Kalman filtering is a recursive mathematical algorithm that estimates the state of a dynamic system by combining predictions from a motion model with noisy sensor measurements, iteratively minimizing estimation error through a predict-and-update cycle. In the context of crowd sensing and motion estimation, it is particularly valuable for tracking pedestrian trajectories and filtering noise from sensor data streams — such as those from IoT devices or camera-based systems — enabling smoother and more accurate real-time state estimation in dense or complex environments. Key variants relevant to the field include the Extended Kalman Filter (EKF), which linearizes nonlinear system dynamics, and the Unscented Kalman Filter (UKF), which better handles highly nonlinear motion patterns characteristic of high-density crowd behavior.
Source Papers
- Motion estimation of high density crowd using fluid dynamics ↗ — Motion estimation of high density crowd using fluid dynamics
- Sensing Technologies for Crowd Management, Adaptation, and Information Dissemination in Public Transportation Systems: A Review ↗ — Sensing Technologies for Crowd Management, Adaptation, and I