Reliable Hardware Architectures of CORDIC Algorithm with Fixed Angle of Rotations

Fixed-angle rotation operation of vectors is widely used in signal processing, graphics, and robotics. Various optimized coordinate rotation digital computer (CORDIC) designs have been proposed for uniform rotation of vectors through known and specified angles. Nevertheless, in the presence of faults, such hardware architectures are potentially vulnerable. In this thesis, we propose efficient error detection schemes for two fixed-angle rotation designs, i.e., the Interleaved Scaling and Cascaded Single-rotation CORDIC. To the best of our knowledge, this work is the first in providing reliable architectures for these variants of CORDIC. The former is suitable for low-area applications and, hence, we propose recomputing with encoded operands schemes which add negligible area overhead to the designs. Moreover, the proposed error detection schemes for the latter variant are optimized for efficient applications which hamper the performance of the architectures negligibly. We present three variants of recomputing with encoded operands to detect both transient and permanent faults, coupled with signature-based schemes. The overheads of the proposed designs are assessed through Xilinx FPGA implementations and their effectiveness is benchmarked through error simulations. The results give confidence for the proposed efficient architectures which can be tailored based on the reliability requirements and the overhead to be tolerated.