RADAR OPEN SYSTEMS ARCHITECTURE AND APPLICATIONS*

Stephen B. Rejto

MIT Lincoln Laboratory Lexington, MA, USA

This work is sponsored by the U.S. Army Strategic and Missile Defense Command, KMR Support Directorate, under Air Force Contract F19628-95-C-0002. Opinions, interpretations, conclusions and recommendations are those of the author and are not necessarily endorsed by the United States Air Force.”

RADAR OPEN SYSTEMS ARCHITECTURE AND

APPLICATIONS*

Stephen Rejto, MIT Lincoln Laboratory Lexington, MA, USA

Introduction

Radar systems are traditionally developed from the

ground up, using proprietary hardware and software architectures. This traditional development model is expensive and requires long design times. Further, because each radar system employs unique architectures and technology, it is difficult and expensive to maintain and upgrade the vast assortment of fielded systems.

Acquisition reform and the proliferation of open systems (OS) and commercial off-the-shelf (COTS) technologies have paved the way for major changes and cost reductions in the development process of defense-acquisition programs. But OS and COTS are about more than saving money: they speed up the development process and provide access to the latest technology advances. Further, OS facilitates the use of common architectures, alternate vendors, and a more competitive acquisition model. A standard open architecture applied to radar systems could streamline the development process and improve future technology-insertion opportunities.

This paper presents the radar open-systems architecture (ROSA) that has been used successfully in building the prototype Cobra Gemini radar and is currently being used to modernize four unique signature radars at the Kwajalein Missile Range (KMR) and three unique radars at the Millstone Hill radar facility. ROSA embraces the OS model by decomposing a radar into functional building blocks constructed using COTS hardware. This decomposition provides loosely coupled operational subsystem components that, when tied together using well-defined interfaces, form a complete radar-processing and -control system. Building blocks can be easily added or modified to allow new technology insertion, with minimal impact on the other elements of the radar system. More importantly, existing radar building blocks can be shared and used to create new radars or to modernize existing “This work is sponsored by the U.S. Army Strategic and Missile Defense Command, KMR Support Directorate, under Air Force contract F19628-95-C-0002. Opinions, interpretations, conclusions and recommendations are those of the author and are not necessarily endorsed by the United States Air Force.”

systems. This modular OS architecture can lead to improvements in time-to-market, reduced cost, and increased commonality.

Open Systems

An open system is a collection of interacting components designed to satisfy stated needs. All components conform to formal interface specifications. Interactions among the components depend on the interface specifications; in particular, the interface specification of all components in an open system is

• Fully defined, • Available to the public, and • Maintained according to group consensus.

An open system approach

• Is an integrated technical and business strategy. • Uses modular hardware and software design. • Buys, rather than builds, individual components.

A key aspect of OS is a focus on decomposition and

interfaces, which provides maximum flexibility in developing and maintaining a system. By decomposing a system into functional units that are connected using open interfaces, developers can select components from a competitive marketplace based on performance, quality, and price. Replacing older parts with new components that adhere to the standard interface provides a maintenance and upgrade solution. A common example of an open system is the personal computer, which provides standard interfaces for disk drives, graphic cards, and other peripherals. By focusing on the interfaces, personal computers can be built using the best new low-cost technology. Customers also benefit by being able to replace or upgrade components independent of a specific vendor.

Another important issue in an OS approach is COTS technology. Maximizing the use of COTS allows developers to benefit from a competitive market and to change quickly to newer, better, and lower-cost components. Developing custom components that adhere to standard open interfaces is perfectly acceptable within an open system where custom components are required; conversely, COTS components that adopt proprietary closed interfaces cannot be part of an open system.

How is the OS approach relevant to large defense systems? For years the Department of Defense (DoD) was the major driver of electronic components used in weapon systems, allowing the DoD to synchronize the modification

or fabrication of new systems with new components. Today’s electronic world is market-driven, and the DoD has little control over the fast evolution of commercial components. With electronic components evolving every 18 months, it is increasingly difficult to build and maintain DoD systems, which have 8- to 15–year cycle times.

In order to benefit from new technology and

reductions in cost, new DoD systems must be designed to accommodate the fast evolution of the commercial market. An OS approach is the solution. By using open, standard interfaces, defense contractors can harness the best, lowest-cost technology for the DoD. If standard interfaces are used, new technology can be easily integrated or used to replace failed components that may no longer exist.

The use of OS is the solution to bridging the commercial market with the DoD and building cost-effective systems that can evolve over a lifetime and adapt to new threats.

