Study on the Optimal Number and Position of Drones in Fire Prevention in Victoria

Abstract

During the 2019-2020 fire season, every state in Australia had devastating fires, with the worst impact on eastern Victoria. In order to help Victoria’s Country Fire Authority (CFA) to monitor and rescue the fire in time, we set up an optimized model of the number of drones combinations, which saves the national cost while monitoring the fire occurrence. We first carry on the K-means cluster analysis to the fire dense points according to the fire occurrence situation in Victoria, Australia, which obtains the possible fire area. Then, we calculate the emergency factor to judge the fire grade, according to the area fire occurrence frequency and the fire intensity. By using the improved ant colony algorithm, we find the shortest path through each fire occurrence point. Therefore, the multi-objective optimization model is established with the minimum total cost and the highest safety factor as the optimization target. And we make the signal propagation distance SSA drones, Radio Repeater drones and fireman under different terrain as the constraint conditions. Through the model, we calculate that the Victoria’s Country Fire Authority (CFA) needs 23 SSA drones and 20 Radio Repeater drones, as well as the location of the relays and the combination of DRONESs in different positions. The occurrence of fire is related to climate and season. Through the data of Victorian fire in the past 20 years, we use ARIMA time series to predict the severity and frequency of fire in different areas over the next decade. Then, we determine the annual monitoring range of each area to estimate the probability of fire. In the case of extreme fire events, we update the scale and frequency of volcanic occurrence, using the model of question one, find the optimal number of drones combinations, calculate the increase in equipment costs, and finally get the need to increase the equipment cost price of $500000(AUD). Considering the different terrain conditions in Victoria, we introduce the altitude correction coefficient to obtain the influence of altitude on radio signal propagation, combing the change of altitude. Then, the influence of terrain on the signal propagation of hovering radio-repeater drones is obtained, and the signal propagation distance under different terrain problem one is changed. The position of the optimized relay is obtained. Finally, we get the number of radio-repeater drones increases to 25. Finally, we examine the sensitivity and stability of the model and prove that our model is accurate and stable for the problems.