Marcelo G. S. Bruno

Consensus-based distributed particle filtering algorithms for cooperative blind equalization in receiver networks

We describe in this paper novel consensus-based distributed particle filtering algorithms which are applied to cooperative blind equalization of frequency-selective channels in a network with one transmitter and multiple receivers. The proposed algorithms employ parallel consensus averaging iterations to evaluate the product of some node-dependent quantities across the receiver network, thus eliminating the need for message broadcasts beyond each receivers local neighborhood. Additionally, parallel minimum consensus iterations are used to assess the convergence of the quantized consensus averages and ensure accordingly the coherence of particle sets across the different network nodes. We verify via computer simulations that the consensus-based schemes exhibit a small performance gap compared to both centralized and communication-intensive broadcast solutions.

Improved sequential Monte Carlo filtering for ballistic target tracking

We present in this correspondence an improved sequential Monte Carlo (SMC) filter for ballistic target tracking with random, time-varying ballistic coefficient. The proposed tracker is a sampling/importance resampling (SIR) filter that uses an optimized importance function to combat particle degeneracy, and also incorporates an additional measurement-driven Markov chain Monte Carlo (MCMC) move step to prevent particle impoverishment. Simulation results show that, using significantly fewer particles than previously reported in the literature for similar tracking problems, the root mean-square error (RMSE) curves for the proposed optimized SIR filter approach the square root of the ideal posterior Cramer-Rao lower bound (PCRLB).

Improved particle filters for ballistic target tracking

We present in this paper two improved particle filter algorithms for ballistic target tracking. The first algorithm is a sampling/importance resampling (SIR) filter that uses an optimized importance function plus residual resampling to combat particle degeneracy, and also incorporates a Metropolis-Hastings (MH) move step to reduce particle impoverishment. The second proposed algorithm is an auxiliary particle filter (APF). Both algorithms show good performance results when compared to the ideal posterior Cramer-Rao lower bound for the mean square estimation error.

Sequential importance sampling filtering for target tracking in image sequences

We propose in this letter a new approach to direct target tracking in cluttered image sequences using sequential importance sampling (SIS). We use Gauss-Markov random field modeling to describe the clutter correlation and incorporate the clutter and target signature models into the design of the SIS tracking algorithm. We quantify the performance of the SIS tracker using a simulated image sequence generated from real infrared airborne radar data and compare it to the performance of a grid-based hidden Markov model tracker. Simulation results show good performance for the proposed algorithms in a scenario of very low target-to-clutter ratio.

Cooperative particle filtering for emitter tracking with unknown noise variance

We introduce in this paper a novel cooperative particle filter algorithm for tracking a moving emitter using received-signal strength (RSS) measurements with unknown observation noise variance. In the studied scenario, multiple RSS sensors passively observe independently attenuated and perturbed versions of the same broadcast signal transmitted by an emitter which is moving through the sensor field and cooperate to estimate the emitter state. The new algorithm differs from previous methods by employing a parametric approximation to reduce the associated communication burden.

Sequential Monte Carlo methods for joint detection and tracking of multiaspect targets in infrared radar images

We present in this paper a sequential Monte Carlo methodology for joint detection and tracking of a multiaspect target in image sequences. Unlike the traditional contact/association approach found in the literature, the proposed methodology enables integrated, multiframe target detection and tracking incorporating the statistical models for target aspect, target motion, and background clutter. Two implementations of the proposed algorithm are discussed using, respectively, a resample-move (RS) particle filter and an auxiliary particle filter (APF). Our simulation results suggest that the APF configuration outperforms slightly the RS filter in scenarios of stealthy targets.

Collaborative emitter tracking using Rao-Blackwellized random exchange diffusion particle filtering

We introduce in this paper the fully distributed, random exchange diffusion particle filter (ReDif-PF) to track a moving emitter using multiple received signal strength (RSS) sensors. We consider scenarios with both known and unknown sensor model parameters. In the unknown parameter case, a Rao-Blackwellized (RB) version of the random exchange diffusion particle filter, referred to as the RB ReDif-PF, is introduced. In a simulated scenario with a partially connected network, the proposed ReDif-PF outperformed a PF tracker that assimilates local neighboring measurements only and also outperformed a linearized random exchange distributed extended Kalman filter (ReDif-EKF). Furthermore, the novel ReDif-PF matched the tracking error performance of alternative suboptimal distributed PFs based respectively on iterative Markov chain move steps and selective average gossiping with an inter-node communication cost that is roughly two orders of magnitude lower than the corresponding cost for the Markov chain and selective gossip filters. Compared to a broadcast-based filter which exactly mimics the optimal centralized tracker or its equivalent (exact) consensus-based implementations, ReDif-PF showed a degradation in steady-state error performance. However, compared to the optimal consensus-based trackers, ReDif-PF is better suited for real-time applications since it does not require iterative inter-node communication between measurement arrivals.

Sequential Bayesian Algorithms for Identification and Blind Equalization of Unit-Norm Channels

In many estimation problems of interest, the unknown parameters reside on spherical manifolds. As most common filtering algorithms assume that parameters have Gaussian prior distributions, their application to such problems leads to suboptimal performance. In this letter, we propose a model in which the unknown unit-norm parameter vectors have Fisher-Bingham (F-B) prior distributions. We show that if the observations relate to the parameters via Gaussian likelihoods, the F-B priors form a conjugate model that yields closed-form, recursive estimators that naturally take into account the restrictions on the unknowns. We apply this model to a communication setup with multiple gain-controlled FIR frequency-selective channels, deriving a novel maximum a posteriori (MAP) channel parameter estimator and a blind equalizer based on Rao-Blackwellized particle filters. As we verify via Monte Carlo numerical simulations, the F-B model leads to superior performance compared to previous algorithms that adopt mismatched Gaussian prior models.

A particle filtering algorithm for cooperative blind equalization using VB parametric approximations

We introduce in this paper a new distributed sequential Monte Carlo (SMC) algorithm for blind equalization of frequency-selective broadcast channels. In the considered setup, multiple receiving nodes sense independently distorted versions of the same broadcast signal and cooperate to recover it. The proposed approach innovates by using parametric approximations based on the Variational Bayes (VB) method that allow the inter-node communication burden to be greatly reduced compared to previous communication-intensive distributed SMC algorithms. We verify via numerical simulations that the proposed method yields better performance than alternative methods that employ ad hoc parametric approximations, while preserving roughly the same computational cost.

A hybrid GMM/SMC diffusion Bernoulli Filter for joint distributed detection and tracking

We introduce in this paper the Random Exchange Diffusion Bernoulli Filter (RndEx-BF), which enables joint target detection and tracking by a network of collaborative sensors. RndEx-BF is a fully distributed algorithm that, unlike consensus-based solutions, does not require iterative internode communication between sensor measurements. Internode communication cost is further reduced by a novel hybrid GMM/SMC implementation of the proposed filter. Experimental results show that RndEx-BF approaches the performance of a flooding-based implementation of the optimal centralized Bernoulli filter with much lower bandwidth requirements.