How do you weigh up the amount of trust you should place in a piece of intelligence before deciding whether to send soldiers on to the battlefield?
It’s a difficult question at the best of times, but is made even tougher when governments start mixing student theses with genuine information from the intelligence services.
Alastair Campbell’s so-called ‘dodgy dossier’ has highlighted the importance of developing systems to evaluate information before it is sent to the front line, to ensure military planners know what data to place most faith in, said Prof Chris Harris, director of the UK’s first national Defence Technology Research Centre (DTC), set up this year.
‘One of the fundamental problems is how we measure and transmit the level of belief we have in a particular piece of information. The dossier is a good example of where data has been produced but there is not much belief in it,’ he said.
The research centre for ‘data and information fusion’ is one of three DTCs recently established by the Ministry of Defence in a bid to improve the UK military’s access to science and technology research. The other two will focus on electro-optic remote sensing and human factors. As part of its remit the centre will investigate the use of artificial intelligence and algorithms to weigh up the likely accuracy of a piece of information, based on factors such as where it has come from and the level of trust soldiers should place in the data as a result.
The centre has initial funding of £10m each year for six years, of which £5m per year will be provided by the MoD, and the remainder by industrial partners BT Exact, Qinetiq and prime contractor General Dynamics.
The UK’s defence research base has undergone a great many changes over the past 20 years, according to Harris, emeritus professor of computational intelligence at the University of Southampton. What began as the Admiralty Research Establishment and the Royal Aircraft Establishment then became the Defence Research Agency, and later the Defence Evaluation and Research Agency, and was finally split into the Defence Science and Technology Laboratory and the now privatised Qinetiq in 2001.
‘The consequence of this long process is that the fundamental science base they had, which was world class 15 or so years ago, has now been eroded,’ said Harris. So the MoD decided to establish a network of DTCs in an effort to redress this balance, with three launched so far, and a second tranche to be established soon. The MoD announced recently that the first DTC of this second phase will cover ‘uninhabited vehicles’.
‘It’s an attempt by the MoD to capture the very powerful research base we have in this country,’ said Harris. ‘It’s also a fairly low-cost way of doing research, because it basically means getting work done by world-class researchers at half the price.’
Harris’s DTC, which also has a number of academic partners including Imperial College and Southampton, Bristol, Cardiff, Cambridge and Cranfield universities, will consider data fusion issues such as sensor management, multiple target tracking, biometrics and autonomous guided vehicles.
‘Data and information fusion is where you have a range of data sources. They may be different sensors or databases, all of which relate to the same issue. For example, I may be in a helicopter and want to know where I am. So I would use GPS, radar and a map, and what I want to do is integrate all those together to give me a single picture of the vehicle – location, velocity and the objects around me. That sounds very simple, but unfortunately it’s not,’ said Harris.
The centre already has 43 projects underway, with around 70 researchers working on them at any one time. These projects begin with the basic problem of managing a set of sensors in a hostile environment, to provide soldiers with the best information possible, he said.
‘Given that there are a finite number of sensors that you can place around a battlefield, where would you place them, when would you switch them on, and how would you get the information to the right person, with the minimum exposure of yourself to the enemy? Because obviously if you litter the whole battlefield with lots of radar equipment, people will know where you are.’
But sensors are just one source of information used by the military when planning a battle or manoeuvre. Information can also come in from military databases, governmental and non-governmental agencies, the mass media and the internet. So the researchers will investigate systems to ‘fuse’ all this data together, to ensure the relevant people have as much information as possible at their fingertips, in an accessible format. This can be crucial, as non-military information sources can be just as important as military databases, particularly in responding to terrorist threats, and need to be viewed alongside each other, said Harris.
‘A good example is the civil and military flight plans. You have to integrate the two together, because if you only check the military plans you could shoot down a civilian aircraft. That actually happened before the previous Gulf war, when the US shot down an Iranian aircraft because they didn’t check the civil flight plans.’
Timing is also particularly important in getting much-needed information from sensors and intelligence sources to those fighting on the front line, and is extremely difficult to get right. It is no good building up the perfect picture of the enemy’s troop deployment and the conditions ahead if it takes so long to reach the front line that your soldiers are already engaged in fighting before it gets there.
This time lag issue is a particular problem for the military, said Harris. When information such as intelligence images are taken from a satellite they must first be transmitted to the satellite ground station, then on to the military base of operations, from there to the field operations base, and then can finally be passed through to the front line. ‘So by the time you have gone through all that, the object you were interested in has moved on, and your data is dated,’ he said.
Harris blames this delay in getting information to the right person quickly enough for the problem of friendly fire, which resulted in five UK deaths during the recent war in Iraq. ‘Soldiers and pilots were tired, in the wrong place at the wrong time, being asked to make a decision in two seconds. It is about providing the right information at the right time – if you don’t then people die.’
Another way to reduce these friendly-fire deaths is to take humans out of the equation altogether, and researchers at the centre are also investigating the use of autonomous guided vehicles. They will look into the use of swarms of low-cost short-range micro autonomous airborne vehicles to carry out surveillance work, particularly where there is a possibility of biological and chemical weapons.
The research team has no plans to actually build the unmanned aircraft, as this would fall outside its data fusion remit. Instead it plans to develop the sensors needed for such aircraft to be used in potentially hostile environments, although funding for the project has yet to be finalised.
Cost will obviously be an issue with the DTCs. The network will inevitably face comparison with the US Department of Defence’s Darpa, which has an annual budget of around $2.5bn (£1.5bn) with which to fund research projects. Even if the MoD was to launch another three DTCs, the network would still not be in Darpa’s financial league, but money is not everything, Harris insisted.
‘Some problems are solved purely by intellect, so it isn’t just a case of throwing money at something,’ he said. ‘The UK does have an incredible record of solving quite difficult problems at low costs, so I think we can compete.’