1. navigational functionality
2. low cost
3. low power
4. communications capabilities for interaction with other UAVs and transmission of sensor data
5. advanced sensor technology pertaining to the specific sensing application
6. integrated network of sensors coordination flight operations as well as data collection functions

Coordinated operation of multiple UAVs, such that the UAVs themselves form a distributed sensor network.

Sensor Development: Appropriate sensory devices and systems are required to allow the development of advanced navigation, environment sensing and vehicle control systems

Collision Detection: Robust high resolution scanners are required to identify path obstruction by objects or other vehicles. The appropriate response to such detection warnings is thus required. In the operational environment, sensing techniques that are operable in dust, fog or rain are also required.

Terrain Modelling: The integration of various sensor types is required to enable modelling and verification of the terrain over which the UAV is travelling. This information can then be utilized for navigation purposes as well as potential surveillance objectives. The technique applied should also be able to correctly interpret temporary obstructions.

High Integrity Vehicle Navigation: The successful deployment of autonomous must be able to guarantee high integrity operation. In other words, the system must operate reliability, and when it fails it must fail in an identifiable, safe manner. By providing a fail-safe mode, the behaviour of the vehicle is always known. In the case of UAVs this may allow for the recovery of the vehicle and prevent it from disrupting the operation of the remaining active UAVs.

Simultaneous Operation of Multiple UAVs: In this case a number of UAVs are required to operate in the same environment in a coordinated manner to conduct a specific task such as coastal surveillance.

In addition to the issues associated with single UAV operation, additional problems arise. Collision avoidance functionality must now incorporate intelligent systems that enable interaction between UAVs allowing for predefined flight patterns and separation distances to be observed.

Sensor scheduling functionality becomes increasingly important due to the influx of information from a network of onboard sensors. Classification and dissemination of this information is required to enable to UAV to carry out its overall objective in the face of conflicting feedback from different sensor. For example, collision avoidance of multiple UAVs may be simultaneously required while determining the optimal response to target identification and tracking operations.

The limited power and associated bandwidth for such communications make sensor scheduling even more important.

The UAV operational sensor network will also interact with application based sensors which may consist of imaging, detection and positioning functions that require transmission to different sub-network.

Target Tracking: The UAVs may incorporate target tracking capabilities to enable behavioural and operational changes, possibly utilizing target detection and tracking functions as an additional tool for navigation. Investigation and implementation of various detection and tracking techniques utilizing different sensor technologies.

Smart sensors and Biometrics: Vision Sensor Background Modelling  underpinning more reliable and robust computer vision tasks is the need to better model background and clutter. This could lead to better sensors that integrate such modelling to yield a better (more useful) signal (higher signal to noise, lower bit rate, more appropriate for the application). Applications span the whole spectrum but the most obvious are surveillance and biometrics.

The significance is immediate from the rather wide application in computer vision  together with the recognition that poor performance when faced with changes in illumination, interference from multiple non-target objects and other clutter, etc.; is a significant, if not the main, impediment to acceptance of automated vision technology for surveillance and biometrics.

The overall scope of the project thus incorporates research for all aspects of distributed sensor networks.

• Intelligent sensors that will perform the necessary information processing operations, along with self-validation and configuration within the network.
• Sensor data fusion in order to obtain more informative data through fusion of multiple sensor types that would otherwise be available using a single sensor or sensor type.
• Sensor scheduling incorporating the necessary optimisation and control techniques.
• Fault tolerance for the sensor network and individual sensors.
• Data aggregation at each network node, where each UAV may represent a network node.
• Self-organisation ability, enabling the continued and possibly modified operation of the remaining UAVs in the event of failure of an individual UAV
• Development of advanced communication protocols that will allow reliable and adaptive operation of UAV communications in a variety of environments

The development this technology and the coordination efforts by different research groups will address the various issues involved in such a complex application. The result will be an advanced surveillance tool the cost of which will make the application a much more viable and cost effective solution when compared to existing techniques.

The coordinated UAVs will enable functions such as rapid terrain monitoring along with surveillance operations through the collection various types of data. This sensor data can then be combined with existing long distance observational techniques such as satellite & radar to provide a comprehensive picture  of a given location or situation.

Target location by self-organising autonomous air vehicles

Figure 1 A mini-UAV fitted with DSTO electronic support equipment

Target location is a problem where the application of multiple sensors that are geographically distributed can determine or improve the location estimate of a target. If these sensors are capable of cooperative behaviour then the information from each sensor can be autonomously fused to provide an estimate of the target position. The individual sensors may be quite unsophisticated, yet the observation system that is created through cooperation and adaptive networking of these sensors provides sufficient process gain to achieve target location accuracies similar to those of expensive centralised sensor systems.

The accuracy of target location estimates depends heavily on the separation distance between the sensors. Large baseline geometry takes advantage of many seemingly unsophisticated bearing measurements that are organised into a coordinated observation system to locate a target.

Team formation is one method to address coordination of distributed sensors, data fusion, sensor resource and energy management, and communication link control based on the concept of cooperating machines. The defence science and technology organisation (DSTO) has the hardware and software algorithms to address the problem of integrating the sensors of a group of networked autonomous mini-Unhabitated Aerial Vehicles (UAVs). The mini-UAVs are tasked with locating targets within a region of interest. The challenge still to be addressed is in make the location estimation system adaptive to a dynamic environment and robust to failure.