Cognitive Based Platform for Advanced Driver Assistance Systems

点击翻译按钮获取中文摘要

Major advances are made in the design and development of novel ground transportation infrastructure and intelligent vehicles systems. This is an integral part for building the next generation of intelligent transportation systems involving smart roads and highways, connected vehicles, intelligent sensors and actuators, advanced communication systems and autonomous vehicles equipped with the latest smart mechatronic devices and knowledge based software. At the same time, additional stringent requirements are made by the car manufacturers and insurance industries for achieving best possible performance in terms of fuel efficiency, driving safety, comfort and environment protection. Advanced driver assistance technologies are among the most notable advancements made towards these goals. This market alone is expected to grow to $260 billion by 2020, as the technologies become mainstream not only for luxury and executive vehicles, but also for higher volume product. Transportation authorities in several countries have already mandated such systems and large research budgets have been allocated for the development of next generation advanced driver assistance systems (ADAS). The proposed research program will investigate and build novel functionalities for a cognitive based ADAS. The system’s proposed capabilities should substantially help enhancing the safety and the comfort of the car's occupants, while continuously monitoring the driver’s current state, the vehicle’s dynamic characteristics and the surrounding environment. To this end, we will develop a context aware based multi-agent system, which accounts for static and dynamic context information involving all types of sensory information coming from the inside and the outside of the car, along with a continuous update on the driver status and the driving environment. To achieve the highest levels of reliability and to infer the most accurate decision-making process, the system must be able to store a wide variety of information, including knowledge from past situations and should possess capabilities for adaptation and driver behavioral learning. Given the various types of sensory signals involved, which need to be interpreted and analysed in real time, an advanced novel approach for multi-modal data and information fusion will be developed as well. The proposed program will tackle a number of challenges in the areas of cognitive design, context awareness and data/information fusion. It should lead to real opportunities for the advancement of knowledge pertaining to the design of next generation ADAS and the training of highly qualified personnel, who will advance these technologies through their own research programs, while contributing to the strategically important Canadian automotive and transportation sectors.