Modeling Attitude Control in Microgravity
Building on earlier empirical and simulation-based work, this month focused on defining the mathematical model that describes how internal leg movements can influence the robot's global orientation in microgravity. The system was modelled as a non-holonomic free-floating body, with no external torque or fixed base. As in the classic falling cat problem, reorientation must emerge entirely from internal actuation - specifically, coordinated changes in the robot's limb configuration. The body was described with 6 DoF, but control efforts focused on roll (ϕ) and pitch (θ). Each leg has 2 DoF (extension and swing), giving 8 independent inputs. Using this, the control problem was defined as an underdetermined linear system, where the Jacobian:
maps joint velocities to changes in roll and pitch rates:
A Constraint-Aware Bandit
In February, the control architecture for Continuity took a substantial step forward with the formal integration of the bandit-based decision layer into the QuectoFSM state machine framework. This was the first implementation of what I termed Constrained Local Learning (CLL): a structure where bandit-based exploration is actively gated by contextual rules, and actions are only permitted when predefined safety and state conditions are satisfied. The multi-armed bandit (MAB) component, now in version 3, moved beyond basic reward selection and hard-coded action cycling. Instead of optimising only low-level parameters (e.g., step height or cycle duration), the new system could also select from a list of gait patterns, each encoded as an arm with an associated performance history. To regulate instability, the bandit’s reward function was overhauled. The original inverse formulation:
was replaced by a log-scaled penalty system, better suited for sharply discouraging large attitude deviations while maintaining sensitivity around small corrections:
