FPGA implementation of fixed point CORDIC-SVD for E-skin systems

In this paper we present a CORDIC based FPGA implementation of Singular Value Decomposition for real time tactile information processing. The design is implemented to handle an arbitrary size two dimensional sensor input matrix. Besides the accuracy of the computation, implementation results figure out a relationship between used resources/latency and input data resolution: for each additional 1 bit input data resolution in the proposed case study, the average percentage slice logic utilization increases of 5.9%, and the average percentage clock cycles increases of 7.68%. This result demonstrates the necessity of computing the Singular Value Decomposition with fixed point arithmetic. Otherwise, silicon area and time delay will increase dramatically.

[1]  Wai-Chi Fang,et al.  A VLSI design of singular value decomposition processor for portable continuous-wave diffusion optical tomography systems , 2011, 2011 IEEE/NIH Life Science Systems and Applications Workshop (LiSSA).

[2]  Wai-Chi Fang,et al.  A pipeline VLSI design of fast singular value decomposition processor for real-time EEG system based on on-line recursive independent component analysis , 2013, 2013 35th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC).

[3]  Shashank Ahire,et al.  Implementation of Fixed and Floating Point Square Root Using Nonrestoring Algorithm on FPGA , 2013 .

[4]  Arturo Garcia-Perez,et al.  Reconfigurable FPGA-Based Unit for Singular Value Decomposition of Large m x n Matrices , 2011, 2011 International Conference on Reconfigurable Computing and FPGAs.

[5]  Ray Andraka,et al.  A survey of CORDIC algorithms for FPGA based computers , 1998, FPGA '98.

[6]  Joseph Zambreno,et al.  An FPGA Implementation of the Hestenes-Jacobi Algorithm for Singular Value Decomposition , 2014, 2014 IEEE International Parallel & Distributed Processing Symposium Workshops.

[7]  Gene H. Golub,et al.  Calculating the singular values and pseudo-inverse of a matrix , 2007, Milestones in Matrix Computation.

[8]  Aurobinda Routray,et al.  Comparative Evaluation of Symmetric SVD Algorithms for Real-time Face and Eye Tracking , 2015, ArXiv.

[9]  Gordon Cheng,et al.  Directions Toward Effective Utilization of Tactile Skin: A Review , 2013, IEEE Sensors Journal.

[10]  Johan A. K. Suykens,et al.  A kernel-based framework to tensorial data analysis , 2011, Neural Networks.

[11]  Paolo Gastaldo,et al.  A Tensor-Based Pattern-Recognition Framework for the Interpretation of Touch Modality in Artificial Skin Systems , 2014, IEEE Sensors Journal.