Wireless Fire Alarm System Of Distributed Intelligent Composite Detection Based On ZigBee

The fire occurred as a higher frequency of disasters often cause huge casualties and property losses, and the existing wired fire alarm system had existed many problems, such as long construction period, amount of investment, maintenance problems, omissions, higher rate of false, etc. Cable mode existence power supply lines and signal transmission lines, when the fire broke out, the line is one of the weakest links, once the fault, it will lead to large areas of signal failure, and not monitor the scene of the fire in the process of fire broke out. For this reason, urgent need of a wireless fire alarm system.In this paper, the protocol stack standard of ZigBee version 1.0 is adapted to constitute a wireless fire alarm system. In each network node, PIC18F4620 is the intelligent network processor. Information from photoelectric smoke detection, temperature detection and CO gas detection is synthesized, then, analyzed by perception neural network model, the judgment of fire alarm is acquired. Based on the judgment, sound and light alarms are performed at the scene. Through the transceiver circuit, which core is CC2420, data is sent to the coordinator which collate and storage data. In the end, sound and light alarms are executed in the fire control centre, and all functions of wireless fire alarm system of distributed intelligent composite detection are realized.Study the structure, principles and methods of version 1.0 of the ZigBee standard, in-depth analysis of the company 1.0-3.5 Microchip version of the protocol stack file system, proposed the amendments and method for use of Microchip companies 1.0-3.5 version of the protocol stack, defined the various documents. defined the interface between the location and interface methods of the various Document.Design the complex detection circuit for fire signal, which the PIC18F4620 and CC2420 are the cores, adapt the peer-to-peer network topology, constituted a form of wireless communications with intelligent handling to the fire alarm sensor networks.Using MATLAB simulation software for the establishment of a wireless fire alarm system of single-layer perception neural network model, comparative analysis of the current more common several fire detection signal processing algorithms, using a detection rate of change with a combination of neural network algorithm. Simulate the experimental data of three sensor circuit, and the preparation of the relevant single-layer perception neural network algorithm intelligent procedures, to quickly and efficiently deal with the purpose of detecting signals.Place the wireless fire alarm system in the accessibility or obstacles space, the analysis of instructions to receive signal strength confirmed that the communication distance of wireless fire alarm system meet the requirements.