Ja Gh

Multi Agent Cooperative robotics project for intelligent exploration of an unknown environment. Coded in Labview, and seamlessly connecting to a Java Intelligent Server via ethernet sockets and connected via wireless Bluetooth to mobile robots coded in C and Labview.

Multi Agent Mobile Robotics System functional architecture. It involves the seamless connection of agents programmed in Java, C and Labview across Bluetooth and ehternet socket connections.

NXT Robot Swarm - First Phase consisted of programming a single NXT robot with LabView, and testing all sensor, motor and communication capabilities with the Labview platform. Simple reactive behavior and remote control were implemented in this early phase.

NXT Robot Swarm, getting the first ultrasonic sensor mappings

Testing simultaneous communication protocol with 2 NXT Robots over USB, before testing over a wireless Bluetooth conbection.

NXT Robot Swarm - First NXT robot was programmed with Labview, another one with C, and the last one with Java. Regardless of the language, a common bluetooth messaging protocol was designed to communicate with the Labview Platform,

NXT Robot swarm: Small window (left): Laptop side code making a and remote connection with the NXT robot. Right and Top: Robot side code, executing on the NXT brick.

NXT Robot Swarm - Rotative ultrasonic sensor configuration

Support Vector Machine/Artificial Neural Network Machine learning program that I coded with Labview

Genetic Optimization Algorithm - Finds optimal solutions for complex problems, by processing data just as natural evolution happens: with populations, individuals, chromosomes, genes, mutation, elitism and reproduction of the fittest.

my original manuscript with the drafts of equations and data structures neede to implement the customizable "Napalm" Artificial Neural Network with LabView.

Implementation of Texture Segmentation Evaluator algorithm (NI Labview Vision Toolkit) - I wrote this code in NI Labview (Vision Toolkit) for the master's Artificial Vision class, to benchmark the effectivity of different texture segmentation algorithm outputs. It calculates an error metric based on the distance of missclasified pixels between the grount truth and the segmented image.

Texture Segmentation Evaluation - I made this Flash presentation for the Artificial Vision class of the master, its about proposing an original generic technique to evaluate different texture segmentation algorithms with a same metric.

Artificial Intelligence Goal Regression Non-Linear Planner (Railroad Domain) - I had to learn Java from scratch to make this project, and throw in lots of programming techniques and data structures to finish it: Linked List, Object Oriented Programming, Unbalanced Trees, Hash Functions, String Handling, File Write/Read, Heuristic regresion functions, and search algorithms, to implement this algorithm for the master's Artificial Intelligence class. The interface is graphic nice and friendly, it even draws states with characters (Full description continues on the next image)

Artificial Intelligence Goal Regression Non-Linear Planner - This text describes the project of the previous image.

Modelling an air bearing suspension as a Spring-Mass-Damper-System with LabView

Modelling an air bearing suspension as a Spring-Mass-Damper-System with LabView

Enter Your Description Here

Implementation of Texture Segmentation Evaluator algorithm (NI Labview Vision Toolkit) - I wrote this code in NI Labview (Vision Toolkit) for the master's Artificial Vision class, to benchmark the effectivity of different texture segmentation algorithm outputs. It calculates an error metric based on the distance of missclasified pixels between the grount truth and the segmented image.

"Napalm" Neural Network - for the Neural Network course of the Intelligent Systems master. Coded in LabView

"Napalm" Neural Network - for the Neural Network course of the Intelligent Systems master. Coded in LabView.

"Napalm" Neural Network - for the Neural Network course of the Intelligent Systems master. Coded in LabView.

Artificial Intelligence course: The topic of the day was the Graphplan algorithm. (Tarragona, Spain, 2012)

Calculating different Statistic descriptors for texture samples - Finding which statistic measure best distinguishes between the given texture samples for an image.

Histogram Plotter (NI Labview) - This simple code I wrote in NI Labview plots the histogram of a grayscale image.

Testing chosen Statistic measuer efectivity to segment textures with ImageJ - Finding which statistic measure best distinguishes between the given texture samples for an image.

Finding Optimal Texture Descriptor Statistics - Finding which statistic measure best distinguishes between the given texture samples for an image.