An intelligent swarm-based guidance and path planning algorithm for the Unmanned Arial Vehicles (UAV) provides the ability to efficiently carry out grid surveillance, taking into account specific UAV constraints such as maximum speed, maximum flight time and battery re-charging intervals to allow for continuous surveillance. The swarm-based flight planning is based on enhancements of distributed computing concepts that have been developed for NASA's launch danger zone protection. The algorithm is a modified version of an ant colony optimization theory describing ant food foraging. Ants initially follow random paths from the nest, but if food is found, the ant deposits a pheromone (modifying the local environment), which influences other ants to travel the same path. Once the food source is exhausted, the pheromone decays naturally, which causes the trail to disappear. When an ant is on an established trail, it may at any time decide to follow a new random path, allowing for new exploration. Using these concepts, in our system for UAV, we use two units, the Rendezvous unit and the Patrol unit. The Rendezvous units will act as pheromone deposit sites keeping a record of trails of interest (extra pheromone that decays over time), and obstacles (no pheromone). The search area is divided into a grid of areas. Each area unit is assigned a pheromone weight. The patrol unit picks an area unit based on a probabilistic formula consisting of parameters like the relative weight of trail intensity, area visibility to the unit, the distance of the patrol unit from the area, and the pheromone decay factor. Simulation of a UAV surveillance system based on the above algorithm showed that it has the ability to perform independently and reliably without human intervention, and the emergent nature of the algorithm has the ability to incorporate important aspects of unmanned surveillance.
展开▼
