Gas inspector mobile robot as carbon monoxide source finder

By increasing the production and emission of chemical compounds and toxic gases due to development of cities, industries and technology, control and monitoring the environment in order to prevent the emissions of hazardous air pollutants have became one of the most important issues that have been conducted. If there is any toxic gas present in the environment, it has to be detected and its source must be localized. Fire is one of the most possible and common disasters,which in many cases occurs accidentally and produces a large portion of toxic gases and smoke, and threatens the environment’s safety and human’s life. Therefore there have been great attentions to fire-related issues including fire detection, localization and extinguishing. Although significant improvements in fire detection systems provide more reliable and faster detection, which is very e↵ective to control the damage at early stages of fire, still localizing the source of fire is a dangerous process, which is primarily done by human operators with high life-threatening risks. Various intelligent mobile systems are proposed for this application but they mostly operate according to the visible characteristics of fire, which in many potentially harmful circumstances like smouldering, they could fail. As carbon monoxide is one of important components of every type of fire and presents in every stage of combustion and doesn’t exist in ambient air, it can be considered as a reliable target gas to localize fire source. In this project an inspector mobile robot is implemented to localize fire as a CO source,and a novel searching strategy including a sensing unit and a searching method is proposed. For sensing unit, an array of three di↵erent kinds of semiconductor sensors of TGS series are used to detect the CO, which are inexpensive and commercially available. Experiments of sensing unit show the ability of this structure to estimate the source location in right, left, front or back side of robot. Navigation of this robot is based on the gas concentration in environment by air sampling in regular intervals. Algorithms used for exploration and localization by mobile robots are mostly been developed to run on a PC, so they are not executed in real-time. Using FPGA is proposed by related works as a solution to have complex computations done in a single chip without the need for an external processing unit and to speed up the localization process. Therefore, Altera Cyclone II FPGA was used as the main processing and control unit because of its availability in the research lab. This robot could localize and declare the CO source successfully in 80 percent of experiments.

Preview PDF FK 2015 70RR.pdf Download (2MB) | Preview

Actions (login required)

View Item