Autonomous vehicle navigation: A comparative study of classical logic and neural network technique
Today is an era in which humans need the help of automated machines to facilitate their hectic lives. Over the past few decades, robotics has been one of the most researched areas in the world. One area of this research is obstacle avoidance for autonomous vehicles. In order to avoid an obstacle, the system may include two important characteristics: obstacle detection and avoidance control. In this thesis a classical logic system and artificial neural network approach is presented. With the help of an Integrated Development Environment, a virtual robot was able to negotiate a maze and avoid obstacles according to the data gathered from its sensors. The latter approach was compared to an Artificial Neural Network (ANN) configuration, which proved to perform with successful results. This thesis demonstrates that the use of classical logic systems and ANN offers a good solution to the problem of obstacle-avoidance while negotiating a maze environment. ANN can be considered to be faster, once the neural network is trained, in response to obstacle-avoidance due to its massive parallel processing.^
Engineering, Electronics and Electrical
Flores, Javier Alejandro, "Autonomous vehicle navigation: A comparative study of classical logic and neural network technique" (2009). ETD Collection for University of Texas, El Paso. AAI1468295.