An approach to detect the accident in VANETs using acoustic signal. (June 2020)

• Record Type: Journal Article
• Authors: Suman, Amrit; Kumar, Chiranjeev
• Is Part Of: Applied acoustics. Volume 163(2020)
• Publication Date: 2020-06
• Subjects: VANET -- Road accident -- Acoustic signal; Acoustical engineering -- Periodicals; Periodicals; 620.2
• DOI: 10.1016/j.apacoust.2020.107205
• Languages: English
• ISSNs: 0003-682X

Abstract

Abstract: Event detection using acoustic signal pattern recognition is a fascinating research area among researchers. This paper is based on algorithm development to detect vehicle accidents based on acoustic signals generated during vehicle accidents. Two types of acoustic signals have been taken for event (accident) detection. First is the sound of a vehicle collision, skidding of wheels, rollover of the vehicle, and second is distress and panic call of human. According to the world health organization (WHO), 1, 250, 00-man died globally, and 238, 562-man died in India due to vehicle accidents. One of the significant reasons is expanding quantities of vehicles with the increasing population. As indicated by an overview, 79.02 million autos sold worldwide in 2017, and 81.6 million are anticipated in 2018. Accidents on the streets are also increasing, along with traffic. Many lives are lost because of accidents on the road. A genuine street accident happens regularly, and 16 persons die. That's why a solution to prevent all life in a vehicle accident is required. This paper describes the solution by developing a robust algorithm that can be implemented on the microprocessor. The algorithms and processes of the result are described in the paper. The testing and results show the algorithm is robust, and efficiency is 94.7% for collision, with 88.7% recognition for a human distress call.