Velappa Ganapathy

Fuzzy and Neural controllers for acute obstacle avoidance in mobile robot navigation

Robot navigation is the technique to guide the mobile robot move towards the desired goal where dynamic and unknown environment is involved. The environment is distinguished by variable terrain and also certain objects which are known as obstacles that may block the movement of the robot in reaching the desired destination. Fuzzy Logic (FL) and Artificial Neural Network (ANN) are used to assist autonomous mobile robot move, learn the environment and reach the desired goal. This research study is focused on exploring the four combinations of training algorithms composed of FL and ANN that avoid acute obstacles in the environment. Path Remembering algorithm proposed in this paper will assist the mobile robot to come out from acute obstacles. Virtual wall building method also is proposed in order to prevent the mobile robot reentering the same acute obstacle once it has been turned away from the wall. MATLAB simulation is developed to verify and validate the algorithms before they are implemented in real time on Team AmigoBot™ robot. The results obtained from both simulation and actual application confirmed the flexibility and robustness of the controllers designed in avoiding acute obstacles and a comparison of all the four combinations of algorithms is done to find the best combination of algorithms to perform the required navigation to avoid acute obstacles.

Neural Q-Learning controller for mobile robot

In recent years, increasing trend in application of autonomous mobile robot worldwide has highlighted the importance of path planning controller in robotics-related fields, especially where dynamic and unknown environment is involved. Writing a good robot controller program can be a very time consuming process. It is inevitably wasting of resources and efforts if we have to rewrite the controller over and over again whenever there is emergence of changes in the environment. Reinforcement Learning (RL) algorithms and Artificial Neural Network (ANN) are used to assist autonomous mobile robot to learn in an unrecognized environment. This research study is focused on exploring integration of multi-layer neural network and Q-Learning as an online learning controller. Learning process is divided into two stages. In the initial stage the agent will map the environment through collecting state-action information according to the Q-Learning procedure. Second training process involves neural network training which will utilize the state-action information gathered in earlier phase as training samples. During final application of the controller, Q-Learning would be used as the primary navigating tool whereas the trained neural network will be employed when approximation is needed. MATLAB simulation was developed to verify the validity of the algorithm before it is real-time implemented on the real world using Team AmigoBot™ robot. The results obtained from both simulation and actual application confirmed on-spot learning ability of the controller accompanied with certain degree of flexibility and robustness.

Utilization of Webots and Khepera II as a platform for Neural Q-Learning controllers

The Webots commercial mobile robot simulation software and Khepera II miniature mobile robot have always been popular tools in research centers and universities. In this paper, the two items will be utilized as a platform for the investigation of Neural Q-Learning controllers. Webots remains as the primary simulation software where the simulated environment and robot are modeled. To cater for a wide variety of experiments, the simulation developed for the Khepera II is equipped with GUIs and various features. These functions allow the user to configure different environment and robot settings for different experiments. Then, the simulation is validated by comparing the behavior of the simulated and actual robot. As a result, a total of four controllers is proposed and tested on this platform. The designed controllers include both sensor and vision based controllers. These controllers are capable of exhibiting obstacle avoidance or wall following behaviors. In addition, an obstacle avoidance controller which is based on a combination of sensor and visual inputs via a fuzzy logic controller was proposed. Experimental results collected facilitate comparison and discussion of the algorithm and it further reveals that the mobile robot could successfully acquire the desired behavior.

Development of robot assisted stroke rehabilitation system of human upper limb

This project is focusing on the development of robot-assisted stroke rehabilitation system of human upper limb and hand elbow movements. Realizing the complexity of human upper limb, the study is limited to human upper limb, consisting of 3 glenohumeral joints (abduction-adduction, flexion-extension, and rotation) and elbow joint (flexionextension). In this paper the rehabilitation robotic system is proposed to assist a patient to train his hand movements to a desired position by considering other parameters such as speed and joint torque. In this system, each movement of the glenohumeral joints are scheduled based on the feedback signal obtained from the respective muscle groups through EMG signal interface. The kinematic model is proposed based on the literature survey and assumptions are made to reduce the complexity. Denavit-Hartenberg method is used for the positional analysis to determine the end position in 3D space while the Lagrange-Euler method is used for the dynamic analysis. A PID computed torque controller was designed by utilizing the equation of motion. The purpose of the controller is to compensate the dynamical imperfection and the presence of disturbance. The project is developed using MATLAB SimMechanics, in which human arm and rehabilitation robot is modeled.

