Intent Inference and Syntactic Tracking with GMTI Measurements

IEEE Trans. Signal Process.

Gao

2018

Self Cite

Syntactic tracking aims to classify a target's spatiotemporal trajectory by using natural language processing models. This paper proposes constrained stochastic context free grammar (CSCFG) models for target trajectories confined to a roadmap. We present a particle filtering algorithm that exploits the CSCFG model structure to estimate the target's trajectory. This metalevel algorithm operates in conjunction with a base-level target tracking algorithm. Extensive numerical results using simulated ground moving target indicator (GMTI) radar measurements show useful improvement in both trajectory classification and target state (both coordinates and velocity) estimation.Keywords-syntactic tracking, constrained stochastic context free grammar (CSCFG), particle filter, Earley Stolcke parser 1 Traditionally, given track information, a radar (human) operator examines target trajectories to determine anomalous behavior. This paper develops natural language models and meta-level signal processing algorithms for estimating anomalous trajectories. Such "middleware" forms the interface between the physical signal processing layer and the radar operator; see also [7] for alternative models for intent using bridging distributions.

Section: A Syntactic Tracker Architecturesmentioning

confidence: 99%

Syntactic Enhancement to VSIMM for Roadmap Based Anomalous Trajectory Detection: A Natural Language Processing Approach

IEEE Trans. Signal Process.

Gao

2018

Self Cite

“…where P{z k |x k , a k , e k } is the likelihood of the sensor measurement given by (3) and P{x k , a k , e k |z 1:k−1 } is the predicted pdf given by (4).…”

Section: A Bayesian Filteringmentioning

confidence: 99%

“…A variety of approximate filtering algorithms have been proposed in the literature using particle filters [1], [2]. Our formulation builds on [1] by additionally incorporating mode-dependent state dynamics that are designed to take into account trajectory constraints [3]. A similar use of SCFGs in a different application (action recognition) appears in [4].…”

Section: Introductionmentioning

confidence: 99%

Syntactic track-before-detect

Fanaswala

2013 5th IEEE International Workshop on Computational Advances in Multi-Sensor Adaptive Processing (CAMSAP)

2013

Self Cite

In this paper, a track before detect approach utilizing trajectory shape constraints is proposed to track dimly lit targets. The shape of the target trajectory is modeled syntactically using stochastic context-free grammar (SCFG) models that arise in natural language processing. These scale-invariant models are subsequently used in enhancing the track before detect algorithm. Stochastic context-free grammars are a generalization of Markov chains (regular grammars) and can model complex spatial patterns with long range dependencies. A novel particle filter based syntactic tracker is proposed and numerical results are presented to show significant improvement over conventional jump Markov models in track before detect.

“…We omit the exact details of the filter operation and refer the reader to the detailed account of the Earley-Stolcke filter in [3]. Additionally, its use for parsing grammars in radar tracking applications is detailed in [4]. We can define inside probabilities The inside probabilities can be understood as the probability that the non-terminal i eventually generates all the observations Z(s), .…”

Section: The Earley-stolcke Filter For Scfgsmentioning

confidence: 99%

Destination-aware target tracking via syntactic signal processing

Fanaswala

2011 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

White

2011

Self Cite

We consider the prediction of a target's destination and simultaneously recover its filtered trajectory. Two novel models for trajectories with known destinations are presented using reciprocal stochastic processes and stochastic contextfree grammars. We present a destination-aware syntactic tracker which uses conventional state-space estimates from a legacy tracker to perform prediction and trajectory estimation. We also provide statistical signal processing algorithms for model prediction and maximum likelihood sequence estimation using the proposed trajectory models. Simulation results show that both models considered in the paper have superior estimation performance compared to conventional hidden Markov modeling and can reliably predict the target's destination.