Ming-Yuan Shieh

Design of an intelligent hospital service robot and its applications

The paper presents the design concepts of an intelligent hospital service robot (IHSR). The IHSR aims at saving human resources and also improving the hospital service. The navigation path planning can be concluded by an information decision making system. It is decided from the distribution of people in the hospital. For obstacle avoidance, the IHSR mounts a CCD camera and its attachments to determine the relative directions of obstacles. There are nine ultrasonic sensors installed on the IHSR to detect the distances between the IHSR and obstacles. An integrated fuzzy controller is proposed as the collision-free navigation control system. In which, there are two fuzzy subcontrollers designed for objective navigation and collision-free control respectively. Simulation results show that the IHSR is indeed able to accomplish navigation control in hospital environment.

ANFIS based Controller Design for Biped Robots

This paper proposes a design method of biped locomotion controller based on ANFISs (adaptive neuro-fuzzy inference systems). In which, there is an ANFIS assigned as the system identifier to determine the model of a biped robot firstly. And then the motion controller, which is a combination of ten ANFISs, aims to perform the biped locomotion controls. One can adopt the moving velocities of biped robot, trajectory of ZMP (Zero Moment Point) and the tilt angles as the inputs of the controller, and the outputs will be the next angles of joints. As the simulation results show, the proposed controller can generate a stable walking cycle for a biped robot. Moreover, the experimental results demonstrate that the performance of proposed controller is satisfactory whenever the robot stays in different postures or moves on a rugged surface.

Development and Implementation of an Artificial Neural Network based Controller for Gait Balance of a Biped Robot

This paper proposes a gait balancing controller for a biped robot. The controller was designed based on a back- propagation artificial neural network (BPANN). Because of the on-line learning ability of BPANN, it allows the controller to generate advisable corrections of each joint for robotic balance according to the signals of gyroscopes. It results in a balanced locomotion whenever the biped robot is walking or standing upon the uneven terrain. There are four experiments of robotic locomotion in different postures and grounds applied to verify whether the controls of robotic gait balance are satisfactory.

Design of an integrated grey-fuzzy PID controller and its application to non-minimum phase systems

An integrated control system which includes the fuzzy PID controller and the grey predictor is investigated on the approaches to non-minimum phase (NMP) systems. Since the grey predictor can predict the undershoot phenomena of NMP systems, it offers the fuzzy PID controller suitable prediction to generate the appropriate control action on undershoot elimination. In addition, the grey predictor provides suitable compensation by a proper step size. Therefore, we design another fuzzy controller to decide the step size of the grey predictor such that the compensation is suitable. The controller design and computer simulations are examined under several different conditions. The results demonstrate that the proposed controller can eliminate the undershoot phenomena of NMP systems effectively.

The optimization of the application of fuzzy ant colony algorithm in soccer robot

This article provides a theory which is based on the Fuzzy Ant Colony Optimization, and then uses this theory to design an optical speed for the football robots, and then we also apply Ant Colony Optimization to design its routes to avoid obstacles. Afterward, we add Generalized Predictive Control to predict the position which its target might appear at the next sampling time. Our main experiment platform is a simulation machine for FIRA five-to-five soccer robots and MATLAB SIMULATION.

Two-stage fuzzy navigation control of a vision-based intelligent shopping service robot

The paper describes a two-stage fuzzy navigation controller for a vision-based intelligent shopping service robot (ISSR). The navigation strategy of the robot merges two distinct situations. The global navigation strategy concludes the traveling path from the information such as the task styles assigned to the ISSR and the people distribution in the supermarket. The local navigation control consists of two stages. The obstacle avoidance stage is based on visual feedback and sonar sensory information to detect obstacles, in which, an integrated fuzzy controller is designed to offer a smoother navigation. The guiding control stage provides objective guidance to avoid that the robot travels adrift in the supermarket. The simulation results show the feasibility of the proposed navigation control system.

Adaptive classification and strategy making system for android soccer games

This paper proposes an adaptive classification and strategy decision-making scheme which aims to integrate the image object identification and classification, the path planning, obstacle avoidance, and intelligent decision making technologies for android soccer games. It classifies the color block labels of the robots and the ball by using a statistical algorithm of kernel discriminant analysis firstly, and then determines the locations of the blocks based on image morphology operations. These color identification results will be offered the android soccer system to determine the detail object data of the robots and the ball for soccer game controls. Based on the training items of three kinds of AndroSot Challenges in 2011 FIRA World Cup, the proposed system intends to make role assignments of robots as attacker, defender, or goalkeeper, and to make decision of robot movements by formation strategies. Finally, the proposed scheme is implemented by actual competitions in 2011 FIRA AndroSot Game. The experimental results show the feasibility and the adaptability of the proposed system.

Gait detection based stable locomotion control system for biped robots

It is a challenge to maintain a steady and stable locomotion when a biped robot navigates an uneven surface or a step. Firstly it needs to detect the gait of the robot and related environmental objectives, and then to perform appropriate controls of stable locomotion. In this paper, a new type of sensing module with flexible force is designed to be mounted under the robotic sole to measure the foot pressure map which is essential for the analysis of zero moment point (ZMP) to assist the biped robot to detect the gait conditions. In addition, a neuro-fuzzy control scheme is also proposed to control the ZMP trajectories and relative balancing by integrating the data of a 3-axis gyroscope to adaptive posture. The experimental results show that the designed control system has improved the biped robot adaptive ability to overcome landform changes and the stability of locomotion while standing or walking on uneven terrain.

Design and implementation of an interactive nurse robot

In this paper, an interactive nurse robot is developed which has worked successfully in home environment. The mobile robot, driven by two motors/wheels, consists of image processing, speech recognition, semantic analysis, local/remote interaction, and autonomous mobile components to provide nursing functions of surveillance, interaction, accompanying, patrol, and urgent announcement. Since the proposed robot is desired to service children and/or elders in home, its appearance and actions should be loveable and safe. The experimental results reveal the interactive nurse robot really reaches the goals of attracting and looking after children.

SOMNN based object segmentation for soccer robot system using adaptive background generation

This paper proposes a SOMNN based object segmentation algorithm to detect movements in android soccer robot games by combining adaptive background generation. Since omni-directional images of the game field need to be performed proper object segmentation for effective motion recognition. The SOMNN based algorithm aims to provide adaptive segmentation even under variant brightness. Moreover, the proposed scheme combining temporal difference and adaptive background generation could improve the detection of salient motions. The experimental results in real 3 on 3 android soccer robot games demonstrate the feasibility of the proposed scheme.