High Bandwidth Connectors for AI at the Edge

Introduction

AI has impacted the way we live our lives in a myriad of different ways − including making it easier to contact friends and family on social media and providing digital assistance via smart homes and mobile devices. You can find it in self-driving vehicles, retail outlets, search engines, and recently, IoT (Internet of Things) devices. Connected devices are traditionally outfitted with sensors and other hardware that gather information and communicate over the internet or Wi-Fi. When coupled with AI, those devices can have that collected data analyzed in near real-time and stored in the cloud for later use.

Typically, AI processes collected data externally within the cloud or data centers, but as microprocessors have become more powerful, the machine-learning processing can be done locally without any outside help. This process is also known as Edge AI, meaning the hardware device does not need to be connected to the internet to process information. The technology has become pervasive throughout the medical, transportation, manufacturing, service industries, and consumer markets, which will help to save lives, reduce costs, improve decision making, and help to create next-gen technologies.

Benefits of Edge AI

According to an article from Avnet on the AI at the Edge Campaign, Edge AI offers significant advantages, including built-in security, less network congestion with decreased data flow, high flexibility, and the ability to be customized to suit the application. It also allows developers to distribute computing across networks by offloading some activities to the edge processors within that network, including routers, gateways, and servers.

That said, Edge AI can only function efficiently if all the hardware is working collectively, even down to the interconnect and connector levels, which often require large bandwidths to handle the amount of data that's being collected and processed. Interconnects can be a huge bottleneck that can, and will, decrease efficiency in the delivery of data if not appropriately designed. Molex offers a line of high-speed, high-bandwidth interconnects and connectors that mitigate the bottleneck across the signal chain in IoT devices, for increased efficiency and performance without degradation.

Connectors and Edge AI

Molex has joined with Avnet for its AI at the Edge Campaign to make significant advancements in nearly every industry, which will enable Edge AI devices to maintain data collection and efficiency via trusted and tested connectors, interconnects, and other technologies designed to provide solutions across a broad spectrum. To that end, Molex is providing the automotive manufacturers with the tools to advance AI applications for connected and autonomous vehicles, including solutions for telematics, Vehicle-to-Everything (V2X), Firmware Over-the-Air (FOTA) security, and more.

Molex offers a myriad of different connector options for automation manufactures and Edge AI devices, such as the Micro-Lock Plus Wire-to-Board System series, which is ideal for compact applications and offers electrical and mechanical reliability in a high-temperature package. It's one of the smallest on the market and features a 1.25mm pitch connector with a positive lock mechanism. The company's Floating Board-to-Board SlimStack FSB5 0.40mm Pitch Floating connector for automotive applications features a large floating range with space-saving and design flexibility for a wide range of applications.

Molex also offers several other connectors for automotive applications, including the Pico Clasp Wire-to-Board System with core 1.00mm pitch, which ranges in circuit sizes from 2 to 50 and provides wide header versions, ideal for space constrained applications. They also offer a variety of plating (including tin and gold) and design options to suit the manufacturer's requirements.

Molex also offers a wide range of industrial solutions for AI applications, including machines that can perform preventative maintenance, and product lines that use sensors and machine learning to collect valuable data to improve efficiency and product integrity. Industrial and manufacturing plants have been transitioning over to Industry 4.0/5.0 technologies with Wi-Fi and Bluetooth communications, higher resolution displays, and AI-equipped "cobots" (collaborative robots) that work alongside humans with high levels of safety. In addition, the company's M8 and M12 connectors and cordsets are ideal for industrial environments as they provide an IP67 NEMA rating (when mated), and are IEC compliant.

As with those automotive Edge AI applications, the technologies used in the industrial sector must meet strict demands for functioning in harsh environments. Manufacturing plants, for example, require robust, high-speed connectors that meet high-performance demands, for which Molex offers a wide selection, including FFC/FPC Easy-On connectors that can handle 125^oC operating temperatures.

Conclusion

As Edge AI continues to advance, it is important to note that being successful at deploying the latest technologies requires understanding the direction of computing, architecture, and building next-gen AI-enabled applications, taking full advantage of the tools available within the Edge ecosystem. Avnet's partnership with Molex for the AI at the Edge Campaign will bring about the latest innovations in smart Edge hardware and software for any industry looking to remain competitive in an ever-evolving world of connected IIoT devices.

Examples of Edge AI Connectors