Radar Open-Systems Architecture

The ROSA model decomposes a radar-processing and -control architecture into individual loosely coupled subsystems. Each subsystem performs specific radar functions and can run completely autonomously. When combined, these building-block subsystems form the entire processing and control architecture for a complete radar. Figure 1 compares the traditional radar architecture with a ROSA. Figure 2 shows a ROSA system.

Radar systems have historically employed tightly integrated designs, custom hardware, and proprietary interfaces. ROSA replaces the tightly integrated design with subsystems for each major radar component. These intelligent subsystems (also called radar peripherals) perform all interface functions between the high-level main computer and low-level radar electronics. This configuration provides an important level of abstraction that dramatically increases the level of hardware-independent software within the main radar computer. As the main computer software is not dependent on the underlying hardware, it is very portable from radar to radar.

Communication between the subsystem components and the main computer is key to the success of a ROSA architecture. Subsystems act as a software object that performs specific functions based on control messages; specifically, a high-level control message is passed from the main computer to the subsystems using a single commercial network interface. With every major control cycle (20 Hz), the main computer broadcasts a control

message to all the subsystems. Each subsystem reads the message and performs the requested function: in essence, each subsystem becomes an intelligent peripheral that provides a unique function driven by the control message.

In addition to simplifying design, ROSA systems show benefits in the test and evaluation stages. First, the radar interfaces of each subsystem can often be built using widely-available commercial boards that have already been tested by the manufacturer and are supplied along with diagnostic software. As a result, testing can start at the subsystem level instead of the component level, drastically improving the development cycle of the subsystem.

Second, the distributed architecture in a ROSA system provides a clean mechanism for testing individual subsystem components prior to integration. By providing intelligence within the individual subsystems, test and evaluation can be completed using a modular approach. Modular testing is accomplished by allowing each subsystem to generate its own control messages. (Figure 3 describes this concept.) During development and testing, subsystems generate and drive their own control message. This modular testing provides a very efficient use of resources and allows all subsystems to be developed and tested in parallel. The capability also provides control over specific sections of the radar without requiring the complete system. For example, the antenna control subsystem (ACS) can move the antenna while the transmitter control subsystem can transmit pulses independently while running in “local” mode. The autonomous nature of the subsystems also provides distributed fault isolation: subsystems are responsible for isolating faults within their section of the radar.

One example of a ROSA subsystem is the ACS. The ACS is 90% COTS and consists of a VME-based control computer and VME boards that perform all the interface functions required for a radar antenna pedestal, including the electronic interfaces to the servos, encoders, and status lines, as well as the required servo-control software. The ACS receives high-level position (azimuth and elevation) commands from the main radar computer and then does everything else. The main computer does not need to be cognizant of the underlying antenna electronics or the details of the servo loops. All required information is passed back and forth between the ACS and the main computer using high-level messages.

ROSA Applications

The Air Force Electronic Systems Center, MIT Lincoln Laboratory, and MITRE have recently developed and fielded a dual-band instrumentation radar using the

ROSA approach. The OS approach was key in meeting the low cost and short development schedule (two-year ground-based operations, three-year ship-based operations).

KMR, MIT Lincoln Laboratory, and Raytheon Range Engineering are currently modernizing the four signature radars (ALCOR, MMW, ALTAIR, and TRADEX) located on Roi-Namur on the Kwajalein Atoll. The existing radars are world-class systems supporting seven frequencies from VHF to W-band and an instantaneous bandwidth up to 2 GHz. These radars are used for metric and signature data collection on theater and national missile defense experiments, as well as for space surveillance. The radars also play an important role as surrogates for testing new radar technology that may be appropriate for new weapon-system radars. Examples of such technology include wide-bandwidth waveforms (512 MHz and 1 and 2 GHZ) and frequency jump burst waveforms. Although extremely capable, each KMR radar is a unique, one-of-a-kind system, making all four radars costly to maintain and operate. The goal of the KMR modernization effort is to expand the current world-class data-collection capabilities, reduce operation and maintenance costs, and increase the flexibility of the systems to adapt easily to customers’ needs and to incorporate new radar technology.

An OS architecture is essential in meeting these goals. By decomposing the radar system into ROSA building blocks, a high degree of commonality (greater than 75%) can be achieved among the four radars. Each ROSA component is designed to work as a generic radar subsystem, which allows a subsystem to be moved from one radar to another with minimal impact. As subsystems are used to abstract unique hardware components, the main computer real-time program can also be shared among the individual radars. ROSA Subsystems

Figure 2 shows the ROSA decomposition for the modernized Kwajalein radars. Each subsystem is built using VME or VXI standard components. A very large percentage (approximately 85%) of the radar subsystems is built using COTS components. Only seven custom boards exist for all the radars! As most of the KMR radars are dual-frequency, two copies (shaded boxes) of some subsystems are required. Using two copies of a subsystem for a dual-frequency radar demonstrates the benefit of decomposition: a single-frequency radar could not use a subsystem designed to handle two frequencies. The finer decomposition increases commonality by allowing a single-frequency radar to use a single copy of a subsystem and a dual-frequency radar to use two copies.

