Vetronics designers struggle to fill the gaps where open systems fall short

Commercial off-the-shelf technology can go only so far in helping engineers meet the demanding needs of armored vehicles such as tanks and artillery pieces; vehicle electronics specialists are finding it difficult to satisfy program managers while staying close to industry standards.

Th Mae71743 33

Commercial off-the-shelf technology can go only so far in helping engineers meet the demanding needs of armored vehicles such as tanks and artillery pieces; vehicle electronics specialists are finding it difficult to satisfy program managers while staying close to industry standards.

By John Keller

U.S. Army electronics specialists have been trying for nearly a decade to reroute the direction of combat vehicle electronics from proprietary custom designs to open-systems standards. The needs for reduced system costs and rapid technology insertion make it imperative that they do so. Yet while experts have had some success in using commercially developed vetronics in new main battle tanks, armored personnel carriers, and armored engineering vehicles, the process has been difficult and continues to confront electronics engineers with sometimes-frustrating design compromises.

The central issue involves the kinds of harsh operating environments for which most commercially developed electronic components were never intended. Armored vehicles must withstand some of the most intense extremes of shock, vibration, temperature, and electromagnetic interference anywhere. Take the M1A2 tank from General Dynamics Land Systems Division in Sterling Heights, Mich.; it must not only survive, but also remain fully operational when exposed to nuclear blast and electromagnetic pulse, or when hit with any known aircraft-launched anti-armor weapon. "The tank is designed for survivability, to be put in places no human being should rightly go and get back alive," says Col. Chris Cardine, project manager of Abrams tank systems at U.S. Army Tank-automotive and Armaments Command (TACOM) in Warren, Mich.

It stands to reason that finding open-systems commercial off-the-shelf (COTS) electronic equipment for tanks and their armored cousins is more than a daunting task; it keeps researchers awake at night trying to find the right mix of COTS components and rugged packaging that meet the demands of today`s data- intensive battlefield, and that accommodate systems growth for the future as well.

"The trend will be for more and more COTS, but you will never see an all-COTS vehicle," explains Curtis Adams, vetronics director in the TACOM Research Development & Engineering Center (TARDEC) in Warren, Mich. The electronics development challenge, Adams explains, is finding the right balance of COTS and custom armor-specific technology to keep system costs to a minimum, while ensuring extensive capability and reliability of tomorrow`s combat vehicles.

Driving today`s vetronics design initiatives are several major platforms, including the M1A1-D program to upgrade the General Dynamics M1A1 tank with digital communications capability for the so-called "digital battlefield"; the M1A2 Systems Enhancement Package (SEP) program to improve processors, memory, displays, and the navigation/communications systems in the General Dynamics M1A2 tank; the M1A2 ComPack program to backfit many M1A2 tanks with M1A2 SEP-type command and control capability; the M2A3 Bradley Fighting Vehicle upgrade from the United Defense LP Ground Systems Division in Santa Clara, Calif; the General Dynamics U.S. Marine Corps Advanced Amphibious Assault Vehicle (AAAV); the Crusader self-propelled artillery system from the United Defense LP Armament Systems Division in Minneapolis; the General Dynamics Wolverine heavy-assault bridge; the United Defense Grizzly Breacher mine-clearing system; the M109A6 Paladin self-propelled howitzer upgrade from the United Defense LP Combat Systems Division in York, Pa.; and the Future Scout and Cavalry System, which is being competed between two teams led by United Defense and Lockheed Martin Corp. in Orlando, Fla.

Four pillars of COTS

Efforts to insert open-systems COTS components into vetronics architectures center on four areas: central processor, circuit card form factor/backplane data bus, high-speed data bus, and software interfaces. Under current plans, the first two are fairly straightforward; the second two present designers with several thorny problems.

The clear consensus among vetronics designers for central processor, backplane data bus, and printed circuit card form factor are the IBM/Motorola/Apple PowerPC 603/604 microprocessor, VME 64, and 6U VME. This architectural approach is part of the M1A2 SEP, M2A3 Bradley, AAAV, Crusader, Grizzly, and Wolverine programs. M2A3 Bradley vetronics designers originally specified the MIPS microprocessor, but are in the process of changing to the PowerPC to mirror other vetronics programs.

PowerPC has emerged as the embedded general-purpose processor of choice not only for vetronics, but also for many other military and aerospace programs because of its powerful performance and extensive software support. The use of VME follows a longtime military electronics trend and offers extensive industry acceptance, software support, known and predictable performance, and compatibility with many generations of electronic equipment.

