In this research, we propose a scheme to design the control system of a small unmanned rotorcraft flying in ground effect. Several H∞ controllers can be designed to stabilize the rotorcraft at desired altitudes respectively. However, one of these controllers can stabilize the rotorcraft only in the corresponding region limited small. We introduce several domains of attraction in the state space of the control system concerning the H∞ controllers. In order to perform the altitude control of the rotorcraft flying in various altitudes near ground, we present a switching control law with these H∞ controllers and domains of attraction. The feasibility of the presented scheme is verified through several experiments.
This paper describes vision-based guidance control of a small-scale unmanned helicopter. In previous research, helicopters were usually controlled by sensors such as GPSs and gyros. However, it is difficult to mount those sensors on small-scale helicopters that have little payload. Accordingly, small lightweight cameras were used to control the helicopter in this research. The applications of the cameras are for observation of the environment and the measurement of the position and attitude of the helicopter. The system of mounted cameras on the helicopter is useful for an indoor surveillance system using the helicopter as a mobile camera. The efficacy of the constructed system was verified through experimentations of automatic hovering control and guidance control.
In this paper, a novel approach is proposed to recover human body pose from 3D voxel data. The use of voxel data leads to viewpoint-free estimation, which benefits in that reconstruction of a training model is needless in different multi-camera arrangements. The chief advantage of our approach is speed, which enables real-time processing when capturing 8 VGA size images in 30 [fps] . Our approach is mainly based on an example-based approach. Human posture candidates are constructed beforehand, and the most appropriate posture candidate is selected per frame by comparing the similarity between 3D voxel data and posture candidates. Derivation of similarity is formulated by introducing a histogram-based feature vector that represents the 3D context of human body. In addition, a fast near-neighbor search metric is installed prior to the evaluation process, to reduce the computational cost and ensure real-time processing. Estimation stability is also improved by a motion graph, which adds a smoothing effect to the motion sequence. We demonstrate the effectiveness of our approach with experiments on both synthetic and real image sequences.
In this paper, the development of a robot which has a prototype of“sensor flesh”made of soft polyurethane foam and disributed three-axis force/torque sensors is presented. In order for robots to be working around human, it is important to enable them to be touched by humans, and to sense its contact states throughout the body. Several humanoids have“skin”with tactile sensors, but they cannot sense the distribution of force vectors. Their thin“skin”are not sufficiently soft to emurate the contact states as humans. In this paper, we took an approach of thick force vectors sensing exterior, which enables the emulation of contact states involving deformation of thick “flesh”. A comparative discussion about possible materials for the soft cover is described, and problem about movability of joints and thermal design are explored. Then, implementation of tactile sensing system is described. In the last part of this paper, the ability of sensors which react to force vectors and local deformation is shown.
Achieving energy-efficient dynamic walking has become one of the main subjects of research in the area of robotic biped locomotion. It has been clarified that approaches based on passive-dynamic walkers accomplish it. In general, however, passive dynamic walking is realized by only the legs and the effect of an upper body has not been clarified. Based on the observations, this paper deeply investigates what effects an upper body had on the performances and stability of dynamic biped locomotion. We first consider adding an upper body, which is introduced as a simple 1-link torso, by means of a bisecting hip mechanism so as not to destroy natural dynamics of the biped model. In the second, we analyze the robot's driving mechanism and apply underactuated virtual passive dynamic walking as a method for generating efficient dynamic gait. We confirm that highly efficient dynamic walking is realized with a specific resistance of 0.01, and investigate the effects of physical parameters of the upper body through numerical simulations.
This paper interprets an outfielder's action pursuing and catching a fly ball in terms of the behavior of a closed loop system by taking Chapman's hypothesis into the system connecting perceptions and actions. Through the analysis, we make it clear that the hypothetical trajectory Chapman showed was a special dynamic solution of the closed loop system. Moreover, it is shown that the proposed feedback control law makes it possible to generate a pursuing trajectory automatically that a fly ball can be caught at the right place at the right time. It is also shown that the pursuing trajectory gets closer to the one Chapman showed as a feedback gain increases. In addition, we compare the proposed feedback control law with Proportional Navigation (PN) which is the most common navigation technique for tracking moving objects, and show that the proposed method performs favorably over PN.
In general, either a position or a velocity command is utilized during manual teleoperation. However, when teleoperating a free-flying robot in space, it is difficult to ensure the commanded position or velocity because of the specific system dynamics. An acceleration command is much more appropriate. However, it is very difficult to handle the acceleration command for the operator. In this paper, a new predictive display technique named“Predictive Motion Display, PMD”is introduced to overcome the problem. The experimental results show that the proposed method decreases task completion time and enhances safety. In addition, the effectiveness is discussed based on a simple feedback control model.