Motivation and Goals

The aim of this project is to allow an UAV to safely navigate in a previously unknown and GPS- denied environment without requiring pre-made maps or precoded information to avoid collisions. Using on-board cameras, which are indispensable components of an UAV, the autonomous navigation will be achieved by detecting and tracking visual features and objects in order to gift an UAV of a collision free navigation. In order to reduce the on-board payload data processing, the UAV will send the acquired data to an external processing unit (laptop). Autonomous indoor navigation can be extremely useful in search and rescue operations, surveillance tasks of areas hazardous to humans, traversal of paths inaccessible to ground vehicles and quick deployment.

In order to achieve this goal, the system needs to meet the following requirements:

  • The computational complexity of the system should be independent of the environment size
  • Firm and robust feature detection and object tracking
  • Fast data processing to satisfy real time constrains
  • Brisk obstacle avoidance
  • Flexible navigation