Technical Controllability

Definition

Technical controllability is a design-level necessary condition on a free-flying space manipulation robot (FSMR): each generalized coordinate must be drivable in the commanded direction and at a prescribed rate by its own bounded actuator, irrespective of the other control actions. Formally (Rutkovskii & Glumov 2023, after Glumov et al. 2001), starting from rest at a configuration , applying the maximum control to coordinate (with all ) must produce an acceleration of the same sign as and bounded away from zero. The source proves that whether this holds in a neighborhood of depends only on the design parameters (mass–inertia geometry) of the mechanical system — not on the control-bound vector . This source is genuinely free-flying: the body is actively oriented by gas-jet and electromechanical actuators, so the property concerns both body angular motion and joint motion, not a reactive free-floating base.

Key Equations

Symbols per notation.md.

The robot’s angular motion, linearized about a reference configuration , is

where is the configuration inertia (the source’s ; the inertia symbol in notation.md is , so beware the clash — here is the source’s control-force vector, not inertia) and maps controls to generalized forces.

Technical controllability at requires, for each , from zero initial conditions,

independent of , where are characteristic acceleration values fixed by the mechanism. (The symbol here is the source’s characteristic acceleration; do not confuse it with the quaternion scalar part in notation.md.) When this condition holds over the whole working domain, the joint-block coupling is negligible and the manipulator inertia block may be treated as diagonal.

Source Support

  • rutkovskii2023control — defines technical controllability as a necessary performance condition for an FSMR, gives the linearized model (Eq. 3) and the per-coordinate sign/acceleration test (Eq. via ref [18], Glumov et al. 2001), and uses it to justify neglecting inter-joint coupling (diagonal ) during soft installation.
  • ffsm_dynamics — technical controllability is a property of the coupled FSMR equations of motion ; it certifies that the dynamics are drivable per-coordinate.
  • dynamic_coupling — the condition is exactly a guarantee that the off-diagonal (body↔arm, joint↔joint) coupling terms do not overpower a coordinate’s own actuator, so coupling can be neglected where it holds.
  • coordinated_control — it underwrites the joint use of gas-jet body actuators and electromechanical joint actuators for angular stabilization (motion exchange between body and links).
  • free_floating_dynamics — contrast regime: the source notes the free-floating mode (body orientation control disabled) brings narrowing of the working area and dynamic singularities; technical controllability is asserted for the actuated free-flying mode.
  • generalized_jacobian — the free-floating Generalized Jacobian Matrix (GJM) is the kinematic-rank object whose loss of rank is a dynamic singularity; technical controllability is a distinct, dynamics/design-level drivability condition rather than a Jacobian-rank test.
  • configuration_dependent_stability_domain — the source pairs technical controllability with a stability domain in joint-angle space used to pick the initial configuration and admissible range during installation.

Open Questions

  • The source’s per-coordinate test is built on a simplified angular motion model with small joint rates (Coriolis/centrifugal products neglected). How tight is the condition for our fast-cruise inspection trajectories where those terms are not negligible?
  • The proof (ref [18]) ties controllability to design parameters alone, not to . Does this design-only criterion transfer cleanly to our circumcentroidal coordinate formulation, where the relevant inertia is rather than the raw block?
  • Treating as diagonal once the condition holds is justified for soft, low-rate installation; is the decoupling assumption safe for our coordinated base+arm tracking, or only for quasi-static placement?