High-Speed Fault-Tolerant Finite Impulse Response Digital Filter on Field Programmable Gate Array

doi:10.1007/s12204-020-2214-z

Abstract

Abstract: Some fast finite impulse response (FIR) filters use a large number of look-up tables (LUTs) to configure distributed random-access memories (RAMs) and save registers. The distributed RAMs store 2M precomputed sums of M permuted operands in order to simplify the accumulation, which lays similarity to the solution of Boolean satisfiability (SAT) problem. In this work, a high-speed fault-tolerant FIR digital filter on field programmable gate array (FPGA) is proposed for hardware implementation. A shift register and an RAM are used to arrange the data flow. Generally, an N-tap digital filter only requires N embedded multipliers on FPGA. The better performance is due to high-radix words and low-latency operations. A 32-tap 8-bit FIR digital filter enjoys a throughput of 9.17MB/s, taking 109 ns to calculate one convolution. In addition, a fault-tolerant scheme by majority logic is used to correct real-time errors within digital filters.

Key words: finite impulse response (FIR) | digital filter | fault-tolerant | fast architecture

摘要： Some fast finite impulse response (FIR) filters use a large number of look-up tables (LUTs) to configure distributed random-access memories (RAMs) and save registers. The distributed RAMs store 2M precomputed sums of M permuted operands in order to simplify the accumulation, which lays similarity to the solution of Boolean satisfiability (SAT) problem. In this work, a high-speed fault-tolerant FIR digital filter on field programmable gate array (FPGA) is proposed for hardware implementation. A shift register and an RAM are used to arrange the data flow. Generally, an N-tap digital filter only requires N embedded multipliers on FPGA. The better performance is due to high-radix words and low-latency operations. A 32-tap 8-bit FIR digital filter enjoys a throughput of 9.17MB/s, taking 109 ns to calculate one convolution. In addition, a fault-tolerant scheme by majority logic is used to correct real-time errors within digital filters.

关键词: finite impulse response (FIR) | digital filter | fault-tolerant | fast architecture

CLC Number:

TN 911.72

WU Tao (吴焘). High-Speed Fault-Tolerant Finite Impulse Response Digital Filter on Field Programmable Gate Array[J]. J Shanghai Jiaotong Univ Sci, 2021, 26(4): 554-558.

References 10

[1]	SODERSTRAND M A, ESCOTT R A. VLSI implementation in multiple-valued logic of an FIR digital filter using residue number system arithmetic[M]//Residue Number System Arithmetic: Modern Applications in Signal Processing. New York, USA: IEEE Press, 1986: 165-185.
[2]	PONTARELLI S, CARDARILLI G C, RE M, et al. Optimized implementation of RNS FIR filters based on FPGAs [J]. Journal of Signal Processing Systems,2012, 67: 201-212.
[3]	TAY T F, CHANG C H. Fault-tolerant computing in redundant residue number system [M]//Embedded Systems Design with Special Arithmetic and Number Systems. Switzerland: Springer International Publishing,2017: 65-88.
[4]	PONTARELLI S, CARDARILLI G C, RE M, et al. A novel error detection and correction technique for RNS based FIR filters [C]//IEEE International Symposium on Defect and Fault Tolerance of VLSI Systems. Washington,USA: IEEE, 2008: 436-444.
[5]	LIN S, COSTELLO D J. Error control coding [M]. 2nd ed. Upper Saddle River, NJ, USA: Pearson Education,Inc., 2004.
[6]	GAO Z, REVIRIEGO P, PAN W, et al. Fault tolerant parallel filters based on error correction codes [J]. IEEE Transactions on Very Large Scale Integration (VLSI )Systems, 2015, 23(2): 384-387.
[7]	MEHER P K, CHANDRASEKARAN S, AMIRA A. FPGA realization of FIR filters by efficient and flexible systolization using distributed arithmetic [J]. IEEE Transactions on Signal Processing, 2008, 56(7): 3009-3017.
[8]	PARK S Y, MEHER P K. Efficient FPGA and ASIC realizations of a DA-based reconfigurable FIR digital filter [J]. IEEE Transactions on Circuits and Systems II : Express Briefs, 2014, 61(7): 511-515.
[9]	WANG C L. New bit-serial VLSI implementation of RNS FIR digital filters [J]. IEEE Transactions on Circuits and Systems, 1994, 41(11): 768-771.
[10]	MOHAN P A. Residue number systems: Algorithms and architectures [M]. Boston, USA: Kluwer Academic Publishers, 2002.