DY 4 Systems Inc. of Kanata, Ontario, is the hands-down favorite to provide 6U VME Power PC single-board computers for combat vehicles, and has been heavily involved in the M1A2 SEP program. DY 4 has a reputation for extremely rugged board designs, and has a relationship with Thomson CSF of France, whose engineers fabricate a mil-spec version of the PowerPC.

If there is a downside to the PowerPC story for combat vehicles, it is the relatively small supplier base. DY 4 is virtually the only supplier in the world that can supply VME single-board computers built rugged enough, and with rugged enough chips, to meet the environmental needs of tanks, armored personnel carriers, and sophisticated artillery pieces.

Data bus dilemma

Not so clear, however, is the choice of serial data bus to tie vetronics subsystems together into integrated architectures. The M1A2, which is the standard for all integrated vetronics architectures, uses the venerable 1 megabit-per-second Mil-Std 1553 data bus. Some vetronics experts firmly believe the quarter-century-old 1553 is ready for retirement as the primary vetronics data conduit. Others, however, believe the 1553 has a lot of life left in it, and that smart architectural choices could extend its use far into the future.

The current momentum in U.S. vetronics design favors the Fiber Distributed Data Interface, known as FDDI, but an industry/government consensus is far from clear. The accepted approach is to run FDDI protocols over twisted-pair copper cable, instead of optical fiber, for enhanced reliability.

Industry and government officials in Germany, meanwhile, are pushing a different vetronics high-speed data bus - SAE AS 4075, a 50 megabit-per-second bus capable of passing data over copper or optical fiber. At the same time, U.S. avionics designers are falling into line behind the Fibre Channel high-speed data bus, which Adams admits will put pressure on him. Waiting in the wings is a high-speed data bus with a lot of potential but has yet to catch on widely among designers - IEEE-1394, better-known as Firewire.

Vetronics designers say a high-speed data bus is necessary to support future combat vehicle applications of the tactical internet, automatic target tracking, automatic target recognition, sensor fusion, sophisticated identification-friend-or-foe systems, automatic deployment of battlefield obscurants when used together with high-resolution sensors, and millimeter wave radar for self-defense and target acquisition.

Engineers from United Defense, General Dynamics, and their subcontractors favor twisted-pair FDDI for the Crusader, AAAV, and possibly for future M1A2 SEP upgrades. FDDI is also a leading candidate for the high-speed data bus on the Future Scout and Cavalry System - for now. Decisions on high-speed data buses could change rapidly, however, if other data bus approaches. gain prominence within industry and the Pentagon.

For the Future Scout and Cavalry System, VME and the PowerPC microprocessor are "a safe bet," predicts Dennis Bielawski, director of software and control systems at General Dynamics. The ultimate choice for the vehicle`s high-speed data bus, however, is up for grabs between FDDI, Fibre Channel, and Firewire. "There will be lots of sensors and sensor fusion driving the architecture," Bielawski says.

"It`s time to replace 1553, and if it takes FDDI to do that ... " says Paul Richardson, a TARDEC electrical engineer who specializes in vetronics data communications and software. "1553 was developed in 1973 - prior to the 8088 microprocessor," Richardson points out. "It has poor error detection, its addressing is poor compared to what`s available today. Its packet size is too small, and it needs a bus master that is a single-point of failure."

Yet most vetronics experts freely admit that FDDI is far from a perfect solution. Among its positives are its status as a widely supported industry standard, broad industry support for FDDI software protocols, and wide availability of laboratory-grade FDDI interface boards and interface chipsets. But it has its negatives.

"FDDI is an office network and is fairness based, not priority based," Richardson says. "And there is not much demand for a twisted-pair copper FDDI interface." One of the only VME companies that builds twisted-pair FDDI interface boards rugged enough for use in combat vehicles is Vista Controls Corp. in Santa Clarita, Calif., Richardson points out. With only one supplier, vetronics integrators are unlikely to get the levels of pricing and vendor support they would if multiple suppliers were competing for the work.

"What do you do?" asks a frustrated Richardson. "The commercial market is not meeting our needs - unless you stay with 1553." At the same time, there are vetronics experts within the U.S. Army who maintain that 1553 is all a combat vehicle will need for the foreseeable future, and scoff at the notion of installing FDDI or any other high-speed serial bus in today`s combat vehicles.

