Fault Tolerant and Energy Efficient Signal Processing on FPGA Using Evolutionary Techniques

In this paper, an energy efficient approach using field-programmable gate array (FPGA) partial dynamic reconfiguration (PDR) is presented to realize autonomous fault recovery in mission-critical (space/defence) signal processing applications at runtime. A genetic algorithm (GA) based on adaptive search space pruning is implemented, for reducing repair time thus increasing availability. The proposed method utilizes dynamic fitness function evaluation, which reduces the test patterns for fitness evaluation. Hence, the scalability issue and large recovery time associated with refurbishment of larger circuits is addressed and improved. Experiments with case study circuits, prove successful repair in minimum number of generations, when compared to conventional GA. In addition, an autonomous self-healing system for FPGA based signal processing system is proposed using the presented pruning based GA for intrinsic evolution with the goal of reduced power consumption and faster recovery time.

[1]  Ronald F. DeMara,et al.  Layered Approach to Intrinsic Evolvable Hardware using Direct Bitstream Manipulation of Virtex II Pro Devices , 2007, 2007 International Conference on Field Programmable Logic and Applications.

[2]  J.D. Lohn,et al.  Evolutionary based techniques for fault tolerant field programmable gate arrays , 2006, 2nd IEEE International Conference on Space Mission Challenges for Information Technology (SMC-IT'06).

[3]  Ronald F. DeMara,et al.  Scalable FPGA Refurbishment Using Netlist-Driven Evolutionary Algorithms , 2013, IEEE Transactions on Computers.

[4]  Carthik A. Sharma,et al.  Consensus-Based Evaluation for Fault Isolation and On-line Evolutionary Regeneration , 2005, ICES.

[5]  Stephen P. Boyd,et al.  Self-Tuning for Maximized Lifetime Energy-Efficiency in the Presence of Circuit Aging , 2011, IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems.

[6]  Carthik A. Sharma,et al.  Autonomic fault-handling and refurbishment using throughput-driven assessment , 2011, Appl. Soft Comput..

[7]  Steven F. Quigley,et al.  FPGA Implementation for GMM-Based Speaker Identification , 2011, Int. J. Reconfigurable Comput..

[8]  Narayanan Vijaykrishnan,et al.  Toward Increasing FPGA Lifetime , 2008, IEEE Transactions on Dependable and Secure Computing.

[9]  Ronald F. DeMara,et al.  Sustainable Modular Adaptive Redundancy Technique Emphasizing Partial Reconfiguration for Reduced Power Consumption , 2011, Int. J. Reconfigurable Comput..

[10]  Garrison W. Greenwood.  On the practicality of using intrinsic reconfiguration for fault recovery , 2005, IEEE Transactions on Evolutionary Computation.

[11]  Charles E. Stroud,et al.  Roving STARs: an integrated approach to on-line testing, diagnosis, and fault tolerance for FPGAs in adaptive computing systems , 2001, Proceedings Third NASA/DoD Workshop on Evolvable Hardware. EH-2001.

[12]  Adrian Stoica,et al.  Fault-tolerant evolvable hardware using field-programmable transistor arrays , 2000, IEEE Trans. Reliab..

[13]  Ronald F. DeMara,et al.  Autonomous FPGA fault handling through competitive runtime reconfiguration , 2005, 2005 NASA/DoD Conference on Evolvable Hardware (EH'05).

[14]  Andrew M. Tyrrell,et al.  Challenges of evolvable hardware: past, present and the path to a promising future , 2011, Genetic Programming and Evolvable Machines.

[15]  Subhasish Mitra,et al.  Overcoming Early-Life Failure and Aging for Robust Systems , 2009, IEEE Design & Test of Computers.

[16]  Subhasish Mitra,et al.  Overcoming Early-Life Failure and Aging Challenges for Robust System Design , 2013 .

[17]  Carthik A. Sharma,et al.  Expediting GA-Based Evolution Using Group Testing Techniques for Reconfigurable Hardware , 2006, 2006 IEEE International Conference on Reconfigurable Computing and FPGA's (ReConFig 2006).