Joint target tracking and classification with particle filtering and mixture Kalman filtering using kinematic radar information

Abstract

This paper considers the problem of joint maneuvering target tracking and classification. Based on recently proposed Monte Carlo techniques, a multiple model (MM) particle filter and a mixture Kalman filter (MKF) are designed for two-class identification of air targets: commercial and military aircraft. The classification task is carried out by processing radar measurements only, no class (feature) measurements are used. A speed likelihood function for each class is defined using a prior information about speed constraints. Class-dependent speed likelihoods are calculated through the state estimates of each class-dependent tracker. They are combined with the kinematic measurement likelihoods in order to improve the classification process. The two designed estimators are compared and evaluated over rather complex target scenarios. The results demonstrate the usefulness of the proposed scheme for the incorporation of additional speed information. Both filters illustrate the opportunity of the particle filtering and mixture Kalman filtering to incorporate constraints in a natural way, providing reliable tracking and correct classification. Future observations contain valuable information about the current state of the dynamic systems. In the framework of the MKF, an algorithm for delayed estimation is designed for improving the current modal state estimate. It is used as an additional, more reliable information in resolving complicated classification situations.

A Multi-Classification Method of Improved SVM-based Information Fusion for Traffic Parameters Forecasting

“…Some others have employed Kalman filtering technique [7] or neural networks [8] and system identification [9], and more recently a nonparametric paradigm has been adopted via kernel functions [10]. These proposals are not suitable for traffic information classification and fusion except in some special situations (for some network configurations or with high detector coverage).…”

Integrated target tracking and recognition via joint appearance-motion generative models

“…Due to the non-Gaussian property of I t and non-linearity of X t , we resort to the particle filtering approach where the posterior probability is maintained by a weighted sample set. The samples at time t can be generated from those at time (t − 1) by (2), and weights are assigned by (1). A better proposal is derived by incorporating the current appearance observation in the spirit of APF [16] and the estimation can be further improved by utilizing the Gaussian probability of p(K t |I t , K t−1 ).…”

“…6. From these 3-D models we obtain silhouettes of dimension 60 × 80 corresponding to different poses (1 • to 360 • ) of a particular tank. These silhouettes are transformed into gray-scale images using signed distance transform as in [5,7], to impose smoothness of the distance between poses.…”

“…Traditionally, vision research focuses on the development of powerful appearance models [7,18] (i.e. observation model), while signal processing research places emphasis on the system model by focusing on the dynamics of maneuvering targets [1,11]. Recently, there is a trend to combine both appearance and motion cues into a joint tracking and recognition flow [8,21].…”

“…In reality, different targets would be equipped with customized engines and mechanical systems that generate distinct motion patterns. This inspires researchers to develop multiple type-dependent motion models to achieve joint target tracking and recognition [1,11]. However, these approaches require sensors giving direct kinematics measurements, e.g., radar, and do not consider the target appearance.…”