The Design And Implementation Of Life Detection Equipment Based On Speech Recognition

The thesis comes from reconstruction projects of Lushan earthquake Ya’an,Sichuan—social services project on earthquake prevention and disaster relief. Thesishas based on the discussion on the existing audio life detector, combined withintelligent and efficient tendency of audio life detector, furthermore give a briefanalysis of the application of speech recognition in the earthquake relief and give arelated scheme.When the disaster comes, plunge the people into misery and suffering, thebuilding collapsed, traffic blocked, communication interrupt, as one of the mostdestructive disasters, the earthquake has the characteristics of instant, wide spread andhas difficulties in monitoring and forecasting, which cause negative social influence.In recent years the5.12Wenchuan earthquakes and4.20Lushan earthquakes causedsignificant losses in the aspect of economy construction and people’s lives andproperty safety, At present Chinese emergency rescue technologies in earthquake arerelatively lags behind than developed countries, and poor in the related technologicresearches.The earthquake disasters are frequent during these years, and the centralgovernment highly focuses on the earthquake rescues. At present, the disaster reliefindustry of China also needs a large number of rescue equipments; however high techrescue equipments with independent intellectual property rights are relativelyinsufficient. As an important part of public security Earthquake prevention anddisaster relief are related to public welfare of people’s livelihood, social stability thatbringing benefits to people and the people’s lives and property safety and thesustainable development of social economy. The public fields has been included as thekey field in the “National medium and long-term science and technology developmentplan outline (2006-2020)” in which the superior subjects of monitoring and preventionof significant natural disasters has clearly point out, focus on the research and explorethe key technologies of monitoring, precaution warning and emergency disposal basedon the disasters such as earthquake, typhoon storm, flood and geological disasters.With the consistent development of economy, China has proposed relatedresearches from the “fifteen” national science and technology key project to “eleven five” project successively, more and more talents devote to researches and make muchachievements. However because of poor material selection and structure considerationmany self-made rescue equipments turned out to be unsatisfactory, thus how todevelop a rescue equipment that is combined with the practical rescue environmentand with application value become the dominant direction.As an available vital sign, voice can be seemed as an effective signal in theprocess of rescue, at present the audio detectors are widely used during earthquakerescue, but most audio detectors are only in used with the help of operators, forinstance operators need look for space to stretch the probe, or place the audio sensorsabove relatively flat fields, thus the devices are easily influenced by the noise, impactidentification consequently.FPGA (Field Programmable Logic Array) is a new type of high performancecomponent that is developed on the basis of CPLD programmable logic device.The users are able to configure the needed bus interface personally, thanks to theoutstanding stability and low power consumption, FPGA are widely used in industrialcontrol and high speed process, moreover it can highly improved research progress invirtue of SOC which is based on IP (Intellectual Property).The research route: research and design a set device equipped with speechacquisition and speech recognition based on FPGA, through the audio signalcollection the device is capable of noise reduction and speech enhancement by specialalgorithm, then extract Mel frequency cestrum coefficient as the feature vectors andtrain the BP neural network by using the feature vectors, at last complete the speechrecognition.