Optimizing HW/SW codesign towards reliability for critical-application systems

Abstract

This work presents an innovative approach for hardware/software codesign of safety-critical computing systems. The proposed approach is based on system reliability requirements to decide which parts of the system are partitioned into hardware or software. The approach considers as inpiit a complete software description of the design. In our case, we use as the initial description the C language and then, for those parts compiled to hardware, the Handel-C language is applied. After partitioning, we verqy system reliability based on an adaptation of the weak mutation analvsis technique. This technique was originally proposedfor software testing by means of vercving the adequacy of a test vectors set for a given program. We also present a case study in order to illustrate the proposed approach.

Speech Recognition Systems: From a Conventional SW Implementation to a Reliable Noise-Immunity HW-SW Version

“…In a second step, this approach also estimates the final reliability for the system on the design. These procedures are partially automated by a CAD tool (FT-PRO Methodology) [1][2][3] and are performed at the initial steps of the design flow [17,26]. By doing so, the main goal of this approach is to minimize time and design cost (including prototyping, testing and reliability estimation) [3][4][5][6][7].…”

“…The pattern matching and the decision logic steps are the "identification" steps, where the words spelled in the speech signal are recognized by generating a sequence of text words. The observation sequence is evaluated using Hidden Markov Models (HMM), which, as the acoustic reference pattern, plays the main role in the recognition process [16][17][18]. To do so, the pattern matching and decision logic block input data (i.e., the observation sequences from the Signal Analysis Block) are actually an "index" to access the local cache memories associated with each HMM block, as seen in fig.…”

“…Then, recognizing the word. Usually, the pattern matching and decision logic steps have been implemented by the Viterbi algorithm [8,12,17,25] to perform the evaluation of the code-label sequence against the HMM structures. Fig.…”