Unmanned Aerial Vehicles (UAVs) are vital elements in the gathering of Intelligence, Surveillance, and Reconnaissance (ISR) data. UAVs carry a payload of electro-optical sensors plus lasers, radar, or signals intelligence. These sensors generate masses of data that are transmitted securely to the ground over limited-capacity data links. Assistance is needed on the ground to identify and classify targets so that the UAV can direct its sensors and alter its flight profile to track targets of interest. UAVs are in constant use and continuously evolving to detect and counter new threats. Operators and integrators are urgently seeking greater, proven capability, which is being provided by COTS products and related enabling technologies.
A surveillance UAV is essentially an unmanned sensor platform with well developed autonomous flight control allowing it to take off, follow flight plans, avoid obstacles, and land, but with limited mission autonomy. Size, Weight, and Power (SWaP) parameters can be so critical that COTS embedded computing standards might not be prime choice for the payload/sensor processing chain. However, because of the rapid design cycles needed to maintain tactical superiority, the infusion of proven and deployable enabling technologies – specifically COTS based – has become an essential development practice.
Sensor processing chain
There are many elements in a typical sensor processing chain: image stabilization, video compression, image processing, analysis, tracking, classification, and engagement. Electronic stabilization smooths out the effects of vibration and atmospheric disturbance. With video in and out connections, image stabilization can now be accomplished on a COTS module no larger than an SD memory card, for use in the smallest sensor platforms or within the camera mounting itself.
Air-to-ground bandwidth is a key issue. Only the largest UAVs support high-bandwidth SATCOMMs because of the weight and power requirements of terminals. Yet even mid-sized UAVs carry multiple sensors, relying on digital radio to transmit video to troops on the ground for immediate tactical evaluation. To preserve bandwidth, video streams are compressed, typically using H.264 to allow a number of regular 30 Hz TV channels to be transmitted over a data link. The daq8580 module from GE Intelligent Platforms is a complete compression subsystem integrated into a rugged box, supporting two channels of HDTV, or four regular TV, plus a network connection to the radio equipment. Shown in Figure 1, it is already in use by NASA Dryden Flight Research Center (DFRC) for atmospheric research.
Enhanced mission autonomy
Some parts of the processing chain are onboard the UAV, but many – such as image processing, analysis, or classification – are ground based and hence limited by link bandwidth, human resources, security, and military doctrine; this can cause mission-critical control loop times to vary from seconds to hours. Reducing this latency to identify, classify, and track multiple targets from multiple sensors without ground assistance requires a massive increase in onboard computing power, such as that offered by the new generation of many-core processors. An integrated multisensor vision system using these devices, packaged on 3U VPX (VITA 46) modules, is able to analyze incoming video streams in real time, saving air-to-ground bandwidth by sending processed target data to the ground for further analysis and classification. Next generations of many-core processors will bring the goal of autonomous mission and target engagement, with minimal man-in-the-loop intervention, closer to realization.
An additional goal is to extend mission duration with in-flight refueling from tankers or from UAV to UAV. A unique, deployable 3D vision system using a pair of cameras has been demonstrated using COTS modules that will automatically fly the UAV and engage with the refueling drogue, maintaining its position until fueling is complete.
Rugged, open systems embedded computing vendors are extending their portfolios with complementary product lines and technologies to meet many new application niches. These are not necessarily based on open standards such as VPX, as this is not always appropriate packaging for small, light, self-contained UAV subsystems. Rapid evolution of the UAV ISR market, widespread network connectivity, and the demands of SWaP are driving the creation of new breeds of COTS-based technologies with diverse packaging options tailored to their intended deployed platforms.