• Electronics and Telecommunications Trends
• /
• v.32 no.6
• /
• pp.57-65
• /
• 2017

Neural network based AI computing is a promising technology that reflects the recognition and decision operation of human beings. Early AI computing processors were composed of GPUs and CPUs; however, the dramatic increment of a floating point operation requires an energy efficient AI processor with a highly parallelized architecture. In this paper, we analyze the trends in processor architectures for AI computing. Some architectures are still composed using GPUs. However, they reduce the size of each processing unit by allowing a half precision operation, and raise the processing unit density. Other architectures concentrate on matrix multiplication, and require the construction of dedicated hardware for a fast vector operation. Finally, we propose our own inAB processor architecture and introduce domestic cutting-edge processor design capabilities.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.10 no.2
• /
• pp.283-290
• /
• 2006

The cellular neural networks have the structure that consists of an array of the same cell which is a simple processing element, and each of the cells has local connectivity and space invariant template properties. So, it has a very suitable structure for the hardware implementation. But, it is impossible to have a one-to-one mapping between the CNN hardware processors and the pixels of the practical large image. In this paper, a $5{\times}5$ CNN hardware processor with pipeline input and output that can be applied to the time-multiplexing processing scheme, which processes the large image with a small CNN cell block, is designed. the operation of the implemented $5{\times}5$ CNN hardware processor is verified from the edge detection and the shadow detection experimentations.

• Proceedings of the Korean Society of Machine Tool Engineers Conference
• /
• 1997.10a
• /
• pp.225-229
• /
• 1997

This paper describes the camera calibration based-neural network with a camera modeling that accounts for major sources of camera distortion, namely, radial, decentering, and thin prism distortion. Radial distortion causes and inward or outward displacement of a given image point from its ideal location. Actual optical systems are subject to various degrees of decentering, that is, the optical centers of lens elements are not strictly collinear. Thin prism distortion arises from imperfection in lens design and manufacturing as well as camera assembly. It is our purpose to develop the vision system for the pattern recognition and the automatic test of parts and to apply the line of manufacturing. The performance of proposed camera calibration is illustrated by simulation and experiment.

• Electronics and Telecommunications Trends
• /
• v.35 no.3
• /
• pp.66-75
• /
• 2020

As the increasing expectations of a practical AI (Artificial Intelligence) service makes AI algorithms more complicated, an efficient processor to process AI algorithms is required. To meet this requirement, processors optimized for parallel processing, such as GPUs (Graphics Processing Units), have been widely employed. However, the GPU has a generalized structure for various applications, so it is not optimized for the AI algorithm. Therefore, research on the development of AI processors optimized for AI algorithm processing has been actively conducted. This paper briefly introduces an AI processor especially for inference acceleration, developed by the Electronics and Telecommunications Research Institute, South Korea., and other global vendors for mobile and server platforms. However, the GPU has a generalized structure for various applications, so it is not optimized for the AI algorithm. Therefore, research on the development of AI processors optimized for AI algorithm processing has been actively conducted.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.11 no.5
• /
• pp.1597-1600
• /
• 2010

In this paper, a direct controller for nonlinear plants using a neural network is presented. The composed of the control input by using RBF neural networks and auxiliary input to compensate for effects of the approximation errors and disturbances. In the results, using this scheme, the output tracking error between the plant and the reference model can asymptotically converge to zero in the presence of bounded disturbances and approximation errors. Simulation results show that it is very effective and can realize a satisfactory control of the nonlinear system.

• Journal of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• v.13 no.4
• /
• pp.109-119
• /
• 1999

Switched Reluctance Motor(SRM) have been expanding gradually their awlications in the variable speed drives due to their relatively low cost, simple and robust structure, controllability and high efficiency. In this paper neural network theory is used to detemrine fuzzy-neural network controller's membership ftmctions and fuzzy rules. In addition neural network emulator is used to emulate forward dynamics of SRM and to get error signal at fuzzy-neural controller output layer. Error signal is backpropagated through neural network emulator. The backpropagated error of emulator offers the path which reforms the fuzzy-neural network controller's mmbership ftmctions and fuzzy rules. 32bit Digital Signal Processor(TMS320C31) was used to achieve the high speed control and to realize the fuzzy-neural control algorithm. Simulation and experimental results show that in the case of load variation the proposed control rrethcd was superior to a conventional rrethod in the respect of speed response.sponse.

• The Transactions of the Korean Institute of Electrical Engineers A
• /
• v.48 no.4
• /
• pp.419-427
• /
• 1999

We propose an algorithm for obtaining the optimal node number of hidden units in dynamic neural networks. The dynamic nerual networks comprise of dynamic neural units and neural processor consisting of two dynamic neural units; one functioning as an excitatory neuron and the other as an inhibitory neuron. Starting out with basic network structure to solve the problem of control, we find optimal neural structure by multiplication and combining dynamic neural unit. Numerical examples are presented for nonlinear systems. Those case studies showed that the proposed is useful is practical sense.

• Journal of Korea Multimedia Society
• /
• v.17 no.3
• /
• pp.324-332
• /
• 2014

This paper presents a programmable System-On-a-chip for various embedded applications that need Neural Network computations. The system is fully implemented into Field-Programmable Gate Array (FPGA) based prototyping platform. The SoC consists of an embedded processor core and a reconfigurable hardware accelerator for neural computations. The performance of the SoC is evaluated using real image processing applications, such as optical character recognition (OCR) system.

• Journal of the Korea Society of Computer and Information
• /
• v.4 no.4
• /
• pp.94-100
• /
• 1999

This paper concerns the implementation of a digital neural network which performs the high speed operation of Hopfield model's arithmetic operation. It is also designed to use a look-up table and produce floating point arithmetic of nonlinear function with high speed operation. The arithmetic processing of Hopfleld is able to describe the matrix-vector operation, which is adaptable to design the array processor because of its recursive and iterative operation .The proposed method is expected to be applied to the field of real neural networks because of the realization of the current VLSI techniques.

• Proceedings of the KIEE Conference
• /
• 1999.07b
• /
• pp.890-893
• /
• 1999

The intent of this paper is to describe a neural network structure called dynamic neural processor(DNP), and examine how it can be used in developing a learning scheme for computing robot inverse kinematic transformations. The architecture and learning algorithm of the proposed dynamic neural network structure, the DNP, are described. Computer simulations are demonstrate the effectiveness of the Proposed learning using the DNP.