Behaviour based mobile robot navigation technique for real world environments using fuzzy logic system

A key issue in the research of an autonomous robot is the design and development of the navigation technique that enables the robot to navigate in a real world environment. In this research, the focus is to develop the techniques and methods to: (1) design the individual behavior, coordination, and fusion of these behaviors using fuzzy logic expert system. Each behavior design is based on the situation context of applicability (SCA). (2) Design of the controller, which maps the sensors input to the motor output through fuzzy logic inference system. (3) Formulation of the decision-making processes by using fuzzy associative memory (FAM). (4) Obtain experimental results using simulation. The above methods and techniques are applied for active media pioneer robot and tested through simulation.

Fuzzy decision mechanism combined with neuro-fuzzy controller for behavior based robot navigation

This work describes the method to encode fuzzy sets, fuzzy rules and a procedure to perform fuzzy inference into an expert system for behavior based robot navigation. In this paper, we briefly present the design, coordination and fusion of the elementary behaviors for robot navigation using a fuzzy logic expert system. In this work the design of the behavior is based on regulatory control using fuzzy logic and the coordination and integration is defined by fuzzy rules, which define the context of applicability for each behavior. The complexity of robot behavior is reduced by breaking down robot behaviors into simple behaviors or units, and then combined with others to produce more complex actions. In this paper the decision making process of a few behaviors are illustrated specifically for an Active Media Pioneer Robot. The fuzzy logic decision mechanism, used here simplifies the design of the robotic controller and reduces the number of rules to be determined. The decision making process uses fuzzy logic for coordination, which provides a smooth transition between behaviors with a consequent smooth output response. In addition, the new behavior can be added or modified easily. Some of the experimental results are also shown for the obstacle avoidance, wall following and seek-goal behaviors.

Optical Character Recognition Program for Images of Printed Text using a Neural Network

In this paper we present a simple method using a self-organizing map neural network (SOM NN) which can be used for character recognition tasks. It describes the results of training a SOM NN to perform optical character recognition on images of printed characters. 49 features have been used to distinguish between 62 characters (both uppercase and lowercase letters of the English language and numerals). The implemented program recognizes text by analyzing an image file. The text to be recognized is currently limited to characters typed using the Verdana font type, bolded with a font size of 18. The program is capable of handling non-ideal images (noisy, colored text, rotated image). Recognition accuracy is consistently 100% for ideal images, but ranges between 80% -100% for non-ideal images.

Tabu search and simulated annealing algorithms for lot-streaming in two-machine flowshop

The objective of this paper is to propose and evaluate heuristic search algorithms for two-machine flow shop problem with multiple jobs requiring lot streaming that minimizes make span and total flow time. The two heuristic search algorithms evaluated Are tabu search algorithm (TABU) and simulated annealing (SA) algorithm. To create neighborhoods for SA, three perturbation schemes, viz. pair wise exchange, insertion and random insertion perturbation are used and the performance of these on the final schedule is also compared. A wide variety of data sets are randomly generated for comparative evaluation. The results indicate that TABU and SA perform almost similar for the objective of makespan and TABU performs better over SA for the objective of total flow time irrespective of the neighborhood creation schemes.

A novel modular framework for stereo vision

Though the theory of stereo vision is easy to grasp, the practice of computational stereo vision is much more difficult. Especially difficult is the general two-frame stereo vision problem. Many authors have published algorithms which, to a greater or lesser degrees, provide solutions to the two-frame stereo vision problem. The more recent algorithms mostly cite a landmark paper which presents a taxonomy of stereo vision consisting of some steps generally performed by stereo vision algorithms. In this paper, we propose a novel framework - the Modular Stereo Vision Framework - which describes general stereo vision problems in a more general manner than the preceding taxonomy. The central component of this framework is the cost volume, with the framework itself consisting of four stages, with each stage consisting of a number of modules. The mechanics of classifying existing and future methods as belonging to specific modules within this framework are described. Also, a discussion on the merits of the Modular Stereo Vision Framework and the preceding taxonomy is provided, focusing on the generality of the framework to current stereo algorithms and its utility for both developmental and classification purposes.

Mobile Robot Navigation using alpha level fuzzy logic system: Experimental investigations

This paper presents the issues associated with the mobile robot navigation and the establishment of a new technique to navigate the mobile robot in a real world environment. The issues discussed are (i) If multiple obstacles are appeared in the environment with equal distances as perceived from multiple sensors of robot, then the corresponding multiple obstacles are treated as a whole and the robot deviate from obstacles widely and avoid then reaching to the target. As a result of the wide deviation, the robot takes long time and long path to reach the target position. (ii) Behavior rule selection when multiple obstacles are appeared in the environment with equal distances from robot. The robot navigation with optimal path, time and rule selection are more important and critical task, whenever the mobile robots are engaged to search the lives in the event of natural disaster like earthquake etc. A new methodology is proposed and used for resolving the above issues and discussed in this paper. The mathematical aspect of resolving conflicts is presented the following section.