In the last decade, the development and deployment of Uninhabited Airborne Vehicles (UAVs) has increased dramatically. This has in turn increased the desire to operate UAVs in civilian-airspace. Current UAV platforms can be integrated into civilian-airspace, with other air traffic, however they place a high burden on their human operators in order to do so. In order to meet the competing objectives of improved integration and low operator workload it will be necessary to increase the intelligence on-board the UAV. This thesis presents the results of the research which has been conducted into increasing the on-board intelligence of the UAV. The intent in increasing the on-board intelligence is to improve the ability of a UAV to integrate into civilian-airspace whilst also reducing the workload placed upon the UAV's operator. The research has focused upon increasing the intelligence in two key areas: mission planning; and mission piloting. Mission planning is the process of determining how to fly from one location to another, whilst avoiding entities (eg. airspace boundaries and terrain) on the way. Currently this task is typically performed by a trained human operator. This thesis presents a novel multidisciplinary approach for enabling a UAV to perform, on-board, its own mission planning. The novel approach draws upon techniques from the 3D graphics and robotics fields in order to enable the UAV to perform its own mission planning. This enables the UAV's operator to provide the UAV with the locations (waypoints) to fly to. The UAV will then determine for itself how to reach the locations safely. This relieves the UAV's operator of the burden of performing the mission planning for the UAV. As part of this novel approach to on-board mission planning, the UAV constructs and maintains an on-board situational awareness of the airspace environment. Through techniques drawn from the 3D graphics field the UAV becomes capable of constructing and interacting with a 3D digital representation of the civilian-airspace environment. This situational awareness is a fundamental component of enabling the UAV to perform its own mission planning and piloting. The mission piloting research has focused upon the areas of collision avoidance and communications. These are tasks which are often handled by a human operator. The research identified how these processes can be performed on-board the UAV through increasing the on-board intelligence. A unique approach to collision avoidance was developed, which was inspired by robotics techniques. This unique approach enables the UAV to avoid collisions in a manner which adheres to the applicable Civil Aviation Regulations, as defined by the Civil Aviation Safety Authority (CASA) of Australia. Furthermore, the collision avoidance algorithms prioritise avoiding collisions which would result in a loss of life or injury. Finally, the communications research developed a natural language-based interface to the UAV. Through this interface, the UAV can be issued commands and can also be provided with updated situational awareness information. The research focused upon addressing issues related to using natural language for a civilian-airspace-integrated UAV. This area has not previously been addressed. The research led to the definition of a vocabulary targeted towards a civilian-airspace-integrated UAV. This vocabulary caters for the needs of both Air Traffic Controllers and general UAV operators. This requires that the vocabulary cater for a diverse range of skill levels. The research established that a natural language-based communications system could be applied to a civilian-airspace-integrated UAV for both command and information updates. The end result of this research has been the development of the Intelligent Mission Planner and Pilot (IMPP). The IMPP represents the practical embodiment of the novel algorithms developed throughout the research. The IMPP was used to evaluate the performance of the algorithms which were developed. This testing process involved the execution of over 3000 hours of simulated flights. The testing demonstrated the high performance of the algorithms developed in this research. The research has led to the successful development of novel on-board situational awareness, mission planning, collision avoidance and communications capabilities. This thesis presents the development, implementation and testing of these capabilities. The algorithms which provide these capabilities go beyond the existing body of knowledge and provide a novel contribution to the established research. These capabilities enable the UAV to perform its own mission planning, avoid collisions and receive natural language-based communications. This provides the UAV with a direct increase in the intelligence on-board the UAV, which is the core objective of this research. This increased on-board intelligence improves the integration of the UAV into civilian-airspace whilst also reducing the operator's workload.

A forced landing for an Unmanned Aerial Vehicle (UAV) is required if there is an emergency on board that requires the aircraft to land immediately. Piloted aircraft in the same scenario have a human on board that is able to engage in the complex decision making process involved in the choice of a suitable landing location. If UAVs are to ever fly routinely in civilian airspace, then it is argued that the problem of finding a safe landing location for a forced landing is an important unresolved problem that must be addressed. This thesis presents the results of an investigation into the feasibility of using machine vision techniques to locate candidate landing sites for an autonomous UAV forced landing. The approach taken involves the segmentation of the image into areas that are large enough and free of obstacles; classification of the surface types of these areas; incorporating slope information from readily available digital terrain databases; and finally fusing these maps together using a high level set of simple linguistic fuzzy rules to create a final candidate landing site map. All techniques were evaluated on actual flight data collected from a Cessna 172 flying in South East Queensland. It was shown that the use of existing segmentation approaches from the literature did not provide the outputs required for this problem in the airborne images encountered in the gathered dataset. A simple method was then developed and tested that provided suitably sized landing areas that were free of obstacles and large enough to land. The advantage of this novel approach was that these areas could be extracted from the image directly without solving the difficult task of segmenting the entire image into the individual homogenous objects. A number of neural network classification approaches were tested with the surface types of candidate landing site regions extracted from the aerial images. A number of novel techniques were developed through experimentation with the classifiers that greatly improved upon the classification accuracy of the standard approaches considered. These novel techniques included: automatic generation of suitable output subclasses based on generic output classes of the classifier; an optimisation process for generating the best set of input features for the classifier based on an automated analysis of the feature space; the use of a multi-stage classification approach; and the generation of confidence measures based on the outputs of the neural network classifiers. The final classification result of the system performs significantly better than a human test pilot's classification interpretation of the dataset samples. In summary, the algorithms were able to locate candidate landing site areas that were free of obstacles 92.3 ±2.6% (99% confidence in the result) of the time, with free obstacle candidate landing site areas that were large enough to land in missed only 5.3 ±2.2% (99% confidence in the result) of the time. The neural network classification networks developed were able to classify the surface type of the candidate landing site areas to an accuracy of 93.9 ±3.7% (99% confidence in the result) for areas labelled as Very Certain. The overall surface type classification accuracy for the system (includes all candidate landing sites) was 91.95 ±4.2% (99% confidence in the result). These results were considered to be an excellent result as a human test pilot subject was only able to classify the same data set to an accuracy of 77.24 %. The thesis concludes that the techniques developed showed considerable promise and could be used immediately to enhance the safety of UAV operations. Recommendations include the testing of algorithms over a wider range of datasets and improvements to the surface type classification approach that incorporates contextual information in the image to further improve the classification accuracy.