Bacciga Ambrogio, Naliato Renato. Recent Advances in Artificial Intelligence Research

Nova Science, 2013. — 118 p. — ISBN10: 1628088079; ISBN13: 978-1628088076.Artificial intelligence (AI) is 'the study and design of intelligent agents', where an intelligent agent is a system that perceives its environment and takes actions that maximise its chances of success. In this book, the authors present recent advances in the study of artificial intelligence with topics that include a Twitter specific lexicon for sentiment analysis; hybrid unsupervised-supervised artificial neural networks for modeling activated sludge wastewater treatment plants; fast and visible trajectory planning in 3D urban environments based on local point clouds data; and 'smart brakes' - the neuro-genetic optimization of brake actuation pressure.Artificial intelligence (AI) is "the study and design of intelligent agents", where an intelligent agent is a system that perceives its environment and takes actions that maximize its chances of success. In this book, the authors present recent advances in the study of artificial intelligence with topics that include a Twitter specific lexicon for sentiment analysis; hybrid unsupervised-supervised artificial neural networks for modeling activated sludge wastewater treatment plants; fast and visible trajectory planning in 3D urban environments based on local point clouds data; and “smart brakes” - the neuro-genetic optimization of brake actuation pressure.Chapter 1 – Twitter messages are increasingly used to determine consumer sentiment towards a brand. The existing literature on Twitter sentiment analysis uses various feature sets and methods, many of which are adapted from more traditional text classification problems.
In this chapter, the authors introduce an approach to supervised feature reduction using ngrams and statistical analysis to develop a Twitter specific lexicon for sentiment analysis. The authors augment this reduced Twitter specific lexicon with brand specific terms for brandrelated tweets. The authors show that the reduced lexicon set, while significantly smaller, reduces modeling complexity, maintains a high degree of coverage over the Twitter corpus, and yields improved sentiment classification accuracy. To demonstrate the effectiveness of the devised Twitter specific lexicon compared to a traditional sentiment lexicon, the authors develop comparable sentiment classification models using SVM. The authors show that the Twitter specific lexicon is significantly more effective in terms of classification recall and accuracy metrics.
Chapter 2 – Mathematical modeling of wastewater treatment plant process is important for improving its treatment efficiency and thus the quality of the effluent released into the receiving water body. However, due to the highly complex and non-linear characteristics of this biological system, traditional mathematical modeling of the treatment process has remained a challenge. This work presents a hybrid modeling strategy based on the Kohonen Self Organising Map (KSOM) and feed-forward, back-propagation artificial neural networks (BP-ANN) for modeling the activated sludge wastewater treatment plant. The hybrid approach involved a 2-stage process: firstly, the KSOM was used for data preparation, visualisation of high dimensional data and features extraction; and secondly, these features were then used for the training and validation of the BP-ANN. Comparison of this hybrid modeling approach against the straight modeling of the original raw data using BP-ANN showed that the hybrid approach resulted in much more improved model performance.The study demonstrated that the hybrid modeling strategy offers viable, flexible and robust modeling methodology for effectively handling noisy data for environmental systems modeling.Chapter 3 – In this paper the authors present an efficient and fast visible trajectory planning for unmanned vehicles in a 3D urban environment based on local point clouds data. The authors trajectory planning method is based on a two-step visibility analysis in 3D urban environments using predicted visibility from point clouds data. The first step in the authors unique concept is to extract basic geometric shapes. The authors focus on three basic geometric shapes from point clouds in urban scenes: planes, cylinders and spheres, extracting these geometric shapes using efficient RANSAC algorithms with a high success rate of detection. The second step is a prediction of these geometric entities in the next time step, formulated as states vectors in a dynamic system using Kalman Filter (KF). The authors planner is based on the optimal time horizon concept as a leading feature for their greedy search method for making their local planner safer. The authors demonstrate their visibility and trajectory planning method in simulations, showing predicted trajectory planning in 3D urban environments based on real LiDAR point clouds data.
Chapter 4 – Many systems today that need modeling and optimization are non-linear systems or systems whose behavior is strongly influenced by their previous and current state. The most important automotive system having these characteristics is a braking system. The main purpose of braking systems is to control braking torques, allowing a vehicle to decelerate in an optimum manner while maintaining directional stability. The demands imposed on a braking system, over a wide range of operating conditions, are complex. It is especially related to the brakes. The brakes are supposed to provide high and stable values of braking torque over different operating conditions determined by synergistic influence of the actuation pressure and/or sliding speed and/or the brake interface temperature for the specific friction pair characteristics. Since the driver obtains an important feedback of vehicle dynamics and its braking capabilities depending on a brake performance change, it represents an important aspect of a vehicle performance and its quality of use. Sensitivity of braking torque versus influence of the friction couple interaction, under different braking conditions,is one of the most important properties of the disc brake. In this chapter, possibilities for an intelligent dynamic optimization of the brake performance have been investigated. The hybrid neuro-genetic optimization model has been developed for dynamic control and optimization of the disc brake actuation pressure during a braking cycle. This model provided “smart brake” abilities by optimization the value of the brake actuation pressure according to the pressure selected by a driver. The important property of the future smart brakes is related to stabilization and maximization of the brake performance versus the brake pedal travel selected by a driver and current braking conditions. In this chapter, influence of the interrelated parameters, such as a vehicle speed, the brake actuation pressure, and temperature, has been analyzed during a braking cycle in the case of the disc brake. The new optimization model of the brake actuation pressure of a passenger car has been developed.