Figure 4 shows an example of several ROSA subsystems. The exciter is responsible for generating the timing gates and signals used by the radar. The exciter subsystem is driven by high-level target position information (state vectors) from the main computer. The exciter subsystem performs master timing for the radar. For every pulse repetition interval (PRI) up to 2000 Hz, the subsystem extrapolates the state-vector information and computes the appropriate transmit and receive timing. The subsystem then commands the appropriate gates and generates the requested waveform signals using digital waveform generators. The signal is passed through the upconvertor section of the exciter, which outputs the radar signal. With the exception of the upconverter board (which uses different radar frequencies), the exciter system is completely common from radar to radar. It exciter is a generic system capable of generating over 100 waveforms (easily expanded), from continuous wave to 2-GHz bandwidth, at pulse-repetition frequencies up to 2000 Hz. The exciter is PRI-agile, which allows unique control for each pulse over the placement of waveform and range window. In addition, the system provides a flexible test-target capability. The system is built using open VME and VXI standard technology and is 80% COTS, including all the digital and analog components.

Figure 4 also shows the intermediate frequency (IF) receiver and digital pulse-compression subsystems (DPCS). The IF receiver is 85% COTS and is designed around two VXI chassis. The first chassis contains the IF receiver card and other COTS boards to perform PRI-rate automatic gain control, AMP/PHS alignment, bandpass filtering, and downconversion. The second chassis contains COTS VXI test equipment: signal generators, arbitrary waveform generator, noise source, oscilloscope, spectrum analyzer, and power meter. The test equipment can be used to inject and measure signals within the radar, providing automatic fault isolation. Remote operations of the receiver functions are also provided.

The DPCS provides analog-to-digital sampling, digital filtering, inphase-to-quadrature conversion, and pulse compression. Critical to the design is the use of digital receiver technology. Combined with the digital waveform generators in the exciter, the digital receivers provide extreme flexibility in modifying or adding to the 100+ waveform repertoire. The system is scaleable in fast Fourier transform processors. Fully populated, the system provides four-channel, all-range pulse compression for time-bandwidth waveforms as high as 3000. Data rates within the DPCS can reach 480 Mbytes/second sustained. The DPCS is designed around the open VME standard and is 90% COTS.

Main Radar Computer

The main radar computer is responsible for coordinating all aspects of the radar. The main computer’s real-time program (RTP) is developed to be common across all the KMR radars, which is accomplished by writing software that is parameter-driven and hardware-independent. As described earlier, hardware-independence is achieved using subsystems to abstract the underlying radar hardware. Parameter files are used to configure the RTP for a specific radar; these parameters include radar frequencies, available waveforms, and beamwidths.

The RTP performs the following major functions:

• Coherent pulse integration and detection, • Signature and catalog data recording, • Multiple-target tracking, • Target classification, • Display generation and button processing, • Subsystem communication, • Track file maintenance.

All radar functions are under the control of the

autotasker and/or console buttons. The autotasker is a real-time interpreter that processes mission-profile scripts: these mission profiles are developed for individual missions and contain detailed operating instructions for the radar, including contingencies. Manual operations and diagnostics are available from a local console at the radar facility and by a remote console located at the Kwajalein Mission Control Center on the island of Kwajalein, 60 miles away from the radars. Remote access is also provided by a satellite link to a console within the continental United States. Automation, remote operations, and diagnostics all help lower the cost of operations and maintenance of these systems.

Figure 5 shows the complete ALCOR radar system.

Each rack represents an individual radar subsystem. The ALCOR system was received at KMR in December 1999; within only two weeks, and on its very first attempt, the system successfully performed a satellite track. Follow-on systems for the MMW, ALTAIR, and TRADEX radars are scheduled for delivery within the next few years.

Millstone Hill Radars

The Haystack, HAX, and Millstone radars are located in Westford, Massachusetts, and perform space-surveillance functions for the U.S. Air Force. Within the last year these radars began a modernization effort similar to the KMR program. After evaluating the KMR radar modernization effort, the Millstone Hill engineers realized

that they could leverage many of the existing ROSA building blocks. The OS architecture, available subsystems, and generic radar software have drastically reduced the cost for the modernization effort at Millstone Hill.

