Design, construction and modelling of a low cost miniature UAV using machine vision

The price barrier for the implementation of UAVs commercially is substantial, limiting growth in this area. This paper presents results from a project aimed at developing light-weight and low-cost systems for the on-board control of a small scale commercially available helicopter, targeted at the low-end of the market. This would allow UAVs to be used by relatively untrained operators. The control system consists of two major streams, a vision system for high-level control and a low-level attitude control system. The vision system has been developed using a combination of a blackfun processor and an inexpensive cmos camera to implement colour segmentation algorithms. The attitude control system is developed using Fleck wireless sensor nodes developed at CSIRO. These inexpensive nodes perform sensing and control of the helicopter for data acquisition purposes. Flight data has been used to identify and verify a dynamic model of the helicopter