Seiji Uozumi

A quantization method for haptic data lossy compression

In this paper, a quantization scheme for haptic data compression is proposed and the results of its application to a motion copying system are described. The data compression of image and audio data has been researched a lot. However, the data compression of haptic data has not been much researched yet although the amount of haptic data is large in general. The discrete cosine transform (DCT) is used for the lossy compression of haptic data in this paper. A new quantization scheme using characteristics of DCT is proposed. By using a proposed method, a larger signal-to-noise ratio compared to conventional method is achieved. In the experiment, the validity of the proposed method was verified.

Stability analysis of bilateral control by considering environmental impedance and modeling error

This paper analyzes the stability of acceleration-based bilateral control (ABC) using a disturbance observer (DOB) and a reaction force observer (RFOB) by considering the effect of environmental impedance and modeling error. ABC is one of the control method to enable highly precise transmission of haptic sensation between master and slave robots. Conventionally, ABC is analyzed by using slave side environmental impedance and ideal acceleration control and load force estimation. However, the stability of ABC is affected by both master and slave side impedance, cutoff frequency of DOB and RFOB, and modeling error. Therefore, this paper proposes the approach for analyzing stability of ABC by considering these factors which affect the stability of the system. The validity of proposed approach is verified by simulation and experimental results.

Compensation of out-of-order delivery using Kalman Filter for teleoperation system

This paper proposes a new method to compensate out-of-order delivery in teleoperation systems with network. Out-of-order packets generally deteriorate the performance of the system. Re-order buffer is one of the methods to compensate out-of-order delivery. However, designing of the size of re-order buffer is difficult and this buffer makes additional time-delay to wait for the delayed packets. This time-delay degrades the real-time property of the teleoperation system. Therefore, this paper proposes a new compensation method using Kalman Filter. This method considers out-of-order delivery as noise in the system and compensates it in the spatial domain in real time. The experimental results show the validity of the proposed method.

Motion control of tendon-driven rotary actuator using Twist Drive system

This paper proposes a method to achieve high-precision control for tendon driven system by using the Twist Drive. Twist Drive is one of transmission system which is consisted a pair of strings and rotary motor. This transmission system converts torque into a pulling force by using the pair of strings twisted on each other. However, it cannot realize pushing motion. Therefore the Twist Drive needs to be combined with an other actuation system for the realization of pushing motion. The problem is solved by tendon drive using Twist Drive system in this paper. High-precision control is achieved by modal-space disturbance observer. Modal-space disturbance observer is designed to realize the acceleration based control. It also compensates the modal interference. Furthermore work space observer is able to estimate the force in the work space without any force sensors. The proposed method was applied to rotational angle control and torque control of the tendon driven rotary actuator. The effect of the proposal is verified through experiments.

Consideration on function mode design for motion construction

With the development of robot technology, the activities of robots are expected to enter the human environment. Robots in human environment need the large-scale system with multi-degrees-of freedom for environmental adaptation. Therefore, it is necessary to design a large-scale system efficiently and simply. Tsuji et al. proposed function based control method. This method decouples a large control system into small independent components called “function” extracted by the “function mode”. However, design method of function mode has not been clarified. Therefore, this paper considers the design method of function modes by fragmentation of the function mode. By using this method, independent control of complex tasks can be achieved. This method can be a guide for designing complex functions and improve the adaptability to a designers intention and unknown environment. Proposed method is implemented in hybrid control system using three robot arms and each robot arm has two degrees of freedom. The validity of proposed method is verified by the experimental results.

Disturbance suppression method for position-sensorless motion control of DC brushed motor

This paper proposes position-sensorless disturbance observer (PSDOB) for DC brushed motor (abbreviated as DC motor later). The aim of this method is to suppress disturbance for position-sensorless motion control of DC motor. At first, block diagram of motor system is transformed and disturbance torque is defined as modeling errors between real motor model and nominal model. Next, disturbance torque is estimated from current reference, voltage reference, and measured current. Finally, estimated disturbance torque is fed back to suppress the disturbance torque. Therefore, robust acceleration control can be realized by utilizing PSDOB. The utility of the proposed method is experimentally verified.

Verification of Twist Drive system and its application to haptic robot hand

The objective of this study is to propose the master-slave tendon-driven haptic robot hand system using Twist Drive. Twist Drive is one of transmission systems which is consisted of a pair of strings and rotary motor. This transmission system converts torque into a high pulling force by using a pair of strings twisted on each other. In this paper, Twist Drive system is verified at first. Next, Twist Drive system is implemented to the master-slave tendon-driven robot hand system to realize the downsizing of the actuator part. In order to realize haptic transmission, a method which can achieve 4ch-bilateral control and tendon-driven control, simultaneously is implemented. Here, to achieve the tendon-driven control, a method to keep the wire tension using compliance control based on hybrid angle is realized. The effect of the proposed system is verified through experiments.

Stereo vision based robot navigation system using modulated potential field for implant surgery

Image navigation systems and robotics are leading technologies to improve the accuracy of dental surgery. In this paper, stereo vision based navigation system is implemented to the three degrees of freedom implant surgery assistant robot. The aim of this research is to achieve cooperative guide system for teaching the information of “where to start cutting”. To achieve this aim, modulated potential field based on Lennard-Jones potential field is proposed for realizing following three requested performances. (1) Surgeon can freely operate the manipulator when it far from the destination. (2) The manipulator will apply adjusting force to converge correctly when approaching to the destination. (3) The manipulator will apply adjusting force for collision avoidance when passing the destination. This paper uses stereo camera to detect the destination. The validity of the proposed method is confirmed through experiment using three degrees of freedom parallel link manipulator.