In this paper, we focus on the concept of time to contact and propose a simple sensor system for detecting the time to contact, and apply it to autonomous control of a flying robot. To discuss the effectiveness of the proposed sensor system and a controller using the sensor system, we develop a flying robot simply without CPU(s) or high spec sensors. We employ only 5 CdS cells and simple analogue circuits for controlling it, and we realize behaviour for chasing a moving light bulb with stable flying. To demonstrate the effectiveness of the proposed sensor system and flying robot, simulations and experiments have been conducted and chasing behaviours have been realized. I. INTRODUCTIONRecently, robots which operate autonomously in the real environment have attracted considerable attentions, and autonomous control methods inspired by animals have applied to various robots.One of the most important points of the control methods of animals is simplicity and robustness. For example, birds and insects can fly freely while avoiding obstacles and can catch a prey by suitable timing, in spite of the fact that they do not have sensors for detecting long range distance and their positions. The reasons why they can fly so adaptively have not been cleared completely yet. However, it is considered that information of distance or position is not always required for the adaptive behaviour [1-3]. In particular, the concept of time to contact, called tau-margin, is proposed in ecological psychology, and it is considered that time to contact is employed for expressing positional relation among own body and obstacles [1][2][3].In conventional studies in robotics, various autonomous flying robots have been proposed. However, they have CPU(s) and many sensors, which are distance sensors, gyro system, cameras and so on, so the system is complex and simple control framework like animals has not been realized [4][5][6][7][8][9][10].In this paper, we focus on the concept of tau-margin and propose a simple sensor system for detecting the time to contact and apply it to autonomous control of a flying robot. To discuss the effectiveness of the proposed sensor system and a controller using the proposed sensor system, we develop a flying robot very simply without CPU(s) or distance sensors. We employ only 5 CdS cells and simple analogue circuits for controlling it, and we realize behaviour for chasing a moving light bulb with stable flying.