Baranowski's ISS Tube Framework Cuts State Estimation Error by 22.4x

State estimates used in sampled monitoring and automation require bounds that stay valid between measurements—a classic control challenge. Jerzy Baranowski's new paper, 'Uniform High-Probability ISS Tubes for Sampled-Data State Estimation,' tackles this head-on with a finite-horizon input-to-state-stability (ISS) tube and observer co-design framework for continuous-time observers driven by sampled-and-held outputs. The sampled-data error model explicitly separates three disturbance channels: process disturbances, sampled measurement noise, and intersample mismatch. By propagating a horizon-level disturbance-envelope event through an ISS estimate, the framework guarantees simultaneous containment of the entire error trajectory.

The core technical achievement lies in using quadratic dissipation inequalities to derive both ellipsoidal and componentwise tubes, then solving a semidefinite program to minimize the normalized tube width across all three channels. A structured nonlinear extension preserves known nonlinear dynamics. Results are striking: co-design reduces the worst normalized half-width by 31% in a linear compartment benchmark and by a factor of 22.4 in a flexible-joint benchmark. This work, submitted to Elsevier, provides a principled way to design observers with provable bounds—critical for safety-critical systems like robotic manipulators, chemical reactors, and autonomous vehicles where sampled-data control is the norm.