Jerome P. Lynch

Chun Lo, Yechao Bai, Mingyan Liu et al.

When faulty sensors are rare in a network, diagnosing sensors individually is inefficient. This study introduces a novel use of concepts from group testing and Kalman filtering in detecting these rare faulty sensors with significantly fewer number of tests. By assigning sensors to groups and performing Kalman filter-based fault detection over these groups, we obtain binary detection outcomes, which can then be used to recover the fault state of all sensors. We first present this method using combinatorial group testing. We then present a novel adaptive group testing method based on Bayesian inference. This adaptive method further reduces the number of required tests and is suitable for noisy group test systems. Compared to non-group testing methods, our algorithm achieves similar detection accuracy with fewer tests and thus lower computational complexity. Compared to other adaptive group testing methods, the proposed method achieves higher accuracy when test results are noisy. We perform extensive numerical analysis using a set of real vibration data collected from the New Carquinez Bridge in California using an 18-sensor network mounted on the bridge. We also discuss how the features of the Kalman filter-based group test can be exploited in forming groups and further improving the detection accuracy.

Aaron Appelle, Jerome P. Lynch

Large-scale video generation models have demonstrated high visual realism in diverse contexts, spurring interest in their potential as general-purpose world simulators. Existing benchmarks focus on individual subjects rather than scenes with multiple interacting people. However, the plausibility of multi-agent dynamics in generated videos remains unverified. We propose a rigorous evaluation protocol to benchmark text-to-video (T2V) and image-to-video (I2V) models as implicit simulators of pedestrian dynamics. For I2V, we leverage start frames from established datasets to enable comparison with a ground truth video dataset. For T2V, we develop a prompt suite to explore diverse pedestrian densities and interactions. A key component is a method to reconstruct 2D bird's-eye view trajectories from pixel-space without known camera parameters. Our analysis reveals that leading models have learned surprisingly effective priors for plausible multi-agent behavior. However, failure modes like merging and disappearing people highlight areas for future improvement.

Aaron Appelle, Jerome P. Lynch

Recent high-performing image-to-video (I2V) models based on variants of the diffusion transformer (DiT) have displayed remarkable inherent world-modeling capabilities by virtue of training on large scale video datasets. We investigate whether these models can generate realistic pedestrian movement patterns in crowded public scenes. Our framework conditions I2V models on keyframes extracted from pedestrian trajectory benchmarks, then evaluates their trajectory prediction performance using quantitative measures of pedestrian dynamics.