Ryota Kurozumi

A design of CMAC based intelligent PID controllers

PID control schemes have been widely used for most industrial processes which are represented by nonlinear systems. However, it is difficult to find an optimal set of PID gains. On the other hand, Cerebellar Model Articulation Controller (CMAC) has been proposed as one of artificial neural networks. This paper presents a new design scheme of intelligent PID controllers whose PID gains are generated by using CMACs. The newly proposed control scheme is numerically evaluated on a simulation example.

Implementation of an obstacle avoidance support system using adaptive and learning schemes on electric wheelchairs

With advance of an aging society, the persons who are physically handicapped have their respective needs about mobility assist with their living conditions. Moreover, operating an electric wheelchair indoors in confined spaces requires considerable skill. This paper presents an obstacle avoiding support system for an electric wheelchair, using reinforcement learning. The obstacle avoidance is semi-automatically supported by the minimum vector field histogram (MVFH) method. The MVFH modifies the user manipulation and assists the obstacle avoidance. In the proposed scheme, the modification rate is adjusted by the reinforcement learning according to the environment and the user condition. The newly proposed scheme is numerically evaluated on a simulation example.

Development of training equipment with an adaptive and learning mechanism using balloon actuator-sensor system

This paper proposes training equipment using a balloon actuator-sensor system (BASS) for persons unable to move their hands because of injury or disease. BASS is able to control the stiffness adaptively using an adaptive learning impedance controller. The pneumatic actuator has excellent compliance and flexibility, which is good for a human-mechanical system. However, it is also nonlinear, hence high precision control is difficult. Therefore, a CMAC-PID control scheme is installed. Finally, the BASS control performance is evaluated in a control experiment.

Path planning for mobile robots using an improved reinforcement learning scheme

The current method for establishing travel routes provides modeled environmental information. However, it is difficult to create an environment model for the environments in which mobile robot travel because the environment changes constantly due to the existence of moving objects, Including pedestrians. In this study, we propose a path planning system for mobile robots using reinforcement-learning systems and cerebellar model articulation controllers (CMACs). We selected the best travel route utilizing these reinforcement-learning systems. When a CMAC learns the value function of Q-learning, it improves learning speed by utilizing the generalizing action. CMACs enable us to reduce the time needed to select the best travel route. Using simulation and real robots, we performed a path-planning experiment. We report the results of simulation and experiment on traveling by online learning.

Experimental validation of an online adaptive and learning obstacle avoiding support system for the electric wheelchairs

With the advance of an aging society, people who are physically handicapped have specific needs concerning mobility assistance in relation to their respective living conditions. Moreover, operating an electric wheelchair indoors in confined spaces requires considerable skill. This paper presents an obstacle avoidance support system for an electric wheelchair, using reinforcement learning. The obstacle avoidance is semi-automatically supported by the Minimum Vector Field Histogram (MVFH) method. The MVFH modifies the user manipulation and assists the obstacle avoidance. In the proposed scheme, the modification rate is adjusted by reinforcement learning according to the environment and the user condition. The newly proposed scheme is numerically evaluated on a simulation example. Furthermore, the proposed scheme was applied to an experimental electric wheelchair, and the effectiveness of the proposed technique was verified in a real operating environment.

Attitude control using CMAC for electric wheelchairs equipped with hydraulic cylinder

We developed a posture control system for an electric wheelchair equipped with active suspension, using CMAC which is a neural network type of control. Testing of the posture control of the wheelchair, which considers a learning result as an output to the target value inputted on-line using the learning function of CMAC (cerebellar model arithmetic controller), was performed. In this paper, testing of the learning control which holds the posture horizontally both by the simulation and in the experiment by the system, is performed, and the validity of CMAC is verified.

Identification of Visually Impaired Person with Deep Learning

The purpose of this study is to identify visually impaired persons by analyzing still pictures of walking of a visually impaired person and that of a healthy person using deep learning. Still images of walking are taking still pictures from video images. Shoot from sideways and diagonally with two video cameras. The number of images (with 1000 or 2000) and the dropout (three, two, or one time) was changed and analyzed. Because the study focused on only visually impaired persons (totally blind persons) and the healthy person’s study machines of two patterns in the experiment, a correct answer rate of 99.9% for every 2000 images and 2 times of the dropout number was obtained.

Development of a postural supporting device using an adaptive and learning balloon actuator-sensor system

In this paper, a postural supporting device using a balloon actuator-sensor system (BASS) is proposed for the person with disability who is unable to move their body because of injury or disease. The BASS is able to control the position and stiffness adaptively using an adaptive learning impedance controller. The pneumatic actuator has excellent compliance and flexibility, therefore that is good for the human-mechanical system. However, it also has nonlinearity, hence high precision control is difficult. Therefore, in this paper, the CMAC-PID control scheme is installed. The control performance of the BASS is evaluated by the control experiment.

Development of a support system avoiding obstacles for electric wheelchairs using adaptive and learning schemes

Abstract With advance of an aging society, the persons who are physically handicapped have their respective needs about mobility assist with their living conditions. Moreover, operating an electric wheelchair indoors in confined spaces requires considerable skill. This paper presents an obstacle avoiding support system for an electric wheelchair, using reinforcement learning. The obstacle avoidance is semi-automatically supported by the Minimum Vector Field Histogram (MVFH) method. The MVFH modifies the user manipulation and assists the obstacle avoidance. In the proposed scheme, the modification rate is adjusted by the reinforcement learning according to the environment and the user condition. The newly proposed scheme is numerically evaluated on a simulation example.

A Design of PID Controllers Fused CMACS with Neural Networks

Abstract Several neural-net based PID controllers have been proposed for nonlinear process systems. However, they have been not so widely used in process industries due to the considerably computational cost. This paper presents a new intelligent PID tuning scheme, whose PID tuner is constructed by the fusional structure of a cerebellar model articulation controller and a neural network. This PID tuner gives us the higher learning efficiency which has not been realized by the conventional neural-net based controllers, and it enables us to tune PID gains in an on-line manner. The behaviour of the proposed scheme is examined by a simulation example for a chemical reactor model.