Multiple Model Particle Filter for Traffic Estimation and Incident Detection

Abstract

This paper poses the joint traffic state estimation and incident detection problem as a hybrid state estimation problem, in which a continuous variable denotes the traffic state and a discrete model variable identifies the location and severity of an incident. A multiple model particle smoother is proposed to solve the hybrid estimation problem, in which the multiple model particle filter is used to accommodate the nonlinearity and switching dynamics of the traffic incident model, and the smoothing algorithm is applied to improve the accuracy of the estimate when data are limited. The proposed algorithms are evaluated through numerical experiments using CORSIM as the true model. The proposed algorithm is also compared with a standard macroscopic traffic estimator via particle filtering and the California incident detection algorithm. The results show that jointly estimating the state and incidents in one algorithm is better than two dedicated algorithms working independently.

Back to the Future: Predicting Traffic Shockwave Formation and Propagation Using a Convolutional Encoder-Decoder Network

“…Edie [21] estimates the average density k(A) for a time-space block of A (e.g. 100 feet by 1 second) based on equation (4). In this equation, |A| is the area of the time-space block A, and t(A) stand for the total time spent by all the vehicles going through block A.…”

Robust adaptive divided difference filter based on forgetting factors and its applications

“…In the works of Dey et al and Sabet et al, the novel parameter estimation methods were obtained based on divided difference filter (DDF), extended KF (EKF), and unscented KF, respectively. Interacting multiple model is another type of strategy for systems with biases or parameter uncertainties . Based on the multiple‐model approach, Li and Jilkov and Li and Jia considered the same motion model with a Markovian switching parameter, which has been widely used in the target tracking community for handling maneuvers of moving targets.…”

“…Based on the multiple‐model approach, Li and Jilkov and Li and Jia considered the same motion model with a Markovian switching parameter, which has been widely used in the target tracking community for handling maneuvers of moving targets. A multiple model particle smoother was studied to solve the hybrid estimation problem in the work of Wang et al In the work of Yu et al, multiple state models were used to represent the different ballistic missile dynamics in three flight phases, and a state‐dependent interacting multiple model approach based on Gaussian particle filtering was developed. In the work of Xiang et al, a class of widely linear quaternion multiple‐model adaptive estimation algorithms based on widely linear quaternion KFs and Bayesian inference was proposed to track time‐variant model uncertainty.…”

Comparing traffic state estimators for mixed human and automated traffic flows

“…Beginning in the early 1980's, a modified version of Payne's macroscopic model has been used for a variety of Kalman-based estimators [20][21][22][23]. Nonlinear variants of Kalman filtering [24][25][26][27][28][29] and particle filtering [30][31][32][33][34][35] have also been applied to modifications of the Lighthill Whitham Richards (LWR) partial di↵erential equation (PDE) and its discretization [36][37][38][39].…”