Multiple model particle filter track-before-detect for range ambiguous radar

A novel approach to detecting and tracking a moving target using synthetic aperture radar (SAR) images is proposed in this paper. Achieved with the particle filter (PF) based track-before-detect (TBD) algorithm, the approach is capable of detecting and tracking the low signal-to-noise ratio (SNR) moving target with SAR systems, which the traditional track-after-detect (TAD) approach is inadequate for. By incorporating the signal model of the SAR moving target into the algorithm, the ambiguity in target azimuth position and radial velocity is resolved while tracking, which leads directly to the true estimation. With the sub-area substituted for the whole area to calculate the likelihood ratio and a pertinent choice of the number of particles, the computational efficiency is improved with little loss in the detection and tracking performance. The feasibility of the approach is validated and the performance is evaluated with Monte Carlo trials. It is demonstrated that the proposed approach is capable to detect and track a moving target with SNR as low as 7 dB, and outperforms the traditional TAD approach when the SNR is below 14 dB.

Detection and tracking of a moving target using SAR images with the particle filter-based track-before-detect algorithm.

Lightweight tracking-by-detection system for multiple pedestrian targets

This paper presents a deep and extensive performance analysis of the particle filter (PF) algorithm for a very compute intensive 3D multi-view visual tracking problem. We compare different implementations and parameter settings of the PF algorithm in a CPU platform taking advantage of the multithreading capabilities of the modern processors and a graphics processing unit (GPU) platform using NVIDIA CUDA computing environment as developing framework. We extend our experimental study to each individual stage of the PF algorithm, and evaluate the quality versus performance trade-off among different ways to design these stages. We have observed that the GPU platform performs better than the multithreaded CPU platform when handling a large number of particles, but we also demonstrate that hybrid CPU/GPU implementations can run almost as fast as only GPU solutions.

Performance evaluation of a 3D multi-view-based particle filter for visual object tracking using GPUs and multicore CPUs

We propose an efficient sequential Monte Carlo probability hypothesis density (PHD) filter which employs the Kalman-gain approach during weight update to correct predicted particle states by minimizing the mean square error between the estimated measurement and the actual measurement received at a given time in order to arrive at a more accurate posterior. This technique identifies and selects those particles belonging to a particular target from a given PHD for state correction during weight computation. Besides the improved tracking accuracy, fewer particles are required in the proposed approach. Simulation results confirm the improved tracking performance when evaluated with different measures.

/pdf/kalman-gain-aided-particle-phd-filter-for-multitarget-5fid5yg2q1.pdf

Kalman-Gain Aided Particle PHD Filter for Multitarget Tracking

Track-before-detect (TBD) based on the particle filter (PF) algorithm is known for its outstanding performance in detecting and tracking of weak targets. However, large amount of calculation leads to difficulty in real-time applications. To solve this problem, effective implementation of the PF-based TBD on the graphics processing units (GPU) is proposed in this article. By recasting the particles propagation process and weights calculating process on the parallel structure of GPU, the running time of this algorithm can greatly be reduced. Simulation results in the infrared scenario and the radar scenario are demonstrated to compare the implementation on two types of the GPU card with the CPU-only implementation.

Particle Filter Track-Before-Detect Implementation on GPU

SUMMARY

We propose a novel resampling algorithm to deal with the sample impoverishment problem of particle filter. The algorithm introduces sigma points into the generic resampling procedure to enrich the representation of the posterior distribution. Compared with other relevant resampling algorithms, such as the exquisite resampling algorithm reported recently, our algorithm has shown an improved estimation performance and reduced computational complexity in the simulation of two different nonlinear systems. Copyright © 2012 John Wiley & Sons, Ltd.

A sigma point‐based resampling algorithm in particle filter

A novel particle-PHD filter algorithm is proposed to deal with the multi-target tracking. It takes into account the most recent measurements by the unscented Kalman filter, not in the step of proposal distribution generation as usual, but in resampling step, to enhance the efficiency of the particle sampling. Simulation results show that the proposed algorithm outperforms the algorithms in the literature in performance but with extremely less computational cost.

Jian Zhou

Papers

Particle Filter Track-Before-Detect Implementation on GPU

A sigma point‐based resampling algorithm in particle filter

Improved Particle Implementation of the Probability Hypothesis Density Filter in Resampling

Improving the efficiency and efficacy of robust sequential bifurcation under data contamination