Composite Fire Detection System Based On Arm

Fire is uncontrollable burning which can lead to great loss of possessions and threats to the public. The research of detecting the fire in the early time is very important to fire alarming in the early time, which can protect the public possessions effectively.Most of the traditional fire-detectors detect the fire by parameters of a single moment, which have many disadvantages such as frequent misreporting or failed-report under the interference of environment, long response to the fire parameters, which can only respond the fire later, which is hard to alarm the fire exactly. To detect the fire more exactly and reliably, it is helpful to research the rule of fire-happening and fire-development based on the firing process information.In this paper we construct a low-power experiment system of fire-detector which takes carbon monoxide and carbon dioxide as token gas. Based on various fire and burning experiments which obtain the gas characteristic of the process feature, we use the gas concentration ratio of carbon monoxide and carbon dioxide as well as the increasing rate of the ratio as the early process feature parameters which can distinguish the misreporting from the fire. Fuzzy clustering is used to identify the pattern of the fire process parameters. We design a fire early alarming criterion which can alarm the fire earlier and make it more exactly.Based on the ARM embedded microprocessor and real-time/multitask Operation SystemμC/OS-Ⅱ,we develop an early fire detecting complex system which can collect the gas in the real time when fire happens and can report the fire process features based on the fire criterion to the system . The results of experiments prove that the system can not only detect the fire exactly and reliably, but also decrease the alarm response time.