M1A2 project manager Cardine insists that no wider data piper than 1553 is necessary to support even future advanced capabilities on combat vehicles. The solution, he says, is to design vetronics systems with data processing capability located as close to sensors as possible, and send only refined information and instructions over the 1553.

Software challenges

Open-systems standards are also falling short of the needs of vetronics designers when it comes to software. The so-called Joint Technical Architecture (JTA) - a program to enforce commonality among military command and control systems, under direction of the Defense Information Systems Agency - mandates the POSIX operating system interface for vetronics systems as well as for many other platforms. The problem for vetronics designers is the lack of standardized real-time POSIX extensions. "There are no standard tests for real-time POSIX compliance," Richardson says.

"The Army JTA requirement is POSIX," Adams explains. "We need to define where we can use POSIX and extend it for real time. We need to define the real-time extensions our designers need." POSIX will be part of the MIA2 SEP program, Crusader, and many other projects. Vetronics designers are faced with the choice of waiting for POSIX standards committees to define the real-time extensions they need - which may be too late for some programs - or striking out on their own to lay down real-time extensions that ultimately may not comply with future standards.

Trends toward rapid technology insertion are also stepping up pressure on vetronics designers who must consider migrating software to several successive generations of processing hardware. That is a challenge easier stated than done.

"To port a real-time system to other hardware, your timing characteristics will change," Richardson explains. "Isolating timing characteristics on hardware is not available yet. They just haven`t been able to isolate timing characteristics from hardware. that means time-critical software must be compartmentalized and addressed one by one when we move it to new hardware."

These concerns require vetronics integrators to be extremely careful when switching software compilers, hardware, and operating systems during each round of technology insertion, Richardson says.

The most current efforts to introduce open systems standards into vetronics development involve application programming interfaces, Adams explains. This is a departure from earlier vetronics work that concentrated on and then abandoned establishing standard software middleware.

Click here to enlarge image

The M1A2 SEP main battle tank is built with technology insertion in mind

Click here to enlarge image

The future United Defense Crusader self-propelled artillery system will feature the PowerPC 603/604 microprocessor, VME 64 backplane data bus, 6U VME circuit cards, and twisted-pair FDDI high-speed data bus.

Click here to enlarge image

Liquid crystal flat-panel displays from Litton Industries are part of upgrades to the United Defense M2A3 Bradley Fighting Vehicle.

Latest M1A2 designed for vetronics technology insertion

By John Keller

The latest version of the Abrams main battle tank is pushing vetronics technology insertion to new heights. The tank, called the M1A2 Systems Enhancement Package (SEP), is as notable for the subsystems not immediately planned for the its vetronics architecture as it is for the subsystems that are.

The M1A2 SEP from General Dynamics Land Systems Division in Sterling Heights, Mich., is to begin production in 1999, and is being designed with enough hardware and software hooks and sockets to approximately double its vetronics capacity as new requirements and technologies emerge.

"We`re keeping the M1A2 SEP vetronics as modular and open as possible for flexibility. It allows us to accommodate technology insertion," says Rick Wyrembelski, program manager of Abrams vetronics at General Dynamics.

While the first M1A2 SEP tanks off the production line are to have new subsystems such as commander`s color tactical displays, improved microprocessors, additional solid-state memory, global positioning system, improved power distribution, and second-generation forward-looking infrared (FLIR) sensor, each tank will be built to accommodate the quick insertion of new subsystems in the future.

The M1A2 SEP architecture will have the capability to accept equipment such as a remote loader`s display, video bus, high-speed serial data bus, helmet mounted displays, improved fire-control processor, battle combat identification system, eye-safe laser range finder, improved slipring, local area network, embedded training, diagnostic expert system, and autoloader.

"Displays are a key element of SEP, so we looked at the best implementations available," Wyrembelski explains. "We wanted a single commander`s display to mimic a workstation that we could display different windows on. But we had to split it into two displays - one color display, and one for the second-generation FLIR."

The tank`s new color screens are active-matrix liquid crystal displays. The monochrome FLIR screen, however, must have the resolution of high-definition television - 1,316 by 480 pixels - and liquid crystal does not provide resolution that fine. Wyrembelski says he is confident that color liquid crystal eventually will have the resolution necessary for the second-generation FLIR, and he is building connectors into the SEP tank today to accept one workstation-type display.

Future needs also may require a color display at the loader`s station inside the tank. "We have a connector that will interface with another display," Wyrembelski says. "We will just need to plug in another display and load the software."

