Distributed Particle Filter Algorithm Research Articles

Distributed Particle Filtering of $\alpha$ -Stable Signals

In order to present an inclusive framework for distributed estimation/tracking of $\alpha$ -stable signals, a novel distributed particle filtering algorithm is developed. This is achieved through the reformulation of the particle filtering operations from the point of view of the characteristic function and is based on the decomposition of the operations of the particle filter so that they can be distributed among the agents of a sensor network while allowing each agent to retain an estimate of the state vector. In contrast to current distributed particle filtering techniques that approximate distributions with Gaussian mixtures through empirical estimates of the second-order statics and are, thus, limited to signals with finite variance, the developed distributed particle filtering approach is suitable for the generality of $\alpha$ -stable signals, allowing the proposed algorithm to be used in a multitude of applications. Finally, the so introduced distributed particle filtering approach is validated through a simulation example.

Improved distributed particle filters for tracking in a wireless sensor network

A novel distributed particle filter algorithm is presented, called drift homotopy likelihood bridging particle filter (DHLB-PF). The DHLB-PF is designed to surmount the degeneracy problem by employing a multilevel Markov chain Monte Carlo (MCMC) procedure after the resampling step of particle filtering. DHLB-PF considers a sequence of pertinent stationary distributions which facilitates the MCMC step as well as explores the state space with a higher degree of freedom. The proposed algorithm is tested in a multi-target tracking problem using a wireless sensor network where no fusion center is required for data processing. The observations are gathered only from the informative sensors, which are sensing useful observations of the nearby moving targets. The detection of those informative sensors, which are typically a small portion of the sensor network, is taking place by using a sparsity-aware matrix decomposition technique. Simulation results showcase that the DHLB-PF outperforms current popular tracking algorithms.

Distributed Particle Filter for Target Tracking: With Reduced Sensor Communications

For efficient and accurate estimation of the location of objects, a network of sensors can be used to detect and track targets in a distributed manner. In nonlinear and/or non-Gaussian dynamic models, distributed particle filtering methods are commonly applied to develop target tracking algorithms. An important consideration in developing a distributed particle filtering algorithm in wireless sensor networks is reducing the size of data exchanged among the sensors because of power and bandwidth constraints. In this paper, we propose a distributed particle filtering algorithm with the objective of reducing the overhead data that is communicated among the sensors. In our algorithm, the sensors exchange information to collaboratively compute the global likelihood function that encompasses the contribution of the measurements towards building the global posterior density of the unknown location parameters. Each sensor, using its own measurement, computes its local likelihood function and approximates it using a Gaussian function. The sensors then propagate only the mean and the covariance of their approximated likelihood functions to other sensors, reducing the communication overhead. The global likelihood function is computed collaboratively from the parameters of the local likelihood functions using an average consensus filter or a forward-backward propagation information exchange strategy.

DART: Distributed Particle Filter Algorithm with Resampling Tree for Ultimate Real-Time Capability

A novel Distributed Particle Filter Algorithm with Resampling Tree, called DART, is proposed in this paper, where particles are resampled by Branch Resampling and Root Resampling in a flexible tree-like structure. Though sampling and weight calculation can be executed in parallel on a group of Processing Elements, resampling is the bottleneck for distributed particle filters since it requires the knowledge of the whole particle set. Conventional approaches to accelerate resampling on distributed platforms often introduce extra procedure other than the standard processing flow and achieve acceleration limited by linear speedup. By introducing the proposed algorithm, where Branch Resampling can be executed in parallel with sampling and weight calculation, the number of particles in the final sequential implemented Root Resampling can be reduced in an exponential relationship with the depth of the tree. With the same linear speedup in sampling and weight calculation steps, the overall speedup achieved in DART surpasses linear boundary and outperforms state-of-art approaches. The corresponding implementation architecture, which possesses unique features of hardware efficiency and scalability, is also presented. The prototype of the algorithm with 8 PEs is implemented on a Xilinx Virtex-IV Pro FPGA (XC4VFX100-12FF1152) under BOT system. With 8192 particles, the input observation can achieve 63.3 kHz at a clock speed of 80 MHz.