Qinetiq is planning to carry out a demonstration of the control of a surrogate UAV from the cockpit of a fast combat jet in March next year as part of ongoing research into cooperative teaming operations between autonomous and manned aircraft.
The March demonstration will see the research organisation’s Panavia Tornado Integrated Avionics Research Aircraft (TIARA) used to control a BAC 1-11 twin-jet aircraft that has been modified to operate as a surrogate UAV with high level autonomy capabilities. The demonstrations will also involve the control of simulated UCAVs.
Qinetiq says the architecture under development is intended to allow for the control of multiple, self-organising UAVs or UCAVs by a single operator.
The approach is intended to facilitate human in the loop involvement in critical mission phases, such as a ground strike by multiple UCAVs in a combat environment. In a civil scenario, that critical human operator role could potentially involve the positive identification and recovery of a survivor in a search and rescue operation carried out by multiple UAVs controlled from a helicopter.
Proving the ability to distribute decision making between the operator and the individual air vehicles to reduce workloads is a key focus of the research programme.
The BAC 1-11 surrogate system has been flying from the Royal Air Force’s Boscombe Downs test and evaluation facility in Wiltshire since January this year in what Dr Jon Platts, Qinetiq’s research team leader for autonomous systems, describes a “series of flights of steadily increasing capability”.
Speaking at the IQPC UCAV 2006 conference in London on 28 November, Platts said the first flights of the modified BAC 1-11 were limited purely to automatic flight management testing of the aircraft itself, with those functions transferred to a modified lap top computer placed into the cockpit. The test pilot on that flight was a USAF officer on posting to Boscombe Downs.
The BAC 1-11 has remained under the control of a pilot and co-pilot on the flight deck for take-off and landing phases and transit to flight test areas during each of the subsequent trials flights. When operating in its autonomous mode the flight crew remain in position to carry out safety monitoring.
In October the test bed achieved fully autonomous flight with the UAV command and control demonstrator station located in the aircraft rear cabin as a prelude to the planned March 2007 demonstration.
Pending the outcome of the March activities, Qinetiq anticipates evolution of the research and technology programme into a maturation and development phase, potentially linked to the UK Ministry of Defence’s planned UCAV demonstration programme to be spearheaded by BAE Systems.
Platts said the autonomous flight control architecture under development is based on the use of intelligent agent technology to facilitate a relatively high level of reasoning and system flexibility in environments that are characterised by high levels of uncertainty. “In old fashioned speak we would call that robustness,” he said.
He said that the research is necessarily retaining a wide focus at present so as to give the MoD a choice of options as part of its expected launch of a national UCAV development effort. “The MoD certainly haven’t made a decision yet as to what of these bits and pieces that is going to form any kind of capability, what we will be able to draw upon and so on, so we have to be pretty open minded as to what our autonomous system can do for us,” he said.
He also said that the notion of the airborne controller was part of a broad spectrum of approaches being explored by Qinetiq: “Having as one extreme a single seat fast jet operator supervising a number of UAVs, UCAVs, and at the other end of the extreme a ground based operator supervising those same UAVs, and explore ways of exchanging between those two.”
The autonomous architecture is developed around a hierarchical model that covers the entire spectrum of the aircraft system. “Those levels of control must be deployed across the entire vehicle,” Platts said. “With these complex high end type systems a range of sensors which must be managed; payload, weapons and so on that must be managed. You are going to have all sorts of health systems and environmental systems on board that must be managed. All potentially have an impact on the level of autonomy that you can deliver.”
On a manned combat aircraft those same management functions are carried out by the aircrew. “It is all about, given a nominal plan, policing that plan as things don’t unfold the way you anticipate. As soon as something needs changing, they redo the plan, rethink about how things should be achieved; they overcome aspects of system failure to still achieve the mission and so on. That is really what the crew are doing at the moment,” he said.
The architecture will “effectively share that [workload] across a datalink between the operator and the system, pushing things that are easily reasoned about by machine intelligence to the vehicle…For us it is very much the idea of an optimised partnership between an intelligent capability, an intelligent platform and the operator.”
The demonstration programme is currently funded by the UK Ministry of Defence under its applied research programme, but with initial funding commencing in 1998. Earlier flight testing of the system has included the use of the Cranfield University developed Observer UAV system.
The initial research focus looked at “a number of high risk operational environments. As thinking as to how to UCAVs and UAVs would be deployed in any strike role or any futuristic mission, they steadily firmed up, and started focusing on time sensitive targeting type missions.
“Overall what we are trying to do is to address operator requirements; mature the selected man-machine intelligent technologies. Whilst we are increasing autonomy, what we are not talking about here is the idea of pressing a button on the side of it, it goes and does a mission, and then comes home again. This is very much keeping it on a short leash and maintaining the human authority in those key areas.”
Given the ethical and political dimensions of warfare, that authority is particularly focused on the interpretation of rules of engagement, target selection and weapon release decisions. In architecture terms however, the actual conditions under which those decisions are made by the operator will be circumstance dependent. Platts told the conference that while “we will mandate that some of these decisions have to be made by the human being, we will mandate that they will have to probably change depending on the mission, depending on the campaign, and possibly on a day to day basis”.
While that level of command is feasible in a two ship manned-unmanned formation, in a multiple air vehicle environment that target level of human in the loop control needs to be abstracted between all air vehicles, Platts said. In turn that has the potential to place large demands on system bandwidth requirements unless an intermediary system is used.
Qinetiq’s approach uses intelligent software agents in that intermediary role so that the specific interpretation of those decisions is consistent across the formation, but in individual platform terms and how its role relates to the wider mission plan. Each level of decision making supports its own level of autonomy which can be enacted by the human operator, or what Platts describes as a “variable autonomy interface”.
In practical terms this approach means that each UCAV in a formation can be configured to seek approvals for every action it takes, or allowed a high degree of freedom that could conceivably reduce the human role down to final authority for weapons release after the aircraft has selected and determined an engagement solution for targets of its own choice.
“This allows us the flexibility to tag various decisions and to sanction when those decisions can be made by the system as opposed to when they can be made by the operator.”
That level of capability was included in the suite carried by the BAC 1-11 UCAV surrogate during its October flight programme.
Extensive use has been made of simulation as the project has progressed, with this also including development of the operator interfaces flown on the BAC 1-11 and being fitted to the Tornado trials aircraft. The BAC 1-11 suite includes a synthetic environment generator which can inject up to three UCAVs in the controller system architecture. Platts said that “as far as the TIARA operator is concerned, he is controlling four UCAVs”, one of which is the surrogate aircraft and the remainder being virtual.
As fitted to the BAC 1-11, the controller station comprises a four screen display providing a situational awareness picture; a sensor payload image display; a package command interface with this incorporating the variable autonomy selection interface; and a standard ground control interface detailing UAV location, course and systems status. For the research effort the sensor payload display has included the use of a burst illumination laser with automated identification capabilities for final target identification prior to engagement.
The Tornado controller station will involve three displays Platts said, with these directly based on the BAC1-11 arrangement but fully integrated into the aircraft rear cockpit for use by a single operator.
The ground based version of the same control station “is relatively minimalist compared to systems which are out there at present”, Platts said, with this aimed at reducing information overload for operators.
Further development of the overall architecture will be multi-dimensional Platts told the conference. He said ongoing work covers a broad spectrum of activities ranging from communications architectures and sensor outputs management through to system clearance and certification, with the later seen as a major challenge: “There is plenty to be done yet before we can rock up with a UCAV that is controlled from a TIARA.”
Source: FlightGlobal.com
