Conga-SMX8 SMARC Modules

Main articles:

• Virtualization. Classification and applications
• What is the Internet of Things (IoT)
• CPU
• SMARC - Generic Small Form Factor Computer Module Specification

2021

SMARC computer-on-module platform with PCIe video camera extension

On October 5, 2021, congatec and MATRIX VISION introduced the computer-on-module (Computer-on-Module, COM) SMARC platform with an extension in the form of a video camera module based on the PCI Express (PCIe) computer bus. Without additional overhead and the need for additional interfaces such as GbE, USB, or CSI MIPI, these images are written directly to the SMARC RAM, with little or no delay and with more bandwidth.

congatec introduced PCIe-based machine vision technology

As it was reported, modules of Sony Pregius video cameras on the basis of MATRIX VISION sensors transfer data of the image with the speed up to 226.5 frames per second (FPS) and with resolution of 1.6 megapixels to the SMARC module of the congatec company which is executed on base processoraintel by Atom. Such a high-speed transmission provides an accurate real-time mode with a clock frequency having a period of approximately four milliseconds. This synchronization is also suitable for Time Sensitive Networking (TSN) instructions to actuators that provide accurate real time at a clock frequency less than one millisecond. Typical uses of the presented solution can be found in industrial machine vision applications in electronics and semiconductors, automotive, food and beverage, pharmaceuticals, packaging and printing. Other target markets include health care, intelligent transportation systems (ITS), and airport security and surveillance systems.

PCIe based camera implementations are designed for low-latency real-time machine vision applications. Firstly, and this is one of the reasons is that unlike GbE, USB or MIPI, there is no additional overhead in such a solution. Secondly, the interface is always initially supported by the processor, which does not always happen with GbE, USB or MIPI. explained Martin Danzer, Director of Product Management at congatec

The ability to use multiple lines in parallel makes PCIe performance highly scalable for multi-camera system solutions, while maintaining low overall system costs. PCIe also provides investment security in the future, since this bus is inextricably connected to the bus. processor x86 Uwe Hagmaier, Head of Research and Development at MATRIX VISION

A full-scale demonstration of the solution will be held in Stuttgart, which will demonstrate work with four camera modules. It will be conducted for SMARC modules Intel Processors with Atom, Intel Pentium Intel and Celeron (code names Elkhart Lake and Apollo Lake). Variants with SMARC modules based on the processor are also available. The NXP i.MX8 M Plus MATRIX VISION mvBlueNAOS family of camera modules uses Sony Pregius and Pregius S series frame photosensor sensors. A wide range of processor architectures available in SMARC are supported by mvIMPACT Acquire. SDK The manufacturer of GenICam GenTL ensures compatibility with existing developments and smooth switching between different hardware platforms. Additional packages for LabVIEW, DirectShow, VisionPro, and Halcon are also available.

Developers interested in evaluating PCIe graphics cards of the mvBlueNAOS family in combination with congatec SMARC modules based on Intel Atom, Intel Pentium and Intel Celeron processors, as well as NXP iiMX8 M Plus processors, can choose between six different camera models with different resolutions - from 1.6 MP (1456 x 1088) up to 24.6 MP (5328 x 4608) and frame rates from 226.5 to 24.1 frames per second.

Starter Kit for conga-SMX8-Plus Embedded Machine Vision Applications with intel MX 8M Plus Processor

On June 8, 2021, congatec announced the expansion of its ecosystem, the MX 8, with a launch kit for embedded intelligent machine vision applications with accelerated artificial intelligence (AI).

Starter Kit for AI Acceleration Embedded Machine Vision Applications

Based on computer the Computer-on-Module (COM SMARC processor) with the MX 8M Plus, this solution is a suitable variant of the startup kit for using NXP the neural processing unit (NPU) integrated into the company's processor. With up to 2.3 TOPS (trillions of operations per second) for deep-learning artificial intelligence, it can run output mechanisms and libraries such as Arm neural network Neural Network and TensorFlow Lite. The kit also integrates with for software the embedded machine vision company Basler, providing OEMs to develop next-generation embedded vision systems with acceleration based on artificial intelligence, a working platform of ready-made applications. Typical applications of this solution range from cost-sensitive automated cash terminals in retail to to trade building safety systems, and from systems for to video surveillances navigation bus surveillance systems. Industrial use cases include vision-based man-machine interfaces identification with user and machine control based on gestures, as well as robotics support for machine vision and industrial systems for visual quality control.

