DRAFT
High performance processing functions (color coded orange in the FRM, Figure 3-1: Airborne Reconnaissance FRM) generally refer to real-time processing operations performed on raw (or pre-processed) sensor data and real-time system control functions typically performed in airborne reconnaissance subsystems.
The term "real-time processing" refers to the ability to process data at computing speeds fast enough to keep up with continuous, sustained sensor data-rates with no data buffering. The term also refers to the ability to time-tag sensor data with great accuracy and precision, which may be measured in nanoseconds (e.g., accuracy of marking an event to an absolute time reference might be ± 50 nanoseconds, and precision might be to the nearest nanosecond). Similarly, the term "real-time system control" refers to the ability to invoke sensor (and other subsystem) commands at an exact instant in time, typically with accuracy measured in microseconds.
Functional descriptions for the digital signal processing, system processing and control, and encrypted storage functional areas are given in the following subsections.
3.5.1 Digital Signal Processing Functions
Standards for this functional area are:
The digital signal processing (DSP) functions are generally "number crunching" mathematical operations performed on the sensor data in real-time. Sensor data is input from the SIGINT, IMINT, and/or MASINT front-ends via the high-speed data flow network. Processed data is output to subsequent functions (e.g., processing, storage, or direct reporting) via the multimedia network. Currently, no standards exist for all DSP functions although some are under development.
Example DSP functions in the SIGINT domain include the following:
Example DSP functions in the IMINT domain include the following:
Example DSP functions in the MASINT domain include the following:
The Common Imagery Processor (CIP) program is directed at standardizing some of the DSP functions for imagery. Originating from the CIGSS initiative, the DARO funded the initial CIP program to develop a proof-of-concept capability for imagery ground/surface systems to demonstrate the feasibility of processing data from multiple IMINT sensor subsystems in a single, common processor. The functions would be similar to those cited above for the IMINT DSP functions. The CIP will interface with sensors of various types (SAR, electro-optical, and IR), and from various manufacturers. If the proof-of-concept is successful, the DARO will assess whether a CIP follow-on program will be pursued as a standard technology for IMINT ground/surface systems.
The services also have been migrating to common SIGINT processors to conserve R&D dollars and lower recurring costs through multi-service purchases. In addition, the SIGINT community's JASA initiatives include standardizing some of the real-time SIGINT processors (e.g., SEI processors, PROFORMA, ELINT, multichannel, etc.). Recent R&D shows promise for a shoe-box size CRAY super-computer that will enhance the capability to process multiple types of signals simultaneously through state-of-the-art technology (e.g., Marquise, S90E). Massively parallel super-computers enable multiple signal processing algorithms to execute concurrently which widens the instantaneous spectral coverage. This type of technology will become more prevalent in airborne reconnaissance systems, especially where mission requirements call for high performance processing in airborne components of the system.
For digital video, MPEG-2 4:2:2 Profile@Main Level (ISO/IEC 13818-1 Systems, 13818-2 Video, and 13818-3 Audio) is the designated standard.
The only other applicable technology standards identified to date are for image compression algorithms. Airborne reconnaissance systems should use image compression algorithms specified in the National Imagery Transmission Format Standards (NITFS). Per that standard, which currently applies only to still images, Joint Photographic Experts Group (JPEG) is preferred.
3.5.2 System Processing and Control Functions
Standards for this functional area are:
This area refers to those functions that are typically implemented in the platform's subsystems to effect overall control of the reconnaissance payload (e.g., sensors and processing subsystems). The principal function is making real-time commands to the front-end and DSP functions * to manage and control the remote sensing process (collection) and real-time processing of the sensors' data. This may involve executing a preplanned target acquisition plan (e.g., in the case of autonomous UAV operations), or it may involve responding to operator command and control received from operator workstations either on-board or remoted through the data link. In the case of operator-invoked commands, latency is a key design concern (i.e., delay from operator action to execution of the command, and return of the expected response back to the operator).
Functionally, as shown in the FRM (Figure 3-1), the system processing and control functions use the high speed data flow network for the real-time interfaces with the front-end and DSP functions. The C2 and multimedia networks are used for interfaces with other functions.
Other system processing and control functions include the following:
Currently there are no technology standards identified for system processing and control functions.
3.5.3 Encrypted Storage Functions
Standards for this functional area are:
Encrypted storage provides the capability to store classified algorithms, data, etc. necessary to support the mission while allowing the platform to be operated at an unclassified level when unkeyed. Currently, there are no technology standards identified for encrypted storage functions.
DRAFT