Analog Neural Network Implementation for a Real-Time Surface Classification Application

Gatet, L.; Tap-Beteille, H.; Lescure, M.

doi:10.1109/jsen.2008.920713

Cited by 29 publications

(6 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The electrical characteristic corresponds to a continuous and differentiable logistic function [30,31,32,33,34], thus allowing for the use of a back-propagation training algorithm. It is worth mentioning that the differential pair is the basis of the proposed activation function circuit because of the symmetrical output characteristic, and the well-defined maximum and minimum saturation levels.…”

Section: Ann-based Microelectronic Circuit For Sensor Conditioningmentioning

confidence: 99%

High-Level Modeling and Simulation Tool for Sensor Conditioning Circuit Based on Artificial Neural Networks

Martínez-Nieto

Marques

Sanz-Pascual

et al. 2019

Sensors

View full text Add to dashboard Cite

For current microelectronic integrated systems, the design methodology involves different steps that end up in the full system simulation by means of electrical and physical models prior to its manufacture. However, the higher the circuit complexity, the more time is required to complete these simulations, jeopardizing the convergence of the numerical methods and, hence, meaning that the reliability of the results are not guaranteed. This paper shows the use of a high-level tool based on Matlab to simulate the operation of an artificial neural network implemented in a mixed analog-digital CMOS process, intended for sensor calibration purposes. The proposed standard tool enables modification of the neural model architecture to adapt its characteristics to those of the electronic system, resulting in accurate behavioral models that predict the complete microelectronic IC system behavior under different operation conditions before its physical implementation with a simple, time-efficient, and reliable solution.

show abstract

Section: Ann-based Microelectronic Circuit For Sensor Conditioningmentioning

confidence: 99%

High-Level Modeling and Simulation Tool for Sensor Conditioning Circuit Based on Artificial Neural Networks

Martínez-Nieto

Marques

Sanz-Pascual

et al. 2019

Sensors

View full text Add to dashboard Cite

show abstract

“…4.a), has been implemented in 0.35 µm CMOS technology ( Fig. 4.b) [23]. It approximates the distance with a satisfaying precision in a range three times wider than the one limited by the 2kπ indecision.…”

Section: Cmos Neural Networkmentioning

confidence: 99%

Embedded System for Distance Measurement and Surface Discrimination Applications

Tap

Gatet

Moutaye

et al. 2018

2018 Conference on Design of Circuits and Integrated Systems (DCIS)

Self Cite

View full text Add to dashboard Cite

The development of robotics and miniaturization has lead to the design of intelligent embedded sensors capable of measuring distance either on wide ranges or with high resolution. Yet, it is impossible to gather both information at the same time performing a single measurement. Furthermore, the enlightened surface type may be detected only under wellknown experimental conditions. Computer processing is also often necessary to analyze the sensors data. In this paper, we report an embedded sensor capable of measuring distances with a resolution of 250 µm in a range of 2.7 m and of discriminating a wide variety of surface types. The embedded system is based on a phase-shift laser rangefinder, whose essential components are a Laser Diode (ML4412N) and its intelligent CMOS laser driver circuit, a CMOS Avalanche PhotoDiode (APD) and its transimpedance circuit and a CMOS Analog Neural Network.

show abstract

“…Hardware implementation of neural networks has been used in applications such as pattern recognition [1], olfaction system [2], test of analog circuits [3], range-finding [4] and smart sensing [5].…”

Section: Introductionmentioning

confidence: 99%

Efficient hardware implementation of threshold neural networks

Zamanlooy

Mirhassani

2012

10th IEEE International NEWCAS Conference

View full text Add to dashboard Cite

Area and Noise to Signal Ratio (NSR) are two main factors in hardware implementation of neural networks. Despite attempts to reduce the area of sigmoid and hyperbolic tangent activation functions, they cannot achieve the efficiency of threshold activation function. A new NSR efficient architecture for threshold networks is proposed in this paper. The proposed architecture uses different number of bits for weight storage in different layers. The optimum number of bits for each layer is found based on the mathematical derivation using stochastic model. Network training is done using the recently introduced learning algorithm called Extreme Learning Machine (ELM). A 4-7-4 network is considered as a case study and its hardware implementation for different weight accuracies is investigated.The proposed design is more efficient considering area × NSR as a performance metric. VLSI implementation of the proposed architecture using a 0.18 μm CMOS process is presented which shows 44.16%, 58.04 % and 67.30% improvement for total number of bits equal to 16, 20 and 24.

show abstract

Analog Neural Network Implementation for a Real-Time Surface Classification Application

Cited by 29 publications

References 41 publications

High-Level Modeling and Simulation Tool for Sensor Conditioning Circuit Based on Artificial Neural Networks

High-Level Modeling and Simulation Tool for Sensor Conditioning Circuit Based on Artificial Neural Networks

Embedded System for Distance Measurement and Surface Discrimination Applications

Efficient hardware implementation of threshold neural networks

Contact Info

Product

Resources

About