Dynamics and Vibrations

Sensors based on nonlinear dynamics and active interrogation
Several current micro-sensing technologies are based on vibratory responses, such as bio-chemical detectors (which use mass measurements) and tapping-mode atomic force microscopy. In those technologies, highly nonlinear systems can provide increased sensitivity and selectivity. In that context, we are creating a comprehensive and radically novel sensing paradigm which provides ultra high sensitivity, robustness, as well as multi-functional sensing capabilities (e.g. sensor self-calibration).

MEMS inertial sensor applications and sports training devices
Miniature and wireless MEMS inertial measurement systems to analyze and to teach the fine motor skills required in sports and other applications. For instance, when attached to or embedded in sports equipment, our patented designs enable detailed analysis of athletic performance. To date, developed designs for sports include those for fly fishing, golf (illustration above right), baseball, hockey, bowling, crew and curling. In addition, we are actively extending our concept to support portable gait measurement, knee injury detection and surgical training.

Inverse problems and system identification: modeling and identification of mistuning in bladed disks
Bladed disks such as those found in jet engines are cyclically symmetric structures. However, slight deviations from this symmetry caused by manufacturing, wear, etc. can result in a highly non-symmetric distribution of stresses during the vibration of such systems. Such a phenomenon is also known as vibration localization and leads to extremely high stresses, causing high cycle fatigue and premature failure of the structure. These deviations, known as mistuning, can be identified using reduced-order modeling techniques along with experimental response data. With the mistuning known, more accurate models can be constructed to predict the dynamics of the system (e.g. stresses and aeroelastic responses).

High-sensitivity damage detection in nonlinear systems
Current vibration-based damage detection methods have significant limitations and exhibit low performance when applied to nonlinear and high-dimensional systems. We take a radically different approach, and overcome these existing limitations through the development of robust and highly sensitive nonlinear techniques for identifying the location and level of multiple, simultaneous damages. One of the enabling key advancements is our novel approach for modeling nonlinear systems using augmented linear system models.

Multi-vehicle systems
Driver behavior has significant effects on emerging multi-vehicle patterns. Analyzing car-following models that incorporate driver reaction time, it can be shown that a single driver can bring vehicular traffic to a halt. When a driver brakes harder than a critical threshold predicted by the theory, he/she creates a ripple that gets amplified as it propagates backward on the chain of vehicles leading to a stop-and-go wave. This nonlinear string instability is responsible for a large fraction of traffic congestion occurring on highways. (Gabor Orosz)