A novel Big Data analytics and intelligent technique to predict driver's intent

Modern age offers a great potential for automatically predicting the driver's intent through the increasing miniaturization of computing technologies, rapid advancements in communication technologies and continuous connectivity of heterogeneous smart objects. Inside the cabin and engine of modern cars, dedicated computer systems need to possess the ability to exploit the wealth of information generated by heterogeneous data sources with different contextual and conceptual representations. Processing and utilizing this diverse and voluminous data, involves many challenges concerning the design of the computational technique used to perform this task. In this paper, we investigate the various data sources available in the car and the surrounding environment, which can be utilized as inputs in order to predict driver's intent and behavior. As part of investigating these potential data sources, we conducted experiments on e-calendars for a large number of employees, and have reviewed a number of available geo referencing systems. Through the results of a statistical analysis and by computing location recognition accuracy results, we explored in detail the potential utilization of calendar location data to detect the driver's intentions. In order to exploit the numerous diverse data inputs available in modern vehicles, we investigate the suitability of different Computational Intelligence (CI) techniques, and propose a novel fuzzy computational modelling methodology. Finally, we outline the impact of applying advanced CI and Big Data analytics techniques in modern vehicles on the driver and society in general, and discuss ethical and legal issues arising from the deployment of intelligent self-learning cars.