Cameras and computing power are both becoming more affordable. Companies are building ever larger networks of cameras and there is a drive to distribute the computing locally to the camera. It remains an open problem how to control such smart camera networks. FeedNetback will investigate scalability of distributed camera networks for surveillance. Multiple cameras can be installed to cooperatively monitor the perimeter of sensitive sites. In order to detect and recognise abnormal events, cameras and possibly other sensors must operate in a co-ordinated manner. In a distributed and cooperating system the information is processed locally but the results are shared over the network.

The particular challenges investigated in FeedNetback include:

• Architecture: scale and complexity of the system. Need to communicate large amounts of data efficiently and coordinate several cameras.

• Control & Complexity: The number of network nodes in this application will be very large and therefore the issue of scalability will come to the forefront.

• Control & Communication: enormous quantities of data collected by the cameras that, if communicated naively, are likely to swamp any network with traffic. Therefore the research objectives of Control design integrating networking constraints and Networking protocols for control will play a crucial role in the development.

• Control & Computation: Of primary importance will be the issue of Control aware computing, especially since one of the camera platforms used is based around FPGA and DSP on board processors that have slow development cycles and limited computing flexibility.