This paper is focused on the exploitation of intrinsic nonlinear dynamics toward novel measurement systems and readout methodologies. In particular, sensors that can be represented as nonlinear dynamical systems and are often reducible to systems described by a static nonlinearity are considered; the nonlinear behavior therefore reduces to the dynamics of a system characterized by two or more (meta)stable equilibrium states (or attractors) separated by energetic thresholds to be overcome to transition from one attractor to the other. The presence of a weak unknown target signal is assessed via the monitoring of the "residence times" in the attractors. This operational scenario that is based on the monitoring of suitable "events" avoids an "amplitude-based" readout and provides a very simple and sensitive readoutprocessing scheme. Many noise effects are also mitigated by the intrinsic decoupling between the amplitude domain of the input signal and the event or time domain of the output signal. We present here the general transduction methodology for this class of "residence-times difference" sensors, together with the experimental results obtained from the working versions of these sensors (in particular, a simple fluxgatemagnetometer).We then introduce some novel dynamical behavior that occurs when the active nonlinear (in this case, bistable) elements are coupled using well-crafted coupling topologies. Sensors based on these coupling schemes provide several advantages over their single-element counterparts. We discuss the dynamics of the coupled-element device, summarizing recent theoretical and experimental results. Finally, we describe the construction and performance of working devices (magnetic- and electric-field sensors) based on these concepts.