Font Size: a A A

Research On Intelligent Fire-Alarm Detection System Based On Fuzzy Neural Network

Posted on:2007-12-04Degree:MasterType:Thesis
Country:ChinaCandidate:F R XuFull Text:PDF
GTID:2178360212495458Subject:Measuring and Testing Technology and Instruments
Abstract/Summary:PDF Full Text Request
Fire-alarm has the tremendous threat to the people's life and property, on which people get more and more concerns. With the perfection of the sensor technique, signal processing, artificial intelligence and wireless communicatio- ns technique, multi-parameter compound intelligent fire-alarm detection system obtains quite great progress. Adopting wireless communications technique lea- ves out miscellaneous lead and saves the usage space, having good market application foreground.According to the characteristics of fire-alarm detecting signal, the thesis mainly puts forward a sort of thought of adopting PC and Single Chip Machine (SCM) to constitute second-class measure and control system, applying fuzzy neural network intelligent algorithm, taking Bluetooth model as carrier to build wireless intelligent fire-alarm detection alarm system with point to multi-point. The main content and conclusions are as follows:By analysing and researching to current fire-alarm detection system and new-old residences, this thesis sets up the scheme of multi-sensor information fusion, combining CO, temperature and smoke sensors, designing the correspo- nding fire-alarm signal detection circuit and SCM system circuit. Combining with the Bluetooth protocol criterion, the Bluetooth Host Controller Interface (HCI) has been mainly researched. Using Ericsson ROK101007 model, the frame of wireless intelligent fire-alarm detection system based Bluetooth technique is built and serial interface hardware circuit design during SCM, PC and Bluetooth model is finished. The fire-alarm information is a kind of signal of special type, so the system should adjust the movement parameter automatically according to the variety of environment. Comprehensively considering the merits and shortcomings of neural network and fuzzy control, the thesis builds a novel fire-alarm forecast algorithm model and network learnalgorithm based fuzzy neural network. The down-machine software flow is programmed to realize the data collection of fire-alarm detectors. Taking ROK101007 as carrier, the up-machine software is programmed in Visual Basic 6.0, and the point to multi-point wireless communications system between SCM and PC is designed.By emulating and debugging, this system has higher intelligence, and can judge fire-alarm well. Simultaneity, it is also an attempt in applying the Bluetooth to wireless communications control in the special industry field (e.g. the environment of deleterious gas and dangerous districts), which has good application value.
Keywords/Search Tags:Fire-alarm Detection, Data Fusion, Fuzzy Neural Network, Fuzzy Control, Neural Network, Signal Processing, Bluetooth
PDF Full Text Request
Related items