We propose a novel tactile sensor system, APCS (Air Pressure Collision detection System), to reduce harmful force caused by collision between manipulators and environments. The APCS has a simple structure composed of flexible tubes which can give it a shock absorbing function and pressure sensor for measuring the change of air pressure generated by the collision. It can detect collision at the initial moment of impact force generation by differentiating air pressure wave, and can detect contact (collision with low speed) by measuring the value of air pressure.The flexibility of tubes is effective to reduce impact force and it covers all over a manipulator including joints. In this paper, we describe the principle and condition of collision (contact) detection of the APCS, and confirm its effectiveness of harmful force reduction through simple collision experiments.Then, we install the APCS to 3 DOF manipulator's links and joints and realize safe motions to reduce harmful force in contact and collision situations. From the result of experiments, we confirmed that the APCS is an effective sensor system in reduction of harmful force for personal service manipulators.
This paper presents a recognition method of human daily-life action. The method utilizes hierarchical structure of actions and describes it as tree. We modelize actions by Continuous Hidden Markov Models which output time-series feature vectors extracted based on knowledge of human. In this method, recognition starts from the root, competes the likelihoods of child-nodes, chooses the maximum one as recognition result of the level, and goes to deeper level. The advantages of hierarchical recognition are: (1) recognition of various levels of abstraction, (2) simplification of low-level models, (3) response to novel data by decreasing degree of details. Experimental result shows that the method is able to recognize some basic human actions.
A control method for humanoid robots of mobile manipulation is proposed. A robot autonomously generates its whole body motion and steps so as to increase arm manipulability and robot stability, while performing various manipulation tasks with its hands; that enables dexterous and stable manipulation. An objective manipulation task is given as the desired trajectories of both hands, and external forces may be applied to the hands. For carrying out this task, the hands are always controlled along their desired trajectories by impedance control. Coordinating with this motion of the hands, the robot controls its body and legs so that the hands may be easy to perform the task. It determines the next foothold by evaluating manipulabilities of both arms, direction of stable region and moving direction. An evaluation function consisting of manipulabilities of both arms and static stability of the robot is defined, and the robot steps if the new double support state is better than the current one. Both in double and single support states, the robot controls its body pose using the support leg/legs so that this evaluation function may be optimal. The swing leg is moved to the next foothold in single support state. Various motions of humanoid robot“HRP-1”by the proposed method are simulated using dynamic simulator“OpenHRP”.
A robot hand imitating human hand is developed. The robot hand is driven by a new method proposed by authors using ultrasonic motors and elastic elements. The method utilizes restoring force of elastic element as driving power for grasping an object, so that the hand can perform the soft and stable grasping motion with no power supply. In addition, all the components oare placed inside the hand thanks to the ultrasonic motors with compact size and high torque at low speed. Applying the driving method to multi-DOF mechanism, a five-fingered robot hand is designed and implemented. It has equal number of joints and DOF to human hand, and it is also equal in size to the hand of average adult male. The performance of the robot finger is confirmed by fundamental driving experiments.
Finding the body in uninterpreted sensory data is one of the fundamental competences to construct the body representation that influences on adaptability of the robot to the changes in the environment and the robot body itself. The invariance of sensation seems a promising key information to find the self body since the sensory data are considered to be consistent in self body observation. To discriminate its body from non-body, the robot should complementarily utilize the invariance in multiple sensory data since single sensory data involve noise or a certain ambiguity occurred in the observation process. In this paper, we propose a method to discriminate body from non-body based on a conjecture about the distribution of the variance of sensations in terms of each observing posture. It can be approximated by a mixture of two Gaussian distributions for observing the body and non-body, respectively. After estimating the distribution by an EM algorithm, the robot can discriminate body from non-body by judging which distribution likely causes the variance of sensory data in the current observing posture. Experiments with real robots show the validity of the proposed method.
We propose an energy-based control method in order to improve the stability of continuous brachiation. The target continuous brachiation basically consists of two actions: a swing-back action and a locomotion action. Our strategy is to control the swing-back action so that the robot can pump up the required energy to grasp a target bar depending on the bar interval and design the locomotion action based on the symmetric motion of a pendulum in order to conserve the energy. In this paper, continuous brachiation on various uniform ladders is reported. The proposed controller is implemented on Gorilla Robot 111 developed as the multi-locomotion robot. The experimental results illustrate the validity of our control strategy.
This paper discusses jumping pattern generation for a serial link robot so that it can jump as high as possible under torque limitation. By applying GA for determining torque assignment, we obtain various jumping patterns with respect to the torque limitation under the mass of robot. With the increase of the torque limitation, double-leg based jump, single-leg based jump, and spring-type jump, are generated for high jump. Under the joint angular limitation, we also obtain an interesting solution where one end of the link is first lifted up and the other end finally kicks the ground strongly.
Movements of the upper limbs are complicated, various and indispensable for daily activities. It, therefore, is important for the aged to exercise to keep their upper limb function. Also when something is wrong with the upper limb function because of disease or disorder, rehabilitation along with medical treatment is needed to recover function. Application of robotics and virtual reality technology makes possible for new training methods and exercises on upper limb rehabilitation and for quantitative evaluations to enhance the qualitative effect of training. We involved in a project managed by NEDO (New Energy and Industrial Technology Development Organization as a semi-governmental organization under the Ministry of Economy, Trade and Industry, Japan), “Rehabilitation System for the Upper Limbs and Lower Limbs, ”and developed a 3-D exercise machine for upper limb (EMUL) . In this paper, We described softwares to control the moving trajectory of the patients hand by EMUL. They were able to be applied the system to practical rehabilitation training such as passive exercise training, active-assistive exercise training or active exercise training.