ASICs always have ruled the military and defense market as the reliable devices of choice. That’s not a surprise. The military has exacting requirements for ruggedness, temperature tolerances, and reliability, among others. But recently FPGAs have been making inroads into the space.
Field-Programmable Gate Array (FPGA) is a programmable semiconductor device or integrated circuit (IC). The device could be reprogrammed as per desired application or functionality requirement as against Application Specific Integrated Circuits (ASICs) that are function-specific. FPGAs offer numerous advantages such as easy debugging, rapid prototyping, low cost and lower the risk of product obsolescence. Rise in demand for customizable integrated circuits serves as a key driver to the FPGA market. High demand for power efficient and high performance IC designs is expected to favorably impact market growth.
Telecom accounted for 33% of the market in 2013 and is expected to remain the largest segment over the next six years. FPGA devices are widely used in wireless network applications. In addition, automotive and consumer electronics segment is expected to grow at a rate higher than the global average over the forecast period from 2014 to 2020. In the consumer electronics sector, growing popularity of smartphones, advanced touch screen functions, phablets, tablets etc is expected to drive the demand for FPGA. Advanced Driving Assistance System (ADAS) such as GPS maps, AFS (Advanced Front-lighting System), 3D visualization, collision avoidance system etc make extensive use of FPGA devices. Furthermore, FPGA technology in military and aerospace applications such as RADAR and SONAR, warfare electronics, unmanned vehicles etc is also expected to contribute to market growth.
FPGA in Mission Critical Applications
Defence electronics requirements are expanding the envelopes of performance with high-speed serial IOs, dedicated DSP capabilities and faster memories interfaces. Current generation FPGAs address these aspects and are gaining the centre stage in many aerospace and defense applications like RADAR Data Acquisition and processing, TRMs, Beam forming, Target Simulation, ECM, Signal Intelligence, Communication Intelligence, Telemetry, etc. Many newer families of FPGAs feature high-speed serial interfaces used in emerging aerospace and defense applications like Serial RapidIO (SRIO), PCI Express (PCIe), InfiniBand, Gigabit Ethernet, 10 Gigabit Ethernet etc.
Currently, the FPGA are used in the defence industry instead of ASIC circuits for the same reason as in the civilian sector: If the number of circuits that will be needed to implement the project, less than 100,000, the cost advantages are in using FPGA circuits. Also, the design cycle for FPGA circuits is shorter than the ASIC one. However, there is a group of applications where cost is not the most important parameter. Configuration of the FPGA in field conditions is another advantage for military applications. FPGA in the military communications device can be automatically deleted if device was lost or was captured by enemy.
Although there are no statistics that would monitor the FPGA circuit used in military applications (the market is small compared to the commercial use FPGA), the main FPGA manufacturers declare a significant increase in the use of these devices in military applications in the last decade.
Mr. Frank Liu, Aerospace and Defence Industry Marketing Manager, MathWorks, USA shares FPGA requirements in A&D sectors, “FPGA requirements within the Aerospace & Defence Industry vary widely based on vertical segments, For example, commercial aviation and avionics areas are primarily driven by safety critical applications, and compliance with certification standards such as DO-254 is very important. Similarly, in space-based applications, rad-hard, reconfigurability, power management and mission life-span are some of the key considerations that drive requirements and component selection.
In general, applications that require reliability, sustainability, security and safety critical designs are everywhere and directly drive system development workflow from the top down to the FPGA. To meet many of the design goals, Aerospace and Defence engineers need mathematical and functional correctness, traceability and maintainability of design and code, methodical design approaches, and efficient design verification.
What is driving use of FPGA in A&D? Amr Elashmawi, Vice President of Corporate Marketing – Verticals, Microsemi Corporation discusses 3 trends as below:
- Power: Mobility and the need for performance per watt is a strong driving factor for defense electronics. For communications and sensor systems, the trend is to do as much of the signal processing in digital form as possible. Many systems are performing digitization at RF frequencies directly without down-conversion to an IF. The flexibility and signal processing capabilities of FPGAs make them well-suited for the entire signal processing chain. Further, because FPGAs can be customized to the datapath processing, FPGAs typically have better performance per watt as compared to GPPs and DSPs.
