Aerial Vehicle Search-Path Optimization: A Novel Method for Emergency Operations

Optimization approaches for civil applications of unmanned aerial vehicles (UAVs) or aerial drones: A survey

“…Afterwards, we can capture information on the connectivity of cells and distances between them in a graph (see Figure c). In this way, if we code the cells as nodes of the graph , the coverage path planning problem can be transformed into a traveling salesman or a vehicle routing problem (VRP). If we code the cells as edges of the graph, a Chinese postman problem arises .…”

“…Most publications assume a stationary object, even if the target is a person or an animal, because the speed of the object is usually much slower than that of the drone. In the case of a moving object, the probability map evolves over time . Due to imperfect sensors, a drone detects an object with some probability if it visits the cell where the object is located , so it may make sense to visit some cells several times.…”

Modelado y optimización de misiones de búsqueda de objetivos mediante UAVs

“…En el estado del arte actual pueden encontrarse una amplia variedad de trabajos que proponen diferentes métodos de optimización de las rutas de búsqueda de uno [6,13] o varios UAVs [7,9,11] de acuerdo a alguna función objetivo que tiene en cuenta la incertidumbre del escenario, modelada mediante P (ν 0 ), P (ν t |ν t−1 ) y P (z t u |ν t , s t u ). Aunque es relativamente habitual optimizar la probabilidad de encontrar el objetivo [13] o la entropía [16], en este trabajo en el que se afronta el problema MTS, optimizaremos el valor esperado de encontrar el objetivo ya que es el criterio más adecuado para este problema [6,7,9,11]. Además, debido a la NP-complejidad del problema de búsqueda [15],éste suele abordarse con métodos de optimización aproximados o metaheurísticas (p.e.…”

Motion-Encoded Particle Swarm Optimization for Moving Target Search Using UAVs

“…As time progresses, that probability rapidly decreases due to the attenuation of initial information and the influence of external factors such as weather conditions, terrain features and target dynamics. The main objective in searching for a lost target using UAVs therefore includes finding a path that can maximize the probability of detecting the target within a specific flight time given initial information on target position and search conditions [2,3].…”

“…The initial search map has been modeled as a multivariate normal distribution with the mean and variance being computed based on initial information about the target position [5,6]. In [3,6], the target dynamic is represented by a stochastic Markov process which can then be deterministic or not depending on the searching scenarios. The sensor, on the other hand, is often modeled as either a binary variable with two states, "detected" or "not detected" [5], or as a continuous Gaussian variable [2].…”