SVD of the joint-to-base velocity map
(the reactive-drift mapping above; das2025 names it
, a velocity-coupling matrix — not the Coriolis , and not the inertia
block itself), characterizing it by the normalized Shannon entropy of its
singular-value spectrum plus a directionality cosine
(das2025understanding Eqs. 23–24, 28–29).

Source Support

  • das2025understanding — primary: defines dynamic coupling as the reactive base motion induced by arm motion under momentum conservation; cleanly separates free-flying (actuated base) from free-floating (uncontrolled base); quantifies coupling via the SVD of its joint-to-base velocity map (das2025’s "", not the inertia block itself) and proposes exploiting rather than only minimizing it.
  • wilde2018equations — canonical EOM derivation; introduces the dynamic-coupling inertia matrix and the five-mode regime taxonomy (floating → rotation/translation/full flying) that pins down which momenta are conserved.
  • virgili-llop2016spacecraft — same partition of the generalized inertia matrix; companion modeling/software reference.
  • zhang2020adaptive — frames coupling as the defining difference between space and fixed-base robots; multi-arm case adds arm–arm coupling; TDE+SMC actively decouples/compensates (free-flying base).
  • rutkovskii2023control — free-flying (FSMR) regime, our regime: dynamic coupling between body and manipulators must be accounted for when controlling configuration and base attitude.
  • sukhanov2015dynamic — free-flying dynamic model with the EE-to-target deviation as an explicit coordinate; supports small-joint-speed simplifications of the coupled model.
  • seweryn2008optimization — GJM extended to the four base regimes including free-flying with non-conserved momentum; arm motion disturbs the servicing satellite during rendezvous/docking.
  • ye2019research — ties coupling to reaction-null-space planning; adaptive RNS for free-floating redundant manipulators.
  • virgili-llop2019convex — exploits the coupling structure: decouples a system-wide translation sub-maneuver, then re-couples for internal reconfiguration in a convex capture-guidance scheme.
  • ellery2004engineering — defines the dynamic coupling coefficient ; cites ETS-VII evidence that coupling imposed significant control difficulty; notes dedicated attitude control nulls the coupling.
  • sentis2005control — operational-space treatment of a terrestrial humanoid with a free-floating base (NOT a space base); useful only as a task-prioritization analogue for handling base–task coupling, regime differs.
  • ffsm_dynamics — the full coupled equations of motion in which dynamic coupling is the off-diagonal inertia block.
  • free_floating_dynamics — the regime where coupling manifests as conserved-momentum base drift; the most-studied special case.
  • generalized_jacobian — the GJM folds momentum-conservation coupling into an effective EE Jacobian (a free-floating construct).
  • coupling_inertia_matrix — the () block itself; this page is its phenomenological view.
  • base_disturbance_minimization — the planning/control objective of suppressing the base disturbance that coupling produces.
  • reaction_null_space — the kernel of : joint motions that excite zero base reaction.
  • momentum_conservation — the constraint that, in the free-floating regime, ties base drift to joint motion.
  • dynamic_singularity — configurations where the coupling-folded (generalized) Jacobian loses rank.
  • free_flying_vs_free_floating — the regime distinction that determines whether coupling is a drift to absorb or a term to actively compensate.

Open Questions

  • The clean free-floating identity rests on conserved momentum; with our fully-actuated base the momenta are not conserved (per wilde2018equations’s “flying” mode and seweryn2008optimization). Does the same block still fully characterize coupling once external base wrenches enter, or does the relevant quantity become the residual base-tracking error after compensation?
  • das2025 argues coupling can be exploited (energy-aware trajectories) rather than minimized — but does so for a free-floating base. For a free-flying base, is there still a net benefit to exploiting coupling, or does base actuation make pure compensation strictly preferable?
  • ellery2004 notes dedicated attitude control drives the coupling factor toward zero (). Quantitatively, where is the crossover at which active base compensation costs more (fuel / reaction-wheel saturation) than tolerating residual coupling?