- Flexibility and time to market: The re-programmable nature of FPGAs lends itself well toward fast development and verification, as well as the ability to do field updates.
- Security: With information growth comes security challenges. Security needs to be approached as a layered strategy that includes protections against tampering, supply chain, design security and data security.
He adds, “FPGA requirements will differ depending on the end type of system. If we look at the bulk of the platforms that we service, there are some broad commonalities for fielded systems, including:
- Broad operational envelope, with high reliability requirements. Systems need to operate reliably over temperature and often times will have requirements to mitigate effects of Single Event Upsets (SEU).
- Sensitivity to Size, Weight and Power (SWaP). Systems developed for the soldier (such as communications equipment) are battery powered and thus would have a tight power budget. Other systems, such as avionic control systems, may not be power limited per se, but will have a power budget due to thermal constraints. The rule of thumb for mechanical constraints (size and weight)is the smaller, the better.
- Secret algorithms, highly sensitive key materials for communications
Microsemi offers SmartFusion™2 flash-based FPGAs. These are flash-based FPGAs, with a feature set you would expect in a mainstream FPGA (high speed SerDes, integrated ARM Cortex M3 microcontroller with a host of hardened peripheral communication controllers, memory and DSP-block rich fabric). Because the underlying technology is flash based, it comes with a unique set of benefits – including instant on, SEU immunity, as well as better security. We’ve also engineered the device with advanced security features. This includes resistance to side-channel attacks, state-of-the-art key storage and tamper detection, including a metal mesh around critical circuitry to detect physical tampering.
We also have a product line specifically for high-altitude applications. RTG4™ is our radiation-tolerant FPGA product family for signal-processing applications. RTG4 comes in densities up to 150,000 logic elements and it runs at speeds up to 300MHz. Unlike the company’s traditional one-time-programmable RTAX space-qualified devices, RTG4 is based on flash technology, which makes it reprogrammable.
According to Mr. Frank Liu of MathWorks, “We are seeing two key trends in Aerospace & Defence system development driving FPGA use. First, there is a sharp increase in applications that require complex mathematical algorithms. These applications include next generation Intelligence, Surveillance and Reconnaissance (ISR) and vision systems, wideband tactical radios, advanced integrated sensing systems and RADARs. Second, the increase in application complexity is driving the need for integrated workflows that enable hardware and software co-design for FPGAs.
These trends are incentivizing hardware and software vendors to work closer together to provide more integrated solutions allowing for higher-level of design abstraction. As a result, engineers are benefiting from better connectivity between algorithm and system design environments and downstream FPGA flows leading to substantial efficiency gains in design and verification. We expect these trends to continue and accelerate in the next few years”.
He adds, “Implementing reliable designs on FPGAs requires deep understanding of hardware architectures, fixed-point effects, ability to make efficiency tradeoffs such as mapping to memory vs registers, throughput vs. resource usage, etc. It also requires an understanding of hardware concepts such as pipelining, over-clocking and serial vs parallel design approaches. As design complexity increases, the gap between the mathematic algorithms and the embedded HW and SW code also increases. Under these conditions, traditional methods of system development do not scale and break down very quickly.
The vast majority of the world’s most innovative air, space, naval, and land systems are designed, implemented, and verified with MATLAB, Simulink, and Model-Based Design. The F-35 Joint Strike Fighter, the Orion Multi-Purpose Crew Vehicle, the Mars Exploration Rover, as well as numerous unmanned aerial vehicles, and advanced wireless/ software defined radio (SDR) designs are a few examples of the state-of-the-art designs created with MathWorks products.
Model-Based Design provides the modeling platform for creating an executable specification and simulating the embedded system and plant models. Aerospace and defense companies worldwide rely on Model-Based Design environment enabled by MATLAB, Simulink, and Stateflow from MathWorks for most of their major programs. In addition, HDL Coder (VHDL and Verilog code generation), HDL Verifier (HDL co-simulation and FPGA-in-the-loop verification), and Fixed-Point Designer (automated fixed-point design and optimization) enable engineers to develop complete DSP, FPGA, and ASIC implementations directly from MATLAB and Simulink. The generated VHDL or Verilog is synthesizable, platform-independent, and fully traceable back to the model and requirements.
