Embedded
Xilinx extends ecosystem to reshape the future of embedded vision,
IIoT system design
Broad IP, software, hardware and design services offerings enable smarter, connected, highly differentiated systems based on All Programmable devices. Aaron Behman, director, video & vision, corporate strategy & marketing, Xilinx, and Dan Isaacs, director, industrial IoT, corporate strategy & marketing Xilinx tells us more
S
ystems with unprecedented levels of software-based intelligence, optimised hardware and any-to-any connectivity
are shaping the future of embedded vision and the Industrial Internet of Things (IIoT). A stronger and broader ecosystem, which supports development of IIoT and embedded vision systems based on All Programmable devices, makes it easier for users to develop smarter, connected and highly differentiated systems. The exciting applications emerging in industrial/embedded vision and the IIoT cut across the industrial, scientific, medical, pro A/V, consumer, aerospace and defence, and automotive market segments. In areas like these, a key barrier to using the superior performance and performance/watt characteristics of All Programmable devices has been the programming model. C/C++ users are more accustomed to writing code for CPUs and GPUs. With the emergence of High-Level Synthesis (HLS) for software-defined hardware and SDx environment for software-defined systems development, many more system developers can make use of software- defined All Programmable SoCs and MPSoC families. Such devices are now as easy to use as CPUs and GPUs, but with superior performance.
Embedded vision and IIoT systems have a lot in common, both start with sensing and data acquisition. For embedded vision
systems, data takes the form of a series of images or a video stream, but sensed data for IIoT systems also encompasses a long list of additional sensed parameters.
Rising need for sensor fusion Several embedded vision and IIoT systems require sensor fusion - the processing and merging of data from multiple and different types of sensors into actionable intelligence. For embedded vision systems, multiple video streams may be combined to produce more usable or more-useful video streams. For example, vehicle-based vision systems often combine video streams from four, five, six or more video cameras to produce a single bird’s eye view that gives the driver 360° 2D planar or 3D spherical vision. Vision systems drive local displays but also send locally processed video to the cloud for further processing, for combination with other video streams and for storage. IIoT systems may combine video with additional sensed data to define needed actions. For example, the new CPPS- Gate40 Smart Gateway from System-on- Chip engineering (SoC-e) incorporates a variety of I/O ports commonly used in industrial control systems, combined with local, high-speed data processing, and places the resulting data on a dual- redundant optical Ethernet ring using High Availability Seamless Redundancy/Parallel Redundancy Protocol (HSR/PRP).
Figure 2: This 4K multichannel/multisensor video surveillance system taps the safety and security capabilities of the Zynq UltraScale+ MPSoC
A defining characteristic of IIoT systems is the ability to use sensed data for high- speed, real-time control. This is not possible if relying on cloud-based processing and decision-making. Although it can be achieved using a CPU or GPU, systems based on SoCs and MPSoCs offer significant advantages and benefits, such as: highest performance/watt; sensor fusion capability; any-to-any connectivity; multilevel security and multilayer safety.
Advanced driver assistance system An advanced driver assistance system (ADAS) combines video data from several video cameras and additional vehicle sensor data, including inertial navigation and even GPS map data, to make decisions about braking, steering and driver alerts. This design leverages the heterogeneous processing capabilities of the quad-core ARM Cortex-A53 application processor and dual-core ARM Cortex-R5 real-time processor. Any-to-any connectivity enables the system to communicate with any sensor interface, including MIPI for the video cameras. Superior hardware-based video-processing performance allows it to handle more video channels and provides a programmable number of video streams.
4K video surveillance
Figure 1: This ADAS design leverages the heterogeneous processing capabilities of the ARM Cortex cores in the Zynq UltraScale+ MPSoC
30 July/August 2016 Components in Electronics
For high-performance video surveillance, the MPSoC connects to multiple sensors, including different types of video cameras. MIPI-interfaced video cameras and displays are supported and various I/O interfaces connect other sensor types. The performance/watt metric for this Chameleon All Programmable system is around 5x better than comparable systems designed with CPU/DSP/GPU-based silicon. In addition, safety and security features, including ARM TrustZone capabilities and hardware-based AES encryption, are especially useful for security applications like this.
Smart-grid substation automation A substation automation system targeting smart-grid design involves multiple Ethernet streams from various sensing components monitoring substation parameters. It needs to connect to a large number of interface units throughout the substation using standardised IEEE-1588 Precision Timing Protocol (PTP) and IEC 62439 HSR/PRP. Data from the various sensor sources can be processed in high-speed IP blocks. Alternatively, processing algorithms can run on one or more of six available ARM processor cores, depending on performance requirements.
Industrial automation
A typical application for industrial control might take the form of a motion controller, programmable logic controller (PLC) or human-machine interface (HMI) system. Here, the MPSoC integrates an entire system that would otherwise require four chips (CPU, functional-safety processor, shaft encoder and FPGA for high-speed power modulation and motor control) into one device. The result is a 30 per cent improvement in performance/watt and a substantial reduction in system PC board real estate. Such an industrial control system benefits from any-to-any connectivity and from the functional-safety features embodied in dual-core ARM Cortex-R5 processors.
The ecosystem lowdown These examples make ample use of hardware and software IP from vendors and their ecosystem member companies. Chameleon platforms that pick and choose which IP to use within each product ease the task of creating advanced, intelligent systems for IIoT.
www.xilinx.com www.cieonline.co.uk
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60