| 白倉 尚貴 | D, 中間発表 | ロボティクス | 小笠原 司, 杉本 謙二, 高松 淳, 丁 明, Gustavo Garcia |
|
title: Manipulation with Physical Interaction using a Drone
abstract: Manipulation with drone is expected to be used for work in high places and underwater. In previous research, grasping and manipulating objects has been achieved with manipulators mounted on flying drones. Moreover, the manipulating objects with underwater drone have been also succeeded by remote manual control. These drone systems can be developed from not only full scratch development but also existing open-source flight controllers. However, available functions are limited because existing flight controllers are not designed for use in control with physical interaction to environments such as a push motion. In this study, the system is constructed separately divided into drone control and manipulation control, and the manipulation with physical interaction is realized using the available part of the existing flight controller. This study focuses on the pushing motion as manipulation with physical interaction and constructs the pushing motion system with a flying and an underwater drone. The effectiveness of the proposed system is validated with simulation and real drone. language of the presentation: Japanese | |||
| URIGUEN ELJURI PEDRO MIGUEL | D, 中間発表 | ロボティクス | 小笠原 司, 杉本 謙二, 高松 淳, 丁 明, Gustavo Garcia |
|
title: Combining symbolic and motion planning to execute rearranging tasks in a daily-life environments with a robot
abstract: I propose to combine the symbolic and motion planner to execute a rearranging task. In a conventional approach, the symbolic and motion planner are separated. Which creates a problem when the instructions can not be executed by the robot. This creates a loop where the symbolic planner is called multiple times, because of errors in the motion planner. Our approach solves this problem. At the moment of generating the instructions, the symbolic planner verifies with the motion planner if it is possible to execute the instructions. I propose to use a database of reachable poses of the end effector together with a state checker that confirms if the instructions are valid. In this presentation, I will report the results about combining the symbolic and motion planner and the difference with the conventional approach. language of the presentation:English | |||
| 永谷 智貴 | M, 2回目発表 | ロボティクス | 小笠原 司, 杉本 謙二, 高松 淳, Gustavo Garcia |
|
title: Optimizing cameras placement for human safety to a manipulator
abstract: I propose cameras placement optimization method which considers human and robot behavior patterns in addition to the cameras' field of view. The convenience store clerk robots require both safety and efficiency. Previous research related to this has focused on generating the actions that the robot should take. However, these systems assume a fully monitored human state but occlusion occurs in the real environment. Therefore, it is necessary to optimize the placement of the cameras. I use existence probability maps generated from human behavior patterns as constraints for optimization. I think that this method allows monitoring human state efficiently with a smaller number of cameras language of the presentation: Japanese | |||
| 久保 基 | M, 2回目発表 | 知能コミュニケーション | 中村 哲, 安本 慶一, 国田 勝行(BS), 田中 宏季 |
|
title: Information Extraction from Tourism Behavior Data using Tensor Factorization
abstract: In order to promote inbound tourism, it is essential to first analyze the behaviors and destinations of tourists and understand their trends and factors. Tourists' actions include various factors and information, and it is not easy to analyze them all at once. This time, we configured the 4th floor tensor (time zone / location / day of the week / month) from the multi-dimensional tourism behavior log data obtained from the smartphone app. The analysis was performed using tensor factorization. In this presentation, we report the execution time and estimation error for the main methods of tensor decomposition, and take up some examples of behavior analysis of foreign tourists obtained by each method. language of the presentation: Japanese | |||
| 古志 将樹 | M, 2回目発表 | サイバネティクス・リアリティ工学 | 清川 清, 安本 慶一, 酒田 信親 |
|
title: Augmented Concentration: Concentration Improvement by Visual Noise
Reduction with a Video See-Through HMD
abstract: We propose a concentration improvement technique by using a video see-through head mounted display (HMD). Our technique reduces the visual noise or disturbing visual stimulus, such as moving objects near the central visual field, by lowering its visual saliency in real-time. Earphones and noise cancelling headphones are often used to shutdown auditory noise from surroundings when we need to concentrate on the job. Several studies have proven the effectiveness of such noise reduction on improving the concentration and learning efficiencies. We apply this analogy to the visual noise and propose a visual noise reduction HMD to improve concentration. In this article, we report on two preliminary user studies we conducted and future plan based on this study. language of the presentation: Japanese | |||
| 大久保 至道 | M, 2回目発表 | サイバネティクス・リアリティ工学 | 清川 清, 安本 慶一, 酒田 信親 |
|
title: Behavior Modification by Feedback of Mirror Self-image with Impression Manipulation
abstract: This study aims to create a visual feedback system to encourage ideal self-behaviors by objective self-awareness. In the previous experiment, it was shown that the feedback of mirror self-image which is little bit delayed raises the consciousness that tries to bring the real self closer to the ideal self. However, it was suggested that the awareness of self-affirmation might be lowered by recognizing of the difference between the real self and the ideal self. In this paper, we adopted the visual feedback which has two functions, delaying and impression manipulation, to encourage ideal self-behaviors without lowering the self-affirmation. The effect of this feedback has been verified by an interview test. language of the presentation: Japanese | |||