Certification kits are available for safety-critical applications or standards-compliant code qualification. For more information, visit http://www.mathworks.com/solutions/fpga-design/
Advantages of the FPGA
Possibility to adapt to any standard. In the case of modernization of the armed forces are always undergoing some risk. The armed forces are forced to in order to ensure interoperability using standards, but it is not always possible to correctly predict which standard will be dominant. The advantage of systems based on programmable circuits FPGA is flexibility and the ability and capacity to adapt to any standard. This can eliminate the losses incurred by the introduction incorrectly selected technologies and standards into the armed forces.
The length of the design process. Flexibility and adaptability offered by FPGAs are directly related to the length of the design process. ASIC circuits require a long design process – typically 14 to 24 months. The average time needed to implement the design for FPGAs is 6 to 12 months, which includes the specification, implementation, verification, validation and prototype production. This is a significant time savings, which is important in the implementation of the critical applications and length of the design process sensitive applications and adaptation to new standards. Necessary design changes during the phase of using the device are implemented by reconfiguration of the FPGA. In the case of ASIC circuits, it is necessary to implement the complete design process from the beginning, which also includes the specification and verification.
Changing behaviour of the digital system without exchange of the hardware. FPGA can change the behaviour of the digital system without replacing hardware. All changes can be implemented through software means alone and reconfiguration can be performed even in field conditions. The flexibility is inherent in the FPGA, can reduce the price of the design and the price of used hardware. The ability to change the behaviour of the circuit and field conditions allows for a design with extremely flexible life cycle. The result is the possibility of rapid assimilation of new standards, allowing the armed forces fulfilment of set tasks and implementation of innovation with minimal delay. FPGA can be updated at any time, locally or via remote access. With partial reconfiguration can provide support, service, and an update on field conditions.
Increased performance of the computer system. Although part of the FPGA circuit is reconfigured, the rest of the circuit performs a defined action. This can eliminate the shortfall in functionality and performance of the circuit, which is reconfigured. Partial reconfiguration can allow implementation of multiple applications in a single FPGA. The implementation of multiple applications is switched at the time.
Sharing hardware. The implementation of multiple applications performing partial reconfiguration of one FPGA is a condition for allowing the sharing of hardware. The benefits that flow from it are as follows: reduce the number of necessary facilities, reduced power consumption, smaller size of the necessary circuit boards and lower overall financial costs of implementing the periphery.
The shorter the time needed for reconfiguration. The time required to configure the FPGA is directly proportional to the size of the configuration bitstream. Partial reconfiguration allows for a small modification of FPGA functionality without reconfiguring the entire circuit. Change only part of the configuration bitstream compared with reconfiguring the entire circuit can produce shorter total configuration time.
Use of the FPGA in the Military Applications
Unmanned aerial vehicles. FCS (Future Combat System) is one of the many modernization programs. Its aim is to provide cutting edge technology to the armed forces to enable them to dominate in complex environments of the modern battlefield. Part of this program includes a family of modular UAVs, which can be linked to a common network. Unmanned is intended to carry out activities that pose a potential threat to people and save lives. These devices must be capable of performing search, survey and also must have the ability active influence to the enemy. An example would be the situation when pilotless means gets into unexpected situations and remote operator needs to evaluate the video and audio in order to correctly assess the situation and to choose the right choice. To ensure that it is necessary to implement audio and video processing, which requires high-speed digital signal processing. Effective implementation of operations such “robots” is subject to the ability of listening and seeing, but also the ability to perform many orders. The abilities need to do intensive signal processing and use principles of artificial intelligence and require powerful hardware. Using conventional DSP processors to provide these capabilities has some limitations. Conventional DSP processors have a fixed and unchanging architecture that contains one to four units of MAC (Multiply and Accumulate) with a fixed width of the data. Architecture then defines throughput, which determines the speed of processing. To increase computational efficiency of conventional DSP is required to increase their operating frequency to the maximum, which reduces the requirements for system design. Subsequently, several DSP processors must be included in processing, causing problems with energy and space on the target platform. The biggest advantage of using FPGA is its flexibility. For the application circuit can be configured so that the processing was carried out in parallel with the desired degree of parallelism. This may increase the maximum data throughput and optimize system performance. FPGA reconfigurability is a feature that is in demand for unmanned aerial vehicles. Plan prepared in advance for the use of UAVs in touch with the reality of the environment in which it is used, it must be changed frequently. FPGA has an ability to adapt to changes in behaviour flexibly and quickly to increase the probability of survival on the battlefield. FPGA reconfigurability enables end users to upload a new configuration for carrying out certain tasks within a specified time. Instead of implementing the system with a dedicated hardware for each specific task, it is possible to design a system with a single FPGA, which is used to implement multiple applications or tasks.
Unattended Ground Sensors. Unattended Ground Sensors are represented by the family of devices for implementation of remote sensing, that are interconnected by wireless network. Autonomous ground sensors (APS) can be used in defence of the designated perimeter for detecting and identifying targets and early warning. The size (the smallest size) and minimum power consumption resulted in a high degree of integration. Some sensors (sensor type FLIR – Forward Look Infrared) require intensive processing of information received, as opposed to requirements that are not asked. The amount of information that is sent to the communication line is also limited, and its compression is required. Information is processed before receiving, to avoid sending image that has not been changed. Flexible customization of computer performance can affect power consumption and system size. The solution to these requirements is a system with a high degree of integration. The last generation FPGA on-chip contains many features that were previously performed by software (soft core). High-speed I / O are a good example and some FPGA manufacturers integrate on-chip banks gigabit transmitters / receivers, Ethernet interface and PCI Express. Using these circuits can be integrated with external devices and sensors can be realized with a minimum of discrete components. Internal sources that would be used to implement these functions can be used to implement other parts of the design. FPGA can be implemented as an integrated solution. For most sensors is sufficient to capture the periodic “snapshots” over a defined period of time. It is necessary to generate a continuous stream of information from these sensors, which can represent movement of persons or vehicles..
Software Defined Radio. To secure voice, data and video communications for U.S. forces, coalition forces and allies unit was defined vision of the Ministry of Defence to implement unified communications in real time. Transformation programs (JTRS – Joint Tactical Radio System) was created to eliminate the constraints and gaps between communication systems that can prevent the fulfilment of that vision. JTRS system is designed to ensure interoperability between modern and older systems through software programmable radios designed to ensure modularity, scalability, backwards compatibility and networking capability. Software-defined radio (SDR) includes a software programmable operating environment and can support multiple waveforms, made by a single system. The main constraint for the practical implementation of such systems are costs and requirements for power consumption. The SDR is currently implemented in a dedicated model, which requires a set of sources for each implemented channel. Each channel represents one type of modulation defined radio. Each modulation (waveform) is realized by sources to carry out processing. Typical single-channel SDR modem includes AD and DA converters, FPGA, DSP and versatile processor (GPP – General Purpose Processor). This represents at least four discrete devices, which must be multiplied by the number of channels SDR. The result is increasing costs and demands on energy consumption with increasing the number of channels. This implementation of the SDR is not effective for general use. Another way is to implement a model of shared resources. This model is able to implement several modulations that are using one set of equipment for processing. This method is fundamentally different from the previous method, which required appropriations for each channel. FPGA, which is partially reconfigurable, allows the sharing of hardware. Partial reconfiguration can allow reprogram selected parts of the FPGA circuit to implement the operations defined in user-defined time after the initial configuration of the circuit. Accordingly, it is possible to implement several modulations in one FPGA. These circuits allow the user to dynamically change the modulation used without damaging the current modulation. Using the FPGA reduces the two main constraints that prevent the spread of the use of SDRs. Funding is investigated by integrating the components required to implement multi-channel SDR. Power consumption is also lower because the unused computing resources are eliminated. The result is longer battery life, which is used to power mobile SDR system.
