In this era of digital world, we all have smart phones. There are so many features of smart phone which entertain us. But sometime it may become hazardous for us. Like uses of smart phone while crossing road and railway lines, either by indulging with head phone for listening music or something else, it may snatch our lives abruptly. According to the survey of road safety department of India around 1.35 lakhs people died every year in road accident and 10% of them due to mobile phones, either by using mobile phone while riding bike and driving cars or by crossing road by pedestrians. As well as there are more than 11,000 railway crossings are unmanned where accident ratio is higher. So, by having all these problems on mind; this paper proposes a system that can prevent accidents causes by smart phones or using mobile phones. This system can recognize train’s horn, and alert mobile phone users prior from any casualties. Here the system uses Hidden Markov Model to train this network to recognize particular sound at real time and able to perform action accordingly before any casualties. The system can minimize the accidents at unmanned railway crossing.

Keywords : Speech Recognition, Hidden Markov Model, Train, Unmanned, Accident, Mobile Phones.