Real-time driving assistant application for Android-based mobile devices

This paper presents a real-time driving assistant application for mobile platforms. Defeating the constraints imposed by such platforms (limited memory, decreased processing power, low camera frame rate) we propose a simple and robust method for implementing various driving assistant features such as obstacle detection, collision warning or lane departure warning. For offering an optimal implementation, we parallelize the most computationally heavy parts by distributing work across the multiple cores available on the device. Using the Qualcomm MARE Parallel Computing library we succeed in more than halving the computation time on quad-core devices in comparison with single core ones.