Thesis title: Autonomous Pest Detection and Control Using Coordinated Autonomous Ground Vehicles and Unmanned Aerial VehiclesSupervisors: Salah SUKKARIEH , Zhe XU
Pests such as weeds and animals cause billions of dollars of lost profit to Australian farmers every year. Research and developments in Precision Agriculture (PA) technology and operations in the last decade have promised to increase the cost-efficiency of herbicide and mechanical weed controls. In particular, Variable Rate (VR) applications of herbicides and microwave weed management can provide both high cost-efficiency and robustness against herbicide resistance. Autonomous weed management systems have been developed with robots such as LadyBird that are more even more cost-efficient and precise than VR spraying. This thesis explores the design and use of the Variable Injection Intelligent Precision Applicator (VIIPA) system to perform the tedious but important tasks of weed detection and control. The detection system is naturally extensible to other pests such as insects an animals that are observable via monocular and hyperspectral cameras.
The action policy is optimised using a priori information about the spatial distribution, prior actions and economic impacts.