Future Electronic technology is rapidly advancing, and digital hardware is quickly replacing analog components. Decomposing radar systems into loosely coupled subsystems provides an evolutionary path for migrating to new technology. By using industry-standard VME and VXI boards, subsystems can be maintained or upgraded. In a similar manner, building radars using subsystems with well-defined open interfaces allows whole sections of the radar to be upgraded or replaced without affecting the other pieces of the system. Because hardware abstraction is provided, whole subsystems can be replaced without modifying the main radar software. It is possible that some day radar subsystems could be built by commercial companies to a common industry specification, which would provide all the benefits that open systems have to offer.

Conclusion

Acquisition reform, OS, and the use of COTS technology have paved the way for major changes and cost reductions in the development process of DoD acquisition programs. Using an OS approach can

• Provide cost reductions, • Simplify new technology insertion, • Promote the use of alternate vendors, • Streamline the development process. Radar systems are traditionally developed using

proprietary architectures, hardware, and software, which creates radar systems that are very expensive to build and maintain. By embracing the OS approach, the ROSA decomposes a radar into functional subsystems, each built using standard commercial components.

ROSA has been used successfully in building the

prototype Cobra Gemini radar and is currently being used to modernize four unique signature radars at KMR and three unique radars at the Millstone Hill radar facility. This modular OS architecture leads to improvements in time-to-market, cost, technology refresh, and commonality across radar systems.

Bibliography 1. Gillis, Matt (1999), “Open Systems Joint Task Force

Gets the Word Out”, PM, July-August, pp. 44-47 2. Rejto, Stephen (1997), ”COBRA GEMINI Radar

COTS Based Architecture”, International Test and Evaluation Association (ITEA) workshop on T&E, August, Bedford, MA

3. Secretary of Defense Memorandum, Subject:

Specifications & Standards – A New Way of Doing Business, 19 June 1994

4. Under Secretary of Defense for Acquisition and Technology Memorandum, Subject: Acquisition of Weapons Systems Electronics Using Open Systems Specifications and Standards, 29 November 1999

Figure 1 Historical and ROSA

Figure 2 ROSA Block Diagram

Figure 3 Communication (local and System mode)

Digital PulseCompression

TransmitterControl

Master TimingWFG

MIC

RO

WA

VE

Recording

Co

nso

le(s

)Main

Computer

Transmitter

Upconverter

Receiver

RadiationSafety

ROSA

AntennaControl

(Navigation)

Antenna

Transmitter

Recording

Antenna ReceiverExciter

COTSCustom

Exciter

Main Computer And Recording

Origin 2000

Origin 2000

Origin 2000

Origin 2000

. . .

Historical Architecture Radar Open Systems ArchitectureMain Computer

LOCAL

SYSTEM

Origin 2000

Origin 2000

Development, local and diagnosticoperations

Full radar operation

SubsystemMain Computer

Control Message

Control Message

Figure 1 Historical and ROSA

Figure 2 ROSA Block Diagram

Figure 3 Communication (local and System mode)

Digital PulseCompression

TransmitterControl

Master TimingWFG

MIC

RO

WA

VE

Recording

Co

nso

le(s

)Main

Computer

Transmitter

Upconverter

Receiver

RadiationSafety

ROSA

AntennaControl

(Navigation)

Digital PulseCompression

TransmitterControl

Master TimingWFG

MIC

RO

WA

VE

Recording

Co

nso

le(s

)Main

Computer

Transmitter

Upconverter

Receiver

Transmitter

Upconverter

Receiver

RadiationSafety

ROSA

AntennaControl

(Navigation)

Antenna

Transmitter

Recording

Antenna ReceiverExciter

COTSCustom

Exciter

Main Computer And Recording

Origin 2000

Origin 2000

Origin 2000

Origin 2000

Origin 2000

Origin 2000

Origin 2000

Origin 2000

. . .

Historical Architecture Radar Open Systems ArchitectureMain Computer

LOCAL

SYSTEM

Origin 2000

Origin 2000

Development, local and diagnosticoperations

Full radar operation

SubsystemMain Computer

Control Message

Control Message

Figure 4 Example ROSA Subsystems

Figure 5 KMR Radar System Electronics

6 channel Receiverand Diagnostics

Exciter - Timing, WFGsand Upconverterr

6 channel Digital Pulse Compression

Figure 4 Example ROSA Subsystems

Figure 5 KMR Radar System Electronics

6 channel Receiverand Diagnostics

Exciter - Timing, WFGsand Upconverterr

6 channel Digital Pulse Compression