2005 Fiscal Year Final Research Report Summary
Research on Autonomous Team Operation of UAVs
| Project/Area Number |
16560690
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Aerospace engineering
|
|---|
| Research Institution | KYUSHU UNIVERSITY |
|---|
Principal Investigator |
HIGASHINO Shin-Ichiro KYUSHU UNIVERSITY, Faculty of Engineering, Assistant Professor, 大学院・工学研究院, 講師 (40243901)
|
|---|
| Project Period (FY) |
2004 – 2005
|
| Keywords | Aerospace Engineering / UAVs / Team Operation / Path Planning / Autonomous Flight |
|---|
| Research Abstract |
An onboard system of which core subsystem are composed of a quite small (3cmx9cm) and very light (approximately 30g) onboard CPU board and a sensor board (also about the same size and weight) and vital for the flight tests for the multiple autonomous UAVs are developed with the autopilot and ground software. The flight testing system is confirmed useful by actual several flight tests using a single vehicle
The method for the single vehicle to determine its flight path before flight and even after flight using Evolutionary Algorithm is developed. In order to make the treatment of obstacles such as terrain, bad weather area, and altitude limits easy, the method of treating those obstacles as the cells in 3-D grids is proposed. Evolutionary algorithm is applied in order to determine quasi-optimal path of an UAV minimizing the evaluation function which is composed of flight distance, time to flight, fuel consumpotion, and so on. Optimal path is determined avoiding the obstacles which are expressed as cells in 3-D grids using the modified 3-D Bresenham's algorith in which one can judge the collision of the straight line and cells in 3-D grids. The method is also modified as the hybrid method of Evolutionary Algorithm and rule-based automatic obstacle avoidance method, and the time for the computation is improved
The method for multiple UAVs in order to assign optimal tasks and their optimal paths at the same time is also proposed. In this method, ssignment of tasks are directly optimized by Evolutionary Algorithm based on the evaluation function composed of the items such as flight distance, time to flight, fuel consumption of each UAV, and the optimal path of each UAV is determined by the method mentioned above for a single vehicle. The effective ness of this method is confirmed by simulation.
|
