Although it is now possible to integrate many millions of transistors on a single chip, traditional digital circuit technology is now reaching its limits, facing problems of cost and technical efficiency when scaled down to ever-smaller feature sizes. The analysis of biological neural systems, especially for visual processing, has allowed engineers to better understand how complex networks can effectively process large amounts of information, whilst dealing with difficult computational challenges.Analog and parallel processing are key characteristics of biological neural networks. Analog VLSI circuits using the same features can therefore be developed to emulate brain-style processing. Using standard CMOS technology, they can be cheaply manufactured, permitting efficient industrial and consumer applications in robotics and mobile electronics.This book explores the theory, design and implementation of analog VLSI circuits, inspired by visual motion processing in biological neural networks. Using a novel approach pioneered by the author himself, Stocker explains in detail the construction of a series of electronic chips, providing the reader with a valuable practical insight into the technology.Analog VLSI Circuits for the Perception of Visual Motion:analyses the computational problems in visual motion perception;examines the issue of optimization in analog networks through high level processes such as motion segmentation and selective attention;demonstrates network implementation in analog VLSI CMOS technology to provide computationally efficient devices;sets out measurements of final hardware implementation;illustrates the similarities of the presented circuits with the human visual motion perception system;includes an accompanying website with video clips of circuits under real-time visual conditions and additional supplementary material.With a complete review of all existing neuromorphic analog VLSI systems for visual motion sensing, Analog VLSI Circuits for the Perception of Visual Motion is a unique reference for advanced students in electrical engineering, artificial intelligence, robotics and computational neuroscience. It will also be useful for researchers, professionals, and electronics engineers working in the field.
Author(s): Alan A. Stocker
Edition: 1
Year: 2006
Language: English
Pages: 242
Tags: Приборостроение;Схемотехника;
Analog VLSI Circuits for the Perception of Visual Motion......Page 4
Contents......Page 10
Foreword......Page 14
Preface......Page 16
1 Introduction......Page 18
1.1 Artificial Autonomous Systems......Page 19
1.2 Neural Computation and Analog Integrated Circuits......Page 22
2.1 Image Brightness......Page 24
2.2 Correspondence Problem......Page 27
2.3 Optical Flow......Page 29
2.4.1 Explicit matching......Page 30
2.4.2 Implicit matching......Page 31
2.5.1 Global motion......Page 33
2.5.2 Local motion......Page 35
2.5.3 Perceptual bias......Page 39
2.6 Outline for a Visual Motion Perception System......Page 40
2.7 Review of a VLSI Implementations......Page 41
3.1 Associative Memory and Optimization......Page 48
3.2 Constraint Satisfaction Problems......Page 49
3.3 Winner-takes-all Networks......Page 50
3.3.1 Network architecture......Page 54
3.3.2 Global convergence and gain......Page 55
3.4 Resistive Network......Page 59
4.1 Model for Optical Flow Estimation......Page 62
4.1.1 Well-posed optimization problem......Page 65
4.1.2 Mechanical equivalent......Page 66
4.1.3 Smoothness and sparse data......Page 68
4.1.4 Probabilistic formulation......Page 69
4.2 Network Architecture......Page 71
4.2.1 Non-stationary optimization......Page 74
4.2.2 Network conductances......Page 75
4.3 Simulation Results for Natural Image Sequences......Page 82
4.4 Passive Non-linear Network Conductances......Page 88
4.5 Extended Recurrent Network Architectures......Page 92
4.5.1 Motion segmentation......Page 94
4.5.2 Attention and motion selection......Page 102
4.6 Remarks......Page 108
5.1 Implementation Substrate......Page 110
5.2 Phototransduction......Page 112
5.2.1 Logarithmic adaptive photoreceptor......Page 113
5.2.2 Robust brightness constancy constraint......Page 116
5.3.1 Temporal derivative circuits......Page 117
5.3.2 Spatial sampling......Page 121
5.4.1 Wide-linear-range multiplier......Page 126
5.4.2 Effective bias conductance......Page 138
5.4.3 Implementation of the smoothness constraint......Page 140
5.5 Layout......Page 141
6 Smooth Optical Flow Chip......Page 144
6.1 Response Characteristics......Page 145
6.1.1 Speed tuning......Page 146
6.1.3 Spatial frequency tuning......Page 150
6.1.4 Orientation tuning......Page 153
6.2 Intersection-of-constraints Solution......Page 154
6.3 Flow Field Estimation......Page 155
6.4 Device Mismatch......Page 159
6.4.1 Gradient offsets......Page 160
6.4.2 Variations across the array......Page 162
6.5 Processing Speed......Page 164
6.6 Applications......Page 167
6.6.1 Sensor modules for robotic applications......Page 168
6.6.2 Human–machine interface......Page 169
7.1 Motion Segmentation Chip......Page 174
7.1.1 Schematics of the motion segmentation pixel......Page 175
7.1.2 Experiments and results......Page 179
7.2 Motion Selection Chip......Page 184
7.2.1 Pixel schematics......Page 186
7.2.3 Experiments and results......Page 188
8.1 Human vs. Chip Perception......Page 194
8.1.1 Contrast-dependent speed perception......Page 195
8.1.2 Bias on perceived direction of motion......Page 196
8.1.3 Perceptual dynamics......Page 199
8.2 Computational Architecture......Page 200
8.3 Remarks......Page 205
A Variational Calculus......Page 208
B Simulation Methods......Page 212
C Transistors and Basic Circuits......Page 214
D Process Parameters and Chips Specifications......Page 224
References......Page 226
Index......Page 240
