Search results

Stochastic computing which is an alternative method of the binary calculation has key merits such as fault-tolerant capability and low hardware cost. However, the hardware response time of it is required to be very fast due to its bit-wise calculation mode. While the complementary metal oxide semiconductor (CMOS) components are difficult to meet the requirements aforementioned. For this, the stochastic computing implementation scheme based on the memristive system is proposed to reduce the response time. The purpose of this paper is to provide the implementation scheme based memristive system for the stochastic computing.

The hardware structure of material logic based on the memristive system is realized according to the advantages of the memristor. After that, the scheme of NOT logic, AND logic and multiplexer are designed, which are the basic units of stochastic computing. Furthermore, a stochastic computing system based on memristive combinational logic is structured and its validity is verified successfully by operating a case.

The numbers of the elements of the proposed stochastic computing system are less than the conventional stochastic computing based on CMOS circuits.

The paper proposed a novel implementation scheme for stochastic computing based on the memristive systems, which are different from the conventional stochastic computing based on CMOS circuits.

The purpose of this paper is to improve the performance of E-nose in the detection of wound infection. Feature extraction and selection methods have a strong impact on the performance of pattern classification of electronic nose (E-nose). A new hybrid feature matrix construction method and multi-objective binary quantum-behaved particle swarm optimization (BQPSO) have been proposed for feature extraction and selection of sensor array.

A hybrid feature matrix constructed by maximum value and wavelet coefficients is proposed to realize feature extraction. Multi-objective BQPSO whose fitness function contains classification accuracy and a number of selected sensors is used for feature selection. Quantum-behaved particle swarm optimization (QPSO) is used for synchronization optimization of selected features and parameter of classifier. Radical basis function (RBF) network is used for classification.

E-nose obtains the highest classification accuracy when the maximum value and db 5 wavelet coefficients are extracted as the hybrid features and only six sensors are selected for classification. All results make it clear that the proposed method is an ideal feature extraction and selection method of E-nose in the detection of wound infection.

The innovative concept improves the performance of E-nose in wound monitoring, and is beneficial for realizing the clinical application of E-nose.