Col. Chris Cardine, project manager of Abrams tank systems at U.S. Army Tank-automotive and Armaments Command (TACOM) in Warren, Mich., says the M1A2 SEP tank is a veritable COTS vetronics demonstrator. His design approach: use commercial-grade and ruggedized printed circuit cards and install extremely tough card enclosures that can take the punishment of tank operations, but that also can accept several successive generations of circuit cards.

"The technology of a card changes, but the protection technology doesn`t change all that much," Cardine says. "We have built in spare card slots for growth, and use the same boxes, wiring harnesses, and we will have the capability to change only our subcomponents. That makes a big difference. We will consider any commercial component as long as it meets our environmentals."

That change may sound trivial, but it makes a big difference in how technicians support the tank, he says. "Whenever you change a box, there has to be tremendous testing involved. But if you do not change the LRU, you only must validate performance, and you can validate performance at the component level."

In previous generations of the M1A2, for example, designers used mass memory units custom-designed for the vehicle. But not in the M1A2 SEP. "The mass memory unit is a shock-mounted box. The hard drive in there now is COTS," Cardine says. "I don`t even know who makes the hard drive, and I don`t care."

Click here to enlarge image

MEMS sensors eyed for military tires

MEMS sensors eyed for military tires

By John Rhea

"Smart tires" using MicroElectroMechanical Systems (MEMS) are emerging as a dual-use technology in which military and commercial vehicles would use embedded sensors and actuators to interface directly with microprocessor-based advanced information processing systems to improve maintenance and cut costs.

Under sponsorship of the Defense Advanced Research Projects Agency, engineers at Goodyear Corp. of Akron, Ohio, are developing passive MEMS sensors for truck tires to monitor pressure, temperature, and tire wear and to provide a unique identification code for each tire as an anti-theft measure.

The embedded passive transponders emit RF signals - similar to the bar-coded information on groceries scanned at supermarket checkout counters - when queried by a hand-held scanner device. The information can then be inputted into a personal computer for analysis. In addition to preventing theft, knowing where each tire is and what truck it is on would enable trucking companies to improve their routine maintenance procedures.

Goodyear officials cite trucking industry statistics that fleet operating costs amount to about $4 billion a year, of which 39 percent is for fuel and 18 percent is for tires. Both of these costs could be reduced by keeping tires properly inflated.

In cooperation with Case Western Reserve University in Cleveland, Goodyear experts are working on advanced models of the sensors that would also provide readouts of tire pressure, temperature, and total number of revolutions since the tire was installed. Truck tires typically cost about $300 each.

The sensors are part of a larger DOD effort known as condition-based maintenance that is expected to be more efficient than current scheduled maintenance practices for commercial and military vehicles. MEMS devices would monitor critical temperatures, pressures, flow rates, vibrations, surface wear rates, fluid contaminants, and accelerations so that maintenance could be performed before system or component failure.

Like the U.S. Postal Service, DOD operates a large fleet of vehicles and is a likely market for embedded MEMS sensors in its own trucks and other utility vehicles. Expected benefits include reduced labor, fuel loss, tread wear, tire disposal, vehicle down time, and dependence on foreign oil - for a projected savings of $19 per tire.

Since the average cost of a tire for a large military vehicle is $200, that amounts to savings of nearly 10 percent. DOD officials have a goal that the cost of the sensors can be no more than 1 or 2 percent of the cost of the tire.

But that cost goal isn`t likely to be met if the sensors are limited to military vehicles. This is a case where dual-use technology can drive down costs for military and commercial users alike by creating the necessary economies of scale. There are two million long-haul, seven million short-haul and 33 million local trucks in this country.

COTS progress is incremental in vetronics

By John Rhea

Commercial off-the-shelf (COTS) technologies continue to make inroads into military vehicles in an incremental way despite the potential of huge automotive electronics markets

Automotive markets include the in-vehicle and infrastructure improvements being considered under the U.S. Intelligent Transportation System program, points out Duncan Young, director of worldwide marketing at DY 4 Systems Inc. of Kanata, Ontario.

Young makes the distinction between the high-volume automotive markets, in which costs are critical, and the specialized requirements of military vehicles. He says he doesn`t anticipate a lot of demand for VME board products for auto usage except at the top end, for example, but he does see the need for more vetronics to support voice, data, and video on the mission-critical systems of tanks and light armored vehicles.

