Parallelized Particle Swarm Optimization on FPGA for Realtime Ballistic Target Tracking.

Area of Science:

• * Aerospace Engineering
• * Computer Science
• * Signal Processing

Background:

• * Real-time tracking of high-speed ballistic targets presents significant computational challenges.
• * Conventional particle swarm optimization (PSO) is computationally intensive, limiting its application in real-time systems.
• * Nonlinearity in measurement models and target dynamics further complicate ballistic target tracking.

Purpose of the Study:

• * To develop an accelerated particle swarm optimization (PSO) technique for real-time ballistic target tracking.
• * To address the computational time limitations of traditional PSO in high-speed tracking scenarios.
• * To leverage hardware acceleration for improved tracking performance.

Main Methods:

• * Implementation of a parallelized particle swarm optimization (PSO) algorithm.
• * Utilizing a field-programmable gate array (FPGA) for hardware acceleration of the PSO algorithm.
• * Testing and analysis on a heterogeneous processing system integrating an FPGA.

Main Results:

• * The proposed parallelized PSO achieved successful real-time ballistic target tracking.
• * Tracking results were comparable to conventional PSO methods.
• * Computation time was significantly reduced by up to 3.89× compared to CPU-based PSO.

Conclusions:

• * Parallelized PSO on FPGAs offers a viable solution for real-time ballistic target tracking.
• * Hardware acceleration effectively overcomes the computational limitations of traditional PSO.
• * The proposed method demonstrates substantial improvements in processing speed for high-speed tracking applications.