2007 Fiscal Year Final Research Report Summary
Robot Eye Control System and Its 3D Image Recognition Based on Optical Physiology
| Project/Area Number |
17300060
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (B)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Perception information processing/Intelligent robotics
|
|---|
| Research Institution | Tokyo Institute of Technology |
|---|
Principal Investigator |
ZHANG Xiaolin Tokyo Institute of Technology, Precision and Intelligence Laboratory, Associate Professor (60280971)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
SATO Makoto Tokyo Institute of TechnologyTokyo Institute of Technology, Precision and Intelligence Laboratory, Professor (50114872)
KOIKE Yasuharu Tokyo Institute of TechnologyTokyo Institute of Technology, Precision and Intelligence Laboratory, Associate Professor (10302978)
|
|---|
| Project Period (FY) |
2005 – 2007
|
| Keywords | Image Processing / System Control Engineering / Biology Engineering / Information Technology / Brain & Neuron / Robot Vision / Bio-robotics / Neural Network |
|---|
| Research Abstract |
(1) A binocular motor system model based on human neural pathways of the binocular motor system was described and a camera movement control system using the model was presented. Although the complexity of real neural networks necessitated several restrictions on the proposed model, the system exhibited several characteristics specific to human-eye movement, such as "same point gazing tendency."
(2) A new mathematical model which shows the short-term signal processing functions of a neuron was proposed. This research demonstrates that the membrane of a neuron acts as an adder, and the synapses are signal processing units for the input pulses. The adder and signal processing units are related by the total postsynaptic currents of all synapses, i.e. the pulse frequency on the membrane is in proportion to the total postsynaptic currents, and the postsynaptic current of each synapse can be obtained through convolution integration by a one pulse postsynaptic current function and the pulse fre
… More
quency function of the synapse. This allows us to reach an interesting conclusion that the transfer function of a neuron, from the synapses to the axon, is the sum of the Laplace transform of the one pulse postsynaptic current function of each synapse.
(3) A binocular camera control system using only analog devices was developed. This analog robot demonstrated better dynamic performance than a similarly structured digital system.
(4) An active camera system based on hawk eye structure and neuron system was proposed. The system, for example, can be used for search and rescue helicopters.
(5) A binocular security system was developed. The system established that if a set of two active cameras (PTZ camera) uses the binocular model, not only can we easily obtain detailed information, such as 3D data of the imaged person, but can also improve the dynamic performance for target pursuit.
(6) Image processing algorithms for robot vision was proposed. The algorithms include the moving target search algorithms for saccade, and known target detection algorithms for smooth pursuit. Less
|
Research Products
(19 results)
-
-
-
-
-
-
-
[Presentation] 神経細胞の数学モデル2007
Author(s)
張 暁林
Organizer
第2回医工連携・バイオメカニクス国際シンポジウム
Place of Presentation
東京
Year and Date
2007-10-22
Description
「研究成果報告書概要(和文)」より
-
-
-
-
-
-
-
-
-
-
-
-