A dedicated unit for processing neural algorithms, supported by open source AI software solutions such as TensorFlow, is an efficiency accelerator for many machine vision systems. When all this is integrated as a fully ready-to-use, proven platform hardware and software, including Basler's pylon Camera Software Suite, it allows developers to quickly develop applications for intelligent machine vision with NPU acceleration. explained Martin Danzer, Director of Product Management at congatec

software The Basler pylon Camera Software Suite includes a unified SDK BCON for MIPI USB3 and GigE cameras, and provides access to cameras directly from source code, GUI, or third-party software. The viewer built into the pylon Camera Software Suite is suitable for evaluating the camera. With the integration of the pylon Camera Software Suite into congatec's iMX 8M Plus Artificial Intelligence Accelerated Machine Vision Starter Suite, engineers gain access to important artificial intelligence-enabled machine vision features. Among them are launch, individual image capture and various camera configurations, as well as access to customizable outputs and algorithms based on the Arm Neural Network and TensorFlow Lite ecosystem.

The starter set for accelerated AI machine vision applications includes the entire ecosystem developers need to immediately begin developing applications based on this next-generation platform that offers machine vision and its integration with AI. The set is based on SMARC 2.1 computer on the conga-SMX8-Plus module. To accelerate deep learning algorithms, it is equipped with four Arm Cortex-A53 cores, one Arm Cortex-M72 and NXP NPU controller and comes with passive cooling. The 3.5-inch conga-SMC1/SMARC-ARM carrier directly connects the 13-megapixel Basler dart daA4200-30mci BCON camera for the MIPI camera to the F1.8 f4mm lens via the CSI-2.0 MIPI without any additional converter modules. Along with MIPI CSI-2.0, machine vision cameras with USB and GigE connectivity are also supported. For software, congatec provides a boot SD card with a pre-configured bootloader, thus the Yocto OS with the appropriate BSPs and already optimized for the processor with integrated machine vision software from Basler, allows the system to learn the logical output of AI based on captured images and video sequences.

Сonga-SMX8-Plus

On March 2, 2021, the company, congatec a supplier of technology for embedded and, peripheral computing introduced its - computer on-module with small own consumption, energy made according to specification 2.1 SMARC and equipped with the iMX processor 8M Plus from NXP - the first representative of the iMX family with integrated (neural processor neural processing unit, NPU) developed for modern peripheral industrial applications and devices () Internet of Things IoT running on the basis. machine learning

Conga SMX8 Plus

According to the company, this congatec module is focused on industrial peripheral analytics, built-in machine vision and artificial intelligence (AI) systems. With machine vision and deep learning capabilities, the conga-SMX8-Plus low-power module allows industrial embedded systems to see and analyze their environment for situational awareness, visual inspection and product control, identification, surveillance and tracking, and non-contact gesture-based machines and augmented reality systems.

The technical features quad-core processor of the platform from ARM Cortex-A53 the company NXP include an integrated neural processing unit (NPU) for computing power of the AI layer and a specialized digital signal processor used for image processing (ISP), designed for parallel processing of images with proper resolution and video streams from two built-in interfaces of the MIPI-CSI camera. The vast ecosystem of this SMARC module, such as already fully ready-to-use 3.5-inch carriers, as well as support for Basler video cameras and artificial intelligence software stacks, simplifies the launch of the product to quickly test the concept of its final solution. Vertical markets for these credit card low-energy and artificial intelligence-sized modules can be found, literally, anywhere - from smart agriculture and industrial production to retail, from, trade to and transport smart smart cities buildings.

The congatec SMARC module, based on NXP's iMX 8M Plus processor, additionally equipped with a variety of specialized processors, provides fast, built-in, low-power vision and artificial intelligence applications. Modules from congatec include:

• The Neural Processor (NPU) adds 2.3 TOPS (tera-operations per second) of dedicated AI processing power for four powerful multipurpose ARM Cortex-A53 processor cores.
• Integrated to optimize video quality, the digital signal processor used for image processing (ISP) processes up to three full HD video streams with a screen refresh rate of up to 60 frames per second.
• The Digital Signal Processor (DSP) provides local voice recognition without connection to the cloud.
• Cortex-M7 - as of March 2021, the most productive platform among all Cortex-Mx cores, except for its main function, it can also be used as a fault-tolerant device, provides real-time control together with an Ethernet network port with timing.
• Finally, it is cryptographic the CAAM accelerator module for hardware acceleration enciphering based on electric curves (Elliptic-curve cryptography, ECC) and hardware storage for keys. RSA In addition, Arm TrustZone technology is used, which is a secure area that functions separately from everything else, and which integrates software the domain Resource Domain Controller (RDC) at boot time to isolate execution of critical software safe and High Assurance Boot modes.

