No one can deny that mobile devices are increasingly becoming an essential part of our lives, and being used for information delivery, access and communication. In this paper, a novel assistive drive system with mobile offloading is proposed. Three effective measures are integrated for reliable and early drowsiness detection, namely behavioral, vehicle, and physiological measures. These measures give higher quality and relevant information. Additionally, the proposed system uses mobile devices to process readings. However, with huge amount of data and intensive computations, mobiles cannot deliver results in reasonable times. A possible approach is to offload computations onto the cloud.