A UK team is driving down the size and manpower requirements of tactical UAVs by using image-led navigation, a new technology
Stewart Penney/LONDON
Reducing the size and cost of unmanned air vehicles (UAVs) and improving the technology so fewer operating personnel are needed are key drivers in UAV development. New technologies promise to improve navigation and autonomy while increasing the intelligence gathered by the system.
A team from the UK Defence Evaluation and Research Agency (DERA) and Cranfield Aerospace has developed the Observer UAV to demonstrate image-led navigation as an alternative to the waypoint-based navigation used today.
David Potts of DERA's unmanned air vehicles system integration development says tactical UAVs should not require airfields and need to be small. This drives requirements such as autonomous take-off and landing, which in turn reduces the number of personnel required to operate the system. Waypoint navigation forces the operator to plot a series of points through which the UAV will fly. Image-based navigation allows the operator to tell the UAV what he wants to see and the UAV then manages its own mission. This reduces the need for highly qualified personnel to manage the UAV in flight. Observer is a tactical-size UAV that uses image-based navigation. The aim, says Potts, is to reduce the UAV to a man-pack size that can be carried by an army section or company as part of its standard equipment.
DERA, with Cranfield Aerospace - a company owned by Cranfield University's College of Aeronautics - uses a systems approach to UAV development, says Potts. The research driver, he says, is the interrelation of range, air vehicle size, product quality and cost along with the necessary trade-offs and compromises required so that a practical requirement could be defined.
One compromise is environment, a function of terrain and weather. In Europe, for example, the cloud level is often 1,000ft (305m) and a small airframe has a limited - 20kg (45lb) - payload so is restricted to electro-optical payloads rather than cloud-penetrating radar. The UAV will also likely be equipped with line-of-sight links back to the operator. This, says Potts, will mean that a surveillance capability will be available 90% of the time at 25km (13.5nm). If the range is stretched to 35km, there will be 50% availability. The earth's curvature means that, at 1,000ft, a UAV will not be able to provide any capability at 60km. Also, if range is too great, the UAV will require more endurance, which means fuel capacity will have to be boosted. This, in turn, means the airframe will need to be larger and more complex. Greater range needs even larger, more complex and, therefore, more expensive datalinks, and the need to operate at higher altitude - to overcome the Earth's geography - necessitates a larger sensor.
As with the air vehicle, there are compromises required for the sensor, says Potts. If a user has to identify a main battle tank (MBT) from 1,000ft altitude, then a 500m slant-range sensor will suffice. Such a device will weigh around 1.5kg. A similar system capable of identifying an MBT from 2,300ft would need 2,000m slant range and weigh 35kg.
Potts says the practical implications of a battlegroup system also need to be considered. If a UAV is to be deployed then forward mobility and concealment become concerns, the former so the system can keep up with the frontline as it advances or move to support a different sector. Short ranges allow small UAVs to be used, while the need for a large number of air vehicles means the system will have to be inexpensive. Operation needs to be as automatic as possible.
The 2.4m span Observer is designed to meet these needs, says Potts. The air vehicle weighs around 30kg, has a 2h endurance and, although it can fly at more than 8,000ft, is designed to operate at 1,000ft and locate a target to within a 100m circular error probable (CEP).
A fixed-wing platform - despite needing a larger area for launch and recovery - was selected as less risky and complex than a rotary wing system. The delta wing is gust-insensitive allowing the use of a strap-down sensor which is lighter, cheaper, and simpler as it does not require gyros to stabilise the camera. The gust-insensitive airframe is neutrally stable in all three axes - pitch, roll and yaw. A stable airframe will change its attitude and direction if hit by a gust, requiring a gyro-stabilised sensor to track a fixed position on the ground. The gust- insensitive airframe simply translates, retaining the same attitude and direction.
Autonomous launch and recovery, and a simple control system means fewer skills are needed. The control system, the reconnaissance sensor and navigation are interlinked. In the Observer's nose are four charge-coupled device (CCD) cameras. Three of the cameras provide the navigation image. The cameras are mounted at 30°, 60° and 90° to the horizontal and each has a 30° x 40° field of view. This provides a ground footprint of 1,000m long and 800m wide, says Dave Dyer, business manager of control systems, Cranfield Aerospace. The three images are assembled electronically (see image below) in the ground station, he says, adding that the wide field of view (FoV) image provides good spatial awareness. It also gives greater resolution at the bottom of the screen, providing a better picture in the area of interest. Boresighted with the 90° camera is the fourth camera, which has a narrow FoV and provides a higher resolution view to aid target identification.
Control of the UAV is by three so-called sensor footprint control modes. The first, "observation plan", is similar to en-route navigation, the UAV being directed along a preplanned route. Observer's navigation system, however, moves the sensor's footprint over the area of interest rather than guides the UAV over the target. "We drive the image down the track and the aircraft does what it wants," says Dyer. The UAV's position is monitored using GPS satellite navigation. The intended target track is also known, as is the fixed stare of the sensor. The system then calculates, using geometry, the track the air vehicle should fly.
Mode two is "home to target", which allows the UAV to observe a fixed position by a series of overflights in a petal pattern - the UAV passes over the point of interest, before turning around and returning over the top.
Mode three is "observe" and allows the flight control system to control the sensor (in roll only) and the aircraft so that a fixed point is in the area of interest on the operator's display.
The operator controls the UAV using two touchscreens. One displays a map and the UAV's progress while the other shows the image from the sensor. Touching either screen will direct the sensor - and therefore the UAV - to look at a given point, says Dyer. Around the edge of the screens are a series of buttons to allow the operator to tell the UAV to rejoin its preplanned track or to uplink new flight plans. The simple operation and few commands allow a very low bandwidth datalink to be used that can operate intermittently, says Dyer.
Only limited training is required to operate the UAV using this system, he adds. Autonomous launch and recovery is either preplanned or can be ordered by the operator who would indicate an area on the map or sensor image, and the recovery is by automated parachute development with the UAV landing on an airbag released from the underside. An operating section for an Observer-based operational UAV could be included in a four-wheel drive vehicle, which would tow the launcher and contain the ground station, which has a 1m3 (35ft3) volume. A flight section would probably consist of 10 men compared with the 30 required for the British Army Marconi Avionics' Phoenix.
Although the initial sensor is four daylight CCD cameras, DERA and Cranfield Aerospace are considering an uncooled DERA thermal imaging sensor that weighs less than 5kg.
David Gardner, managing director Cranfield Aerospace, says it is intended that a US partner be signed up to market the image navigation technology for US programmes. He says talks have been held with a number of US companies and the US armed forces, although he says it could take up to two years to finalise a deal.
Potts says another development thread is to reduce the size of the air vehicle so it can be carried in a man-pack. The US forces are seeking a system by 2005 that would have about a 10km range. But other requirements, which would drive the size of the UAV, also need to be formalised, says Potts.
Source: Flight International
