Embedded AI connects at Electronica 2018
Some 80,000 visitors descended on Munich last week for one of the world’s biggest electronics trade shows — Electronica — which takes place every two years. EE Times and Aspencore were out there for the week, and a key part of our participation was running the Embedded Forum, as well as going out and talking to exhibitors and senior executives in the industry to find out what was new and relevant.
The theme this year was “Connecting everything — smart, safe & secure." We present highlights of some of the major chip companies taking part.
STMicroelectronