Orchid, a uk university research programme funded by the Engineering and Physical Sciences Research Council, is to use artificial intelligence methods to coordinate the human and machine elements of a disaster management operation. Humans can provide only partial situational awareness, such as where facilities are and what buildings are damaged. This needs to be augmented by uninhabited air vehicles (uavs) that can overfly a disaster zone to survey the bigger picture.
Imagine a scenario where explosions in a major city have triggered flooding and fires. Pandemonium ensues: crowds throughout the city start evacuating the area, afraid of further explosions and trying to avoid the rising water level. Within minutes, social networking sites around the world are being inundated with live video footage, uploaded by people on the ground from their mobile phones. This peak in data traffic rapidly overwhelms these networks. Conventional media is unable to maintain pace with the rapidly evolving situation on a large scale. This creates a scenario in which emergency responders are rushing toward the scene of the emergency, but their response is hampered by the breakdown of existing communication networks and the paucity of information about what is happening on the ground.
“The key to effective disaster response is compiling accurate and timely situational awareness.”
Challenge Disasters scenarios such as this are characterised by uncertainty and a rapidly changing situation. The main challenge is therefore to reduce uncertainty and respond with agility. Participating in this complex operation, conducted over a large scale, are the first responders who act upon fragments of information provided by a number of sources, such as uavs, cctv feeds, and crowdgenerated content. If the operation is not carefully orchestrated it could lead to bottlenecks, delays, and duplication of effort. Central orchestration, the traditional approach to disaster management, can be slow and inflexible. This is because information has to be uploaded to a command centre, whereupon it is processed to yield recommended actions, which are flowed back to the response team. This is inefficient when the response team is large and geographically dispersed and communications are strained.
Solution The solution, according to orchid, is to use artificial intelligence methods to coordinate the human and machine elements of a disaster management operation. The key to effective disaster response is compiling accurate and timely situational awareness, ie a current picture of what is where, and a projection of how the picture may change over time. Humans can provide only partial situational awareness, such as where facilities are and what buildings are damaged. This needs to be augmented by uninhabited air vehicles (uavs) that can overfly a disaster zone to survey the bigger picture. Orchid envisages a collective response to disasters involving humans and machines. For example, information from members of the public, sent by smartphones, could be aggregated by computers to improve situation awareness. Uavs may provide real-time footage to disaster responders on the ground, who can request specific information from the uavs using hand-held electronic devices. Autonomous ground vehicles may also have a critical role to play in a collective response to disasters, entering areas that pose danger to humans. This mixture of resources produces a more effective response, but only if the
Resources are intelligently allocated to tasks on the ground. The orchid project is developing software underpinned by powerful mathematical methods to achieve intelligent resource allocation. This is done by decentralised coordination between the uavs, enabling them to operate as an efficient team. To do this they pass messages between each other, but the solution is robust to out-of-sync or dropped data. Decentralised coordination means the uavs determine their optimal strategy to collect information requests from humans and distribute them back to the ground.
Timescale The solution has already been tested extensively in field trials located in Sydney, Australia. The trials involved Hexacopter uavs, which have six rotary blades and are just under a metre in width. These proved the concept that uavs can be allocated tasks from the ground and fly as a coordinated unit to action these tasks. The next stage is to re-run the trails with mock disasters in open spaces and dynamic interaction between humans and uavs. This is scheduled to take place in October 2012. A fully operational system is expected in 2014. Several police forces have expressed interest and it is hoped government and ngo organisations will also take up the technology.
The orchid project is a partnership between the universities of Southampton, Oxford, and Nottingham, and bae Systems, Secure Meters Ltd, and the Australian Centre for Field Robotics.
Www.orchid.ac.uk
Author:Dr David Nicholson, School of Electronics and Computer Science, University of Southampton