Analysis of Different Measures to Detect Driver States: A Review

Abstract

To review and understand the state-of-the-art sensors, methods, technologies, and challenges in monitoring the driver safety. Statistics indicate that an ever-increasing number and diversity of accidents occur on roads, owing to reasons ranging from drowsiness to fatigue and inattention, as well as the mental and physical state of the driver, who could be drunk or cognitively distracted. In recent times, automobile industries have been developing new technologies like the Advanced Driver Assistance System (ADAS) to avoid death, injuries, or economic losses during tragic accidents. Most of the ADAS systems developed, however, rely only on a few measures to predict the driver’s mental and physical state, and lack accuracy as well. Therefore, it is essential to understand the technologies at hand so as to develop an efficient ADAS that can predict the driver’s safety states. A detailed analysis on different driver states suggests a need for more research in a real-time environment, compared to the research done in a simulated one, it is imperative to focus on a near real-time simulated environment. Further, using more than one measure to identify each of the driver’s states, through fusion or merging, would help develop a more accurate Driver Safety system. Based on the review, the different driver states can be monitored efficiently that calls for an intelligent and time-conscious data processing algorithm and uses data from non-intrusive sensors, combined with different measures.