Actuator-aware sampled-data envelope supervision for a preview active quarter-car suspension

Preview active suspension must improve ride comfort while preserving limited suspension travel. In embedded implementations, sampled road preview, zero-order hold, and actuator lag can make actuator-idealized stroke screens nonconservative. This paper shows that, for a single-input preview active quarter-car suspension, implementation-matched envelope supervision can be collapsed to a scalar online kernel: an actuator-aware sampled-data predictor generates hard-feasible command intervals, their intersection is projected against the nominal Prev-LQR command, and only local infeasibility activates a scalar soft fallback. Proposition 1 establishes the local implication used by the zero-slack screen at the sampling instants under explicitly stated tightening conditions. On the fixed 81-case in-envelope benchmark, the supervisor reduces the mean proactive-envelope violation rate from 5.80e−4 for nominal Prev-LQR to zero, with no observed exceedances of the audit threshold S max on the reported benchmark, while leaving mean comfort and actuator burden nearly unchanged. Replay profiling on an STM32G431CBTx Cortex-M4F target showed a worst observed execution time of 41 . 3 μ s . Within the single-input quarter-car scope studied here, the results show that actuator-aware sampled-data envelope supervision can be realized as a compact embedded scalar kernel.