Although, clearly, more digital data will have to be moved around within and among military vehicles, fiber optics, the logical choice, suffers from the twin constraints of excessive costs for the slip rings and the difficulty of field maintenance, particularly the connectors, Young notes. However, he does see potential for navigation systems, including feedback loops among the commanders and occupants of the vehicles.

Mature technologies, such as the 1553 data bus, which is cost-effective to field and maintain, continue to dominate the vetronics market, adds Anthony Jordan, product line manager at UTMC Microelectronic Systems, Colorado Springs, Colo. Fiber optic data buses are OK in space, but still have to prove themselves in the "dust and grit" of the battlefield, he says.

UTMC experts have also eyed intelligent transportation as a stimulus for advancing vetronics technologies, but Jordan says he doesn`t envision any significant immediate impact. He sees greater potential for remote terminals based on the 1553 architecture for precision-guided munitions, countermeasures, and vehicle electronics.

COTS is being used extensively in the M1A2 tank Systems Enhancement Package (SEP), and DY 4 is supplying the general-purpose processor (GPP) and mezzanine COTS module as well as a custom video interface. The GPP is used in the tank`s mission processor unit for hull and turret control and in the commander`s electronic unit.

This amounts to four modules all using a common base card ( the DY 4 DMV-176) for ease of field maintenance and reduced inventory requirements. DY 4 experts also anticipated the Army`s selection of the PowerPC 603e for the M1A2 SEP processors and uses 80 MHz versions in a COTS module.

As the M1A2 SEP program evolves to build on the Army`s investment in digital electronics, future options include increased lethality, electric drive, remote weapon station, and a two-man crew - all considered possible with COTS components.

Closer to home, DY 4 supplied similar COTS modules to Computing Devices Canada in Ottawa for the operator station of the Canadian armed forces new Light Armored Vehicle Reconnaissance to enable soldiers to conduct 24-hour, all-weather reconnaissance operations from within the hull.

The computerized operator station incorporates a military-grade VME computer built by DY 4, ruggedized high-resolution CRT monitor, key entry controls, a joystick, and a ruggedized VCR - all in one enclosure.

DY 4 supplied an integrated DMV-978 conduction-cooled chassis containing a DMV-152 single board computer, DMV-202 serial/parallel I/O board, and DMV-776 graphics controller card. This system links the operator station with the radar, video camera, laser range finder, and thermal imaging equipment. To date, 170 units have been delivered.

Click here to enlarge image

Mil-spec VME boards such as the DY 4 Systems SVME/DMV-783 graphics controller, pictured at left, are part of the vetronics architecture of the M1A2 SEP main battle tank.

CANbus offers dual-use potential

By John Rhea

One of the contenders for dual-use, "spin-on" technologies to be transferred from the trucking industry to vetronics is the CANbus (Controller Area Network data bus). Engineers are pioneering CANbus at Mercedes Benz in Germany and investigating it for use in U.S. military vehicles such as the U.S. Marine Corps General Dynamics Land Systems Advanced Amphibious Assault Vehicle (AAAV).

This is decidedly a mature technology and far from the "bleeding edge." Originally developed in the 1980s by semiconductor manufacturers, including Intel, Motorola, Philips, and Siemens, CANbus is a twisted-pair copper interface with a basic transmission rate of 250 kilobits per second and potential extensions as fast as 1 megabit per second.

Interest lagged until Mercedes began using it in trucks for what the military would call non-mission-critical applications, such as transmission and engine controls, instrument panels, and lights - essentially monitoring the health of the vehicle subsystems. Yet leaders in the German defense industry also became interested, and the idea has begun to spread as a low-cost alternative to more exotic data buses.

Engineers at General Dynamics Land Systems of Sterling Heights, Mich., for example, have evaluated it as an alternative to their proprietary Utilitybus interface that handles power distribution and low-level functions on the M1A2 Abrams main battle tank.

CANbus is an International Standards Organization (ISO)-defined serial communications bus that goes from the physical layer to ISO level 2. There are several higher-level protocols in use, and the common one for the trucking industry is Society of Automotive Engineers (SAE) J1939.

J1939 is a Class-C communications network designed to support real-time closed-loop control functions between electronic control devices, which may be physically distributed throughout the vehicle. It uses the CAN protocol, which permits any device to transmit a message on the network when the bus is idle.

Every message includes an identifier, which defines the message priority, which node sent it, and what data were contained within it. The arbitration process that occurs while the identifier is transmitted (using a non-destructive arbitration scheme) helps avoid message collisions. This enables high-priority messages to get through with little delay, because there is equal access on the network for any device.

More in Computers