The SMARC 2.1 modules for machine vision and artificial intelligence applications are equipped with four different quad-core processors, the MX 8M Plus from NXP based on ARM Cortex-A53 have versions as for industrial (from 0 ° C to + 60 ° C), and expanded (-40 ° C to + 85 ° C) temperature range, as well as built-in ECC for LPDDR4 memory up to 6 GB. The modules can control up to three independent displays and provide hardware acceleration of video decoding and encoding, including H.265, so that streams from the proper resolution camera transmitted by two integrated MIPI-CSI interfaces can be sent directly to the network. To store data, engineers will find up to 128 GB of embedded non-volatile eMMC memory on the board, which can also work in safe pSLC mode. Peripherals include 1x PCIe Gen 3, 2x USB 3.0, 3x USB 2.0, 4x UART, and 2x CAN FD and 14x GPIO. For real-time networking, the module offers one Gigabit Ethernet port with TSN support plus regular Gigabit Ethernet. An additional M.2 Wi-Fi and Bluetooth LE card soldered to the module adds wireless capabilities. complement the 2x I2S audio feature set. Supported operating systems include Linux, Yocto 2.0, and Android.

2019: Conga-SMX8-Nano with ix MX 8M Nano processor

On December 3, 2019, congatec introduced the SMARC 2.0 computer module with the niMX 8M Nano processor of NXP Semiconductors, which is also equipped with Cortex-A53 cores. The Conga-SMX8-Nano module defines the lower price segment class for SMARC modules. Due to the high graphics capabilities along with low power consumption and a limited but carefully selected number of input/output (I/O) ports, the NXP iX MX 8M Nano processor, compatible with the NXP iX 8M Mini, is designed for applications with a limited financial budget and is suitable for - independent of the manufacturer, standardized computers on modules (Computer, CoM-on). Based on the logic of the mobile portable device itself, a device with a power consumption of not more than 2 watts is required here. With this offer, end-application engineers can also turn a complex, for example, medical user interface with various mechanical buttons into an intuitive touch-sensitive and easy-to-use graphical interface or provide industrial devices with an interactive screen for which earlier it would be just too expensive a pleasure. All this, and even to systems without user interface (so-called "headless system", literally - headless systems, the computer without keyboard and a mouse) with support protokolaIEEE 1588 Precision Time Protocol (protocol of the exact time) for tactile IIoT (Industrial Internet of things), are possible also at the minimum expenses regarding power food, and at the modest budget, claim in congatec.

As noted in congatec, thanks to the extensive ecosystem of computers on SMARC specification modules, engineers benefit from much more advanced and ready-to-use components, standardized APIs, and complex BSPs. Another important advantage is their wide scalability, which goes far beyond the compatibility of the outputs of the NXP i.MX 8M Nano and Mini variants. Typical applications of the presented modules are graphical interfaces for peripheral devices, which, if necessary, may also include voice assistance and voice control. Vertical markets include wireless and network medical and industrial equipment, and household appliances and electronics, digital signage and mobile or outdoor devices, including infrastructural cities, such as small form factor signage or platforms without a user interface for smart city border gateways, such as electronic charging and short-term rental of vehicles with the possibility of ending the trip at user-friendly points and places.

"The availability and broad scalability of applications are the two main determinants for engineers to migrate to PCs with a module. ARM They want to quickly enter the market, save the cost of one-time research and design work, and optimally balance the price-performance ratio for their low-cost, high-volume product families. With computers on SMARC specification modules, congatec offers all this in a completely ready-to-use form. They can scale application-ready platforms across the entire myMX 8 portfolio and beyond, toward a wide variety of low-power application processors. This hardware-independent advantage with its inherent long-term availability is another important reason why engineers tend to use ARM architecture processors in their end applications, and increasingly rely on proven design principles for standardized "Computer on Modules" modules. noted' Martin Danzer, Director of Product Management at congatec

According to the manufacturer's statement, computers based on SMARC specification modules based on NXP's iMX 8M Nano processors are already fully operational subsystems that come with a comprehensive ecosystem that includes a ready-to-use boot loader implementation (boot loader), pre-qualified BSPs for Linux, Yocto and Android and full media boards to evaluate their capabilities. Support for personal integration from congatec and a wide range of customized technical services simplify the integration of the proposed ARM-based module of the NXP processor for congatec customers.

According to the manufacturer, conga-SMX8-Nano SMARC computer modules with the NXP i-MX 8M Nano processor are designed for applications that require ultra-low power and are sensitive to the cost of graphics interfaces, as well as for industrial control devices without a user interface, including support for the IEEE 1588 exact time protocol for tactile. IIoT According to December 2019, they are available in three versions with a quad-core, dual-core or single-core processor of the ARM Cortex-A53 architecture, each version is accompanied by one Cortex-M7 core. Each version is available for an extended (from 0 ° C to + 60 ° C) and industrial (from -40 ° C to + 85 ° C) temperature range. Optional, processor-integrated GPU GC7000UltraLite 3D supports two shaders as well as OpenGL, OpenCL, and Vulkan specifications. Embedded displays can be connected via a dual-channel LVDS, eDP, or MIPI-DSI port. The necessary large volume of memory of the module is provided at the expense of LPDDR4 up to 2 GB with low energy consumption and non-volatile memory of eMMC 5.1 with a capacity up to 128 GB. Embedded cameras are connected through the MIPI-CSI-2 interface, while four USB 2.0 and three UARTs are modern interfaces necessary for industrial applications. For inter-system connectivity, the module offers one Gigabit port, Ethernet as well as an additional M2/extension. Wi-Fi Bluetooth

2018: Conga-SMX8 with 64-bit ARM processor

On July 31, 2018, the company congatec announced the SMARC 2.0-module conga-SMX8, made on the basis of a 64-bit multi-core - ARMprocessor the ix MX8 family, manufactured by the company. NXP

conga-SMX8 SMARC 2.0 Module

According to the company, the conga-SMX8 modules based on the ARM core Cortex-A53/A72 are the flagship for embedded low-power computer boards, offering the next ARM processor, featuring high performance, flexible graphics and built-in functions, which will be a suitable solution for all types of industrial Internet of Things (Industrial Internet of Things, IIoT) applications. The offered computer modules provide high-performance multi-core computing, as well as advanced graphics capabilities for three independent displays with a resolution of 080p or one screen with a resolution of 4K.

Additional advantages of this industrial-level platform include real-time hardware support and hypervisor, as well as wide scalability and also stability of operation under harsh environmental conditions and an extended range of operating temperatures. All of these features position SMARC 2.0 modules as meeting performance and functionality requirements for low-power embedded industrial and IoT applications, as well as for the mobile market sector.

It was also reported that SMARC 2.0 modules with NXP's mobile MX8 processors, with their hardware virtualization and distribution of computing resources, are of interest to a wide range of stationary and mobile industrial applications, including robotics that require real-time control and motion control systems. Since the modules are designed for an extended range of ambient operating temperatures, from minus 40 ° C to 85 ° C, they can also be used in systems of various commercial vehicles or in infotainment applications in passenger cars, buses and trains, and they can also be successfully used for systems of all electric and autonomous (unmanned) vehicles.

conga-SMX8 modules have up to eight processor cores (2x A72 + 4x A53 + 2x M4F), up to 8 GB of LPDDR4 MLC or pseudo-SLC memory, as well as up to 64 GB of non-volatile memory per module. The extended set of interfaces includes two GbE, including optional IEEE1588, clock synchronization over the computer network, up to six ports, USB including 1x, and USB 3.1 also up to 2x PCIe Gen 3.0, 1x SATA 3.0, 2x CAN bus, 4x UART. An additional Wi-FiBluetooth built-in/module with Wi-Fi 802.11 b/g/n and BLE is also available. Up to three displays can be connected to the module via HDMI 2.0 with HDCP 2.2, 2x LVDS, and 1x eDP 1.4. As for video cameras, the proposed modules support two MIPI CSI-2 video inputs. NXP's SMARC 2.0 processor-based MX8 modules come as complete superkits that are fully integrated into applications, including the Uboot loader and board support packages on, OS Linux Yocto, and. Android

According to representatives of the company, since congatec is a member of the early access program for NXP products, the congatec SMARC modules will be available in mass production at the same time as the launch of the next niMX8 processor family, which is planned by the end of 2018.