Doo-Sung Ahn

Formation control based on artificial intelligence for multi-agent coordination

In this paper, the authors propose a method of cooperative control based on an artificially intelligent system in a distributed autonomous robotic system. In general, a multi-agent behavior algorithm is simple and effective for small number of robots. However, as the number of robots increases, this becomes difficult to realize because a multi-robot behavior algorithm requires multiple constraints and goals in mobile robot navigation problems. As the solution to the above problem, the authors propose an architecture of a fuzzy-neuro system for obstacle avoidance. The controller adopts a simple reactive navigation strategy by combining repulsion from obstacles with attraction to a goal. Simulation results shows that the proposed strategy is effective for multi-robot to avoid obstacles while maintaining a formation.

Interacting mobile robots for tele-operation system using the World Wide Web

This paper discusses the interacting mobile robots for tele-operation system using the World Wide Web. The main components of this paper are user interface, networking, simulation, and robot control. The standard graphic user interface (GUI) is implemented using Java Programing language. The Web browser is used to integrate the virtual environment and the standard GUI (Java applet) in a single user interface. Users can access a dedicated WWW server and download the user interface. Java application has been developed to communicate and control the robot using RF communication. In addition to users direct control a fuzzy logic control is applied. We propose the use of FLC (fuzzy logic controllers) that incorporate expert knowledge in terms of linguistic rules. Experiments have shown that such fuzzy controlled mobile robots can have a better performance than a fine-tuned PID-controlled mobile robots.

Object contour following task based on integrated information of vision and force sensor

This paper presents object contour following task based on integrated information of both vision and force sensor. In this paper, first, using vision sensor (1) the distance to the contour from the center of the image and the angle of the contour is detected by the relative area method that measures the image features of interest directly from the thresholded image without any intermediate steps. (2) A set of contour edge points is detected by canny edge detector. Then, the curvature of contour is calculated by the tangent contour model. Second, force information is obtained from a strain-gauge force sensor. It constantly control contact force between tool and object of task. Finally, object contour following task is performed by a mobile robot with manipulator at a straight line section and a curved line section of object contour.

Control of Robot Manipulators Using Time-Delay Estimation and Fuzzy Logic Systems

A highly accurate model-free controller is proposed for trajectory tracking control of robot manipulators. The proposed controller incorporates time-delay estimation (TDE) to estimate and cancel continuous nonlinearities of robot dynamics, and exploits fuzzy logic systems to suppress the effect of the TDE error, which is due to discontinuous nonlinearities such as friction. To this end, integral sliding mode is defined using desired error dynamics, and a Mamdani-type fuzzy inference system is constructed. As a result, the proposed controller achieves the desired error dynamics well. Implementation of the proposed controller is easy because the design of the controller is intuitive and straightforward, and calculations of the complex robot dynamics are not required. The tracking performance of the proposed controller is verified experimentally using a 3-degree of freedom PUMAtype robot manipulator.

Mobile robot navigation algorithm using a vector-based topological map and virtual Jacobian

Abstract In this paper, a vector-based topological map is to be built consisting of available directions (moving directions) and the distances between nodes. This map represents the whole environment as a skeleton, using nodes similar to vectors independent of geometrical information of obstacles. To build a vector-based topological map, the following are developed: (a) acquisition of reliable range values using local minimum values, (b) fuzzy inference system (FIS) for extracting available directions, and (c) collision avoidance using virtual Jacobians. First, only local minimum values are used among the range of sensor values in order to decrease the effect of specular reflection and calculate available directions using the FIS. Specially, the number of consequent membership functions is determined as the number of obstacle orientations around a mobile robot, enabling the mobile robot to find available directions independently of the geometrical information of obstacles. Finally, to avoid collisions, a virtual Jacobian method is used because a vector-based topological map does not include the geometrical information about obstacles. The proposed algorithms have been verified in the simulation and implemented on a robot in the real environment.

Visual servoing system based on ANFIS (adaptive neuro fuzzy inference system)

Research in this visual servoing field in the past few decades has produced remarkable results, leading to many exciting expectations as well as new challenges. However, because of the complicated calculation of the inverse Jacobian, it is difficult to implement in real time. Therefore, instead of using the inverse Jacobian, this paper employs the ANFIS (Adaptive Neuro Fuzzy Inference System) approach for visual servo control of a robot manipulator. It is based on visual feedback and no prior information about the kinematics of robot and the camera calibration are unnecessary. Firstly, to efficiently control a manipulator, 3D space is divided into two 2D spaces. And then, we acquire training data from each 2D space and ANFIS is learned by the training data. We categorize the robot movement into two kinds of actions. That is, TOWARD action is performed, in the xy plane, by joint 1 and APPROACH action is performed, in the plane orthogonal to the xy plane, by joint 2 and joint 3. The time varying object can be tracked by controlling both actions in each plane and the simulation results show the validation of our approach.

Cooperative algorithm and group behavior in multirobot

In a multi-agent system, the action selection strategy is important for the cooperation and coordination of multi agents. However the overlap of actions selected individually by each robot makes the acquisition of cooperation behaviors less efficient. In addition to that, a complex and dynamic environment makes cooperation even more difficult. So in this paper, we propose a control algorithm which enables each robot to determine the action for the effective cooperation in multi-robot system. We employ a reinforcement learning in order to choose a proper action for each robot in its action subspace. In this paper, robot soccer system is adopted for the multi-robot environment. To play a soccer game, elementary actions such as shooting and passing must be provided. Q-learning, which is one of the popular methods for reinforcement learning, is used to determine what actions to take. Through simulations, the efficiency of own proposed algorithm is verified for the cooperation in multi robot system.