T. Nakamura and M. Asada. Stereo sketch: Stereo vision-based target reaching behavior acquisition with occlusion detection and avoidance. In Proceedings CVPR '96, pages 1314-- 1319. IEEE, San Francisco, CA, June 18-20, 1996.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....under different constraints on motion. Speedups of up to 400 are achieved without significant errors. 1 Introduction Stereo vision is one of the most common and robust vision algorithms used in mobile robot navigation. It has been used for mapping, localization and obstacle avoidance [8] 10] [11]. While performance of stereo has increased with growing computer power, the technique is still limited by the high computational cost of the algorithm. One reason for the lack of performance is that stereo is performed fully in each iteration of the perceptual cycle. This is wasteful because the ....

T. Nakamura and M. Asada. Stereo sketch: Stereo vision-based target reaching behavior acquisition with occlusion detection and avoidance. In Proceedings CVPR '96, pages 1314-- 1319. IEEE, San Francisco, CA, June 18-20, 1996.

.... researchers have proposed reinforcement learning algorithms [14] Using this approach and to reduce the convergence time, Connell et al. integrates knowledge such as the properties of the task, the sensor configurations and the environment [4] Asada et al. decomposes the input state space [20] [28] For simple robot tasks in a unknown and dynamic environment, researchers have applied evolution based learning algorithms to low level control architecture such as LISP like programming languages [17] 24] 22] production rules (classifier systems) 30] 5] 10] and neural networks [3] 7] ....

T. Nakamura and M. Asada. Stereo sketch: Stereo vision-based target reaching behavior acquisition with occlusion detection and avoidance. In Proceedings of IEEE International Conference on Robotics and Automation, pages 1314--1319. IEEE Press, 1996.

....coupled and inseparable [3] Human beings cannot see without eye movements, which suggests that actions signi cantly a ect visual processes and vice versa. There have been several attempts to build an autonomous agent based on a tight coupling between vision (and or other modalities) and actions [20,18,19]. The authors of these experiments contend that vision is not an isolated process but a component of a complicated system (physical agent) which interacts with its environment [4,21,10,8] This is a view quite di erent from the conventional computer vision approaches which have paid little ....

T. Nakamura and M. Asada. Stereo sketch: Stereo vision-based target reaching behavior acquisition with occlusion detection and avoidance. In Proc. of IEEE International Conference on Robotics and Automation, pp. 1314-1319, 1996.

....on situations. Since use of all possible sensory information seems impossible, selection of features obtained by the given capability for feature detection is more important. For example, behavior acquisition based on the visual motion cues [13] and based on stereo disparity and motion cues [14] have been proposed. A learning mechanism for selecting features from the sensory data processing available should be developed. ffl Coping with hidden states is another essential problem although we have not dealt with it here. This corresponds to coping with the temporal complexity of the ....

T. Nakamura and M. Asada. Stereo sketch: Stereo vision-based target reaching behavior acquisition with occlusion detection and avoidance. In Proc. of IEEE Int. Conf. on Robotics and Automation, pages 1314{ 1319, 1996.

....we, human beings, cannot see anything without the eye movements, which may suggest that actions signi cantly a ect the vision processes and vice versa. There have been several approaches which attempt to build an autonomous agent based on tight coupling of vision (and or other sensors) and actions [15, 13, 14]. They consider that vision is not an isolated process but a component of the complicated system (physical agent) which interacts with its environment [3, 16, 8] This is a quite di erent view from the conventional CV approaches that have not been paying attention to physical bodies. A typical ....

T. Nakamura and M. Asada. Stereo sketch: Stereo vision-based target reaching behavior acquisition with occlusion detection and avoidance. In Proc. of IEEE International Conference on Robotics and Automation, pp. 1314-1319, 1996.

....we, human beings, cannot see anything without the eye movements, which may suggest that actions signi cantly a ect the vision processes and vice versa. There have been several approaches which attempt to build an autonomous agent based on tight coupling of vision (and or other sensors) and actions [15, 13, 14]. They consider that vision is not an isolated process but a component of the complicated system (physical agent) which interacts with its environment [3, 16, 8] This is a quite di erent view from the conventional CV approaches that have not been paying attention to physical bodies. A typical ....

T. Nakamura and M. Asada. Stereo sketch: Stereo vision-based target reaching behavior acquisition with occlusion detection and avoidance. In Proc. of IEEE International Conference on Robotics and Automation, pp. 1314-1319, 1996.

....lighting conditions and occlusions is one of the central issues in applying this type of vision hardware. As long as the vision system can cope with the above issues, and capture the images of both the specified area (the target) and the ball, there might be no problem [ Nakamura and Asada, 1995, Nakamura and Asada, 1996 ] To prevent the agent from losing the target, and or the ball (in Level II and III, obstacles, too) an active vision system with panning and tilting motions seems preferable, but this makes the control system more complicated and introduces the spatial memory organization problem for keeping ....

....with a ball carrying (or passing shooting) behavior. One good strategy is assign the sensor roles in advance. For example, sonar and bumper sensors are used for obstacle avoidance while vision sensor is used for the target reaching. One can make the robot learn to assign the sensor roles [ Nakamura et al. 1996 ] 4.2.3. Action As described in section 2, total balance of the whole system is a key issue to the robot design. In order for the system to facilitate various kinds of behaviors, a more complicated mechanical system and its sophisticated control techniques are necessary. We should start with ....

T. Nakamura and M. Asada. Stereo sketch: Stereo vision-based target reaching behavior acquisition with occlusion detection and avoidance. In Proc. of IEEE Int. Conf. on Robotics and Automation, pages 1314--1319, 1996.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC