Recently, many control methods for autonomous robots based on biological mechanisms have been studied. In particular, the concept of time to contact, called the tau-margin, which is studied in ecological psychology, has attracted considerable attention. In conventional studies, various approaches have been used for the determination of the tau-margin, and the tau-margin has been used for the timing control of mobile robots. However, in these studies, robots fitted with a light bulb have been used, and direct light from the bulb has been used to determine the tau-margin. It is impossible to apply conventional methods to robots that use indirect light for determining the tau-margin. Therefore, we have developed a compound eye sensor comprising photodiodes. We employ the framework of optical flow to detect objects, and determine the tau-margin from the apparent size of the detected objects and the temporal changes in size. To demonstrate the effectiveness of the developed compound eye sensor, we control a mobile robot using this sensor, and conduct experiments in which the robot has to avoid an oncoming object. The timing of the robot is controlled by using the tau-margin. We conducted experiments by using an actual robot, and found that the robot could avoid an oncoming object successfully. We concluded that it is possible to determine the tau-margin by using the compound eye sensor which we have developed.
[1]
Kazunori Hoshino.
Visual Systems Inspired by the Insect Compound Eye
,
2005
.
[2]
Kazuyuki Ito,et al.
Timing control of the mobile robot using tau-margin
,
2008,
Artificial Life and Robotics.
[3]
David N. Lee,et al.
Visual Timing of Interceptive Action
,
1985
.
[4]
Isao Shimoyama,et al.
A one-chip scanning retina with an integrated micromechanical scanning actuator
,
2001
.
[5]
Isao Shimoyama,et al.
Fabrication and Performance of Compound Eye Sensors with Scanning Receptor Arrays
,
2001
.
[6]
Michiaki Sekine,et al.
Compound-eye-type micro vision sensor with simple structure
,
2000,
Proceedings. 2000 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2000) (Cat. No.00CH37113).
[7]
Berthold K. P. Horn,et al.
Determining Optical Flow
,
1981,
Other Conferences.
[8]
Takashi Fuse,et al.
A Comparative Study on Gradient-Based Approaches for Optical Flow Estimation
,
2000
.