이동 목표물의 효율적인 위치 추정을 위한 파티클 필터 신호 처리의 GPU 기반 가속화
GPU-based Acceleration of Particle Filter Signal Processing for Efficient Moving-target Position Estimation
1
조회
0
다운로드
1
인용
0
참고문헌
키워드
초록
Time of difference of arrival (TDOA) method using passive sonar sensor array has normally been used to estimate the location of a concealed moving target in underwater environment. Particle filter has been introduced for effective target estimation for non-Gaussian and nonlinear systems. In this paper, we propose a GPU-based acceleration of target position estimation using particle filter and propose efficient embedded system and software architecture. For the TDOA measurement from the passive sonar sensor, we use the generalized cross correlation phase transform (GCC-PHAT) method to obtain the correlation coefficient of the signal using FFT and we try to accelerate the calculation of GCC-PHAT based TDOA measurements using FFT with GPU CUDA. We also propose parallelization method of the target position estimation algorithm using the GPU CUDA to update the state of each particle for the target position estimation using the measured values. The target estimation algorithm was verified using Matlab and implemented using GPU CUDA. Then, we realized the proposed signal processing acceleration system using NVIDIA Jetson TX1 as the target board to analyze in terms of the execution time. The execution time of the algorithm is reduced by 55% to the CPU standalone-operation on the target board. Experiment results show that the proposed architecture is a feasible solution in terms of high-performance and area-efficient architecture.
Abstract
Time of difference of arrival (TDOA) method using passive sonar sensor array has normally been used to estimate the location of a concealed moving target in underwater environment. Particle filter has been introduced for effective target estimation for non-Gaussian and nonlinear systems. In this paper, we propose a GPU-based acceleration of target position estimation using particle filter and propose efficient embedded system and software architecture. For the TDOA measurement from the passive sonar sensor, we use the generalized cross correlation phase transform (GCC-PHAT) method to obtain the correlation coefficient of the signal using FFT and we try to accelerate the calculation of GCC-PHAT based TDOA measurements using FFT with GPU CUDA. We also propose parallelization method of the target position estimation algorithm using the GPU CUDA to update the state of each particle for the target position estimation using the measured values. The target estimation algorithm was verified using Matlab and implemented using GPU CUDA. Then, we realized the proposed signal processing acceleration system using NVIDIA Jetson TX1 as the target board to analyze in terms of the execution time. The execution time of the algorithm is reduced by 55% to the CPU standalone-operation on the target board. Experiment results show that the proposed architecture is a feasible solution in terms of high-performance and area-efficient architecture.
참고문헌 (11)
- M. T. Isik, "A Three Dimensional Localization Algorithm for Underwater Acoustic Sensor Networks", IEEE Transactions on Wireless Communications, 8(9), pp.4457~4463 (2009)
- S. Poursheikhali, "TDOA Based Target Localization in Inhomogenous Underwater Wireless Sensor Network", Proceedings of 5th International Conference on Computer and Knowledge Engineering, pp.1~6 (2015)
- 배경숙, "지능형 서비스 로봇의 주의집중을 위한 시간지연 기반 실시간 음원추적 기술개발", 대한임베디드공학회논문지, 2(4), pp.227~232 (2007), 대한임베디드공학회
- Y. Xu, "Underwater Acoustic Source Localization Method Based on TDOA With Particle Filtering", Proceedings of the 26th Chinese Control and Decision Conference, pp.4634~4637 (2014)
- J. Vermaak, "Nonlinear Filtering for Speaker Tracking in Noisy and Reverberant Environments", Proceeding of IEEE International Conference on Acoustics, Speech, and Signal Processing, 5, pp.3021~3024 (2001)
- F. Gustafsson, "Positioning Using Time-difference of Arrival Measurements", Proceedings of IEEE International Conference on Acoustics, Speech, and Signal Processing, 6, pp.553~556 (2003)
- B. Van Den Broeck, "Time-domain Generalized Cross Correlation Phase Transform Sound Source Localization for Small Microphone Arrays", Proceedings of 5th European DSP Education and Research Conference, pp.76~80 (2012)
- B. Qin, "Subsample Time Delay Estimation via Improved GCC PHAT Algorithm", Proceedings of 9th International Conference on Signal Processing, pp.2579~2582 (2008)
- B. Kwon, "Analysis of the GCC-PHAT Technique for Multiple Sources", Proceedings of International Conference on Control Automation and Systems, pp.2070~2073 (2010)
- C. Choi, "RGB-D Object Tracking: A Particle Filter Approach on GPU", Proceedings of IEEE, 2013, 1084~1091
- G. Hendeby, "A Graphics Processing Unit Implementation of the Particle Filter", Proceedings of 15th European Signal Processing Conference, pp.1639~1643 (2007)