Opportunities and challenges of systems on chip technology interconnected embedded systems and industrial automation
Advances in process technology and the availability of new design tools are expanding the field implementation of integrated systems, which are implemented as a whole of chips on a plate or as a set of modules in an integrated circuit. Technology systems on chips (SoC, System-on-Chip) is currently being extended in industrial automation to create complex intelligent field devices. This trend is accompanied by the adoption of platform-based designs, which facilitates the design and verification of complex SoCs with extensive reuse of hardware and software IP (Intellectual Property). Another important aspect of the evolution of embedded systems is the tendency to interconnect nodes built using specialized network technologies, commonly known as networked embedded systems (NES, Networked Embedded Systems).
technology SoC (System-on-Chip), a revolution of design of integrated circuits (IC) has been made possible by advances in process technology, they combine the main components and subsystems of an electronic product on a single chip or integrated set of chips. This development has been welcomed by designers of complex chips, allowing the maximum possible integration for more performance with lower power consumption and offers other advantages in cost and size. These factors are very important for the design and use of SoC is by far probability, a major development decisions in real time embedded systems.
A processor SoC is a complex integrated circuit or integrated set of chips, which combines the main elements or functional subsystems of a complete product. The most demanding SoC designs include at least one programmable processor and often a combination of a RISC control processor and a DSP digital signal processing. They also include on-chip communication structures, bus or buses of processors and peripherals, sometimes a system bus speed. For SoC processors is very important that the chip has units memory hierarchies and links to external memory.
For most signal processing applications is provided through a degree of hardware acceleration functional unit, which achieves higher performance and lower power consumption. For interconnection with the outside, the SoC design process includes several blocks of peripheral components consist of analog and digital interfaces (for example, buses in the system or board level rear panel). The future SoC technology can incorporate sensors and actuators based on microelectromechanical systems-mechanical, chemical or pro cessing (laboratory on a chip).
Among the most interesting SoC designs, which include hardware and software processors are programmable, real-time operating systems and other hardware-dependent software. Thus, the SoC design and use of means in addition to hardware, design and engineering at the system level, commitments hardwaresoftware and partitions, as well as architecture, design and implementation of software.
typical SoC device for consumer applications
programmable chips Systems
was recently expanded the scope of SoC. In addition to custom integrated circuits (custom IC), application specific integrated circuits (ASIC) or application specific standard parts (ASSPs), the new approach now includes the design and use of reconfigurable logic and complex parts with embedded processors. In some cases they also incorporate elements of other blocks of intellectual property, available on the market, such as processors, memories or specialized capabilities for applications, which are purchased from third parties. Xilinx (Virtex-II Pro Platform FPGA, Virtex-IV) and Altera (PCOS) are some companies that offer these programmable gate arrays FPGA user. The trend towards SoC technology combines high number of reconfigurable logic circuits RISC (reduced instruction set computing) integrated to implement a flexible and adaptable design combinations of hardware and software process.
algorithms with great content and great control logic flow of process data can be divided into RISC control processor, which speeds up the hardware reconfigurable logic. The resulting combination does not maximize performance and minimize energy consumption and costs, compared to custom IC or ASIC / ASSP for the same functions, but instead is very flexible to modify the design in the field and avoid the large costs engineering leading to changes in the field. For both new applications and interfaces and improved algorithms can be transferred to products already in actual operation.
Other products for this field are the processing cores and interfaces formed by blocks of multiplication and accumulation (MAC Multiply-Accumulate), designed to image processing and signs of DSP data flow, and high-speed serial interfaces for communication cable, blocks entrellos serializers / de-serializers SERDES. The SoC integration of programmable chip systems are not specific to an application, but not completely generic.
remains to be seen if the SoC applications will be successful in high volume consumer or be limited to two main areas: rapid prototyping of designs that will be redirected to ASIC or ASSP implementations and high-end items, relatively expensive communications infrastructure, which require flexibility in situ and which can be accept higher costs and higher energy consumption with a lower yield.
Other solutions, such as logical structures of the type metal gate arrays programmable processor subsystems together with residents and other nuclei such as those offered by "Structured ASIC" LSI Logic (RapidChip) and NEC (Instant Silicon Solutions Platform) - SoC are intermediate forms between the process "full mask" and programmable gate arrays by the user. This case has two specific problems: slow building design (from one day to several weeks), more non-recurring engineering work in FPGA (But much less than a full set of masks), lower cost, higher efficiency and less power consumption than FPGAs (around 15-30% worse than ASIC). Other mixed approaches are currently of interest, as the case of ASIC / ASSP with a region in a FPGA chip, to give more choice to design teams. Another variant is the combination of a configurable processor and permanently partially implemented in silicon, along with an FPGA region used for the extension of instructions and other field hardware implementations. Stretch semiconductor company inc, for example, uses Tensilica configurable processor in this type of SoC platform.
LX Tensilica Processor
platforms and programmable platforms
In recent years the design has focused on the SoC complex and reuse of virtual components, the so-called "platform-based design [1, 2], a planned design methodology that reduces time and labor required, in addition to the risks, to design and verify complex SoC . For they are extensively re hardware combinations [3] and software [4] intellectual property.
Unlike the reuse of IP block by block, the design platform based on groups of components assembled to form a reusable platform architecture, which in turn is combined with virtual libraries of hardware and software already tested and characterized for specific applications, forming a platform for SoC integration. Several reasons explain the growing popularity of platform-based design: design more productive, less risky, easier to use virtual components and integrated from other fields of design, architecture and reuse SoC created by experts. Among industrial platforms are complete applications to specific products such as Philips Nexperia and TI OMAP [5], and reconfigurable SOPC-based processors. Processor-based platforms such as ARM PrimeXsys, used by many processors
extended and configured Tensilica, focus on the processor, bus architecture and peripherals required basic and in real-time operating systems and basic control programs. FPGA and SOPC devices are a "metaplatform", ie a platform to create platforms. These devices contain more basic skills generic IP and embedded processors, on-chip buses, special IP blocks like MAC and SERDES and other IP blocks are already qualified. Designers can order these devices at companies such as Xilinx and Altera, and then customize the IP libraries metaplatform with specific domains of application, prior to submission to the equipment resulting designs.
Integrated interconnected
Another important aspect of this evoluciĆ³nson distributed embedded systems, called embedded systems interconnected to highlight infrastructure interconnection and communication protocol. An integrated system is a set of interconnected nodes integrated, spatially and functionally distributed, interconnected by a wired communication infrastructure and / or wireless protocols, which interact with each other and the environment through sensors and actuators. The system may also comprise a master node that coordinates the computation and communication to achieve specific objectives.
nodes integrated controllers or field devices such as sensors and actuators, are responsible for signal conversion, data processing and signal and communication chips. Functionality and the processing capabilities and communication drivers, each day more, have reinforced the widespread tendency to interconnect field devices around specialized networks, often called field area networks, which are typically a digital communication link and two-way multipoint [6]. In general, the use of networks (field area) is upside: the combination of integrated hardware and software is more flexible, the system offers better performance and simplify installation, updating and maintenance.
interconnected integrated systems are used in numerous applications (automotive, trains and airplanes, office buildings and industrial applications) and are mainly involved in the supervision and control. Some representative examples are the networks connecting field devices such as sensors and actuators with field controllers, PLC controllers specifically for industrial automation and electronic control units (ECU) for the automotive industry.
are also used in human-machine interfaces, for example, automobile dashboard displays and SCADA (supervisory control and data acquisition) for industrial automation. Specialized network technologies are as diverse as the fields of application. For example, PROFIBUS, PROFINET or Ether-Net/IP (both real-time communication) for industrial control and automation, LonWorks, BACnet and EIB / KNX for building automation and control, CAN, TTP / C and FlexRay automotive, and Train Communication Network (TCN) for the automation of trains. The diversity of application requirements (real-time flexible / rigorous, safety-critical, network topology, etc..) Requires a variety of solutions and the use of protocols based on various principles of operation. The result has been a plethora of networks for many specific applications.
Typical network architecture in the area of \u200b\u200bindustrial automation field l
Given the communication requirements applications, field area networks, unlike LAN networks-often have low-speed data transfer, handling small data packets and generally have to operate in real time, which may require deterministic data transfer or limited time.
However, in field networks are already common data transfer speeds above 10 Mbit / s, as in the LAN. Area networks for industrial automation field (as opposed to the building automation and control) have no great need for routing functions or control from end to end. Accordingly, these networks are used only layers 11 (physical layer), 2 (data link layer, including implicitly layer media access control) and 7 (application layer, including user layer) model Reference ISO / OSI.
To ensure deterministic response is to use appropriate planning schemes, often in real-time operating systems for specific application domains or real-time applications, "naked" (minimum) and personalizadas.Los interconnected embedded systems for applications where security is critical, such as x-by-wire, electro adopting solutions -mechanic to replace mechanical or hydraulic solutions for electrical / electronic systems, they have a high degree of dependence that ensures system security failures. He pointed to the aircraft flight control and driving cable cars, whose failure can endanger human life, property or the environment.
To avoid these risks is necessary to provide reliable services Safe, on user request. The dependence of x-bywire systems is a major requirement, but also a restriction when taking this type of system.
CRF
Lenny Z. Perez M
0 comments:
Post a Comment