SPOCS: Application Specific Signal Processor for OFDM Communication Systems

Abstract

This paper presents an Application-Specific Signal Processor (ASSP) for Orthogonal Frequency Division Multiplexing (OFDM) Communication Systems, called SPOCS. The instruction set and its architecture are specially designed for OFDM systems, such as Fast Fourier Transform (FFT), scrambling/descrambling, puncturing, convolutional encoding, interleaving/deinterleaving, etc. SPOCS employs the optimized Data Processing Unit (DPU) to support the proposed instructions and the FFT Address Generation Unit (FAGU) to automatically calculate input/output data addresses. In addition, the proposed Bit Manipulation Unit (BMU) supports efficient bit manipulation operations. SPOCS has been synthesized using the SEC 0.18 μm standard cell library and has a much smaller area than commercial DSP chips. SPOCS can reduce the number of clock cycles over 8% 53% for FFT and about 48%~84% for scrambling, convolutional encoding and interleaving compared with existing DSP chips. SPOCS can support various OFDM communication standards, such as Wireless Local Area Network (WLAN), Digital Audio Broadcasting (DAB), Digital Video Broadcasting-Terrestrial (DVB-T), etc.

Low-Power Application-Specific Processor for FFT Computations

“…It should be noted that cycle count of 6002 is obtained with eight memory ports while the proposed processor uses only two. The SPOCS [2] is an ASP with instruction extensions tailored for FFT computations which supports FFT from 64-to 8,196-points, but the implemented system would only support FFT to 2,048-points.…”

“…One example is SPOCS [2], which contains special instructions for FFT computations reducing the implementation complexity and cycle count. SPOCS processor consist: program control unit, address unit with two address generation units, specialized FFT addres generation unit, and address register, optimized data processing unit with two MACs, one ALU, barrel shifter and register file, program memory and two dual ported data memories.…”

The implementation of the clustered‐OFDM‐based transceiver on an FPGA device: A comprehensive comparison

“…The use of ASIP was discussed in [32][33][34][35][36]. Currently, ASIP is applied to low-speed data rates transceivers.…”

Low Complexity Time Synchronization Algorithm for OFDM Systems with Repetitive Preambles

“…As will be demonstrated later, the proposed algorithm achieves good performance for low signal to noise ratio (SNR) and has a low-cost VLSI implementation that would facilitate its integration in a multi-standard processor for OFDM communications [24]. This paper is organized as follows: Section 2 presents the system model used in the rest of the paper.…”

“…This residual CFO can be reduced if more samples are used in the average. When we introduce this residual CFO in (24) and (25) we observe that E{QSS} hardly changes and the threshold of 0.4 still gives a probability of false alarm lower than 1×10 -3 for SNR higher than 3 dB. The probability of detection failure ( ) Thr Q GI < Pr does not worsen with the presence of a CFO lower than 0.1 of the subcarrier spacing, since factor ) , , ( L M ω κ is nearly 1 at this CFO, as can be seen in Fig.…”