Today, the use of drones has expanded, particularly in high-risk and/or inaccessible environments, or situations where the cost of human resource is high. One of the most important uses of these unmanned vehicles is in rescue fields where they carry instruments and resources, or transfer wounded people. One of the most important discussions in this regard is the issue of routing these cars in an unknown environment. The first step for the vehicle to start its mission is to drive around environmental barriers. Environmental barriers can be divided into two categories. The first category comprises barriers that can be located on a map using satellite imagery and known maps. The second category contains obstacles that the car may encounter while navigating the path and but which have not been anticipated. To solve this problem, this research uses the PSO method to optimize offline routing in an environment with a specified map. The vehicle then may encounter new obstacles when moving on the planned path and identify those obstacles using sensors. In this case, using the neural network algorithm, it can obtain an optimal pseudo-path to circumvent the obstacle. In fact, the issue is divided into two sections. The first issue is the optimal routing with the PSO method, and the second section tackles the problem of dealing with unplanned obstacles using the neural network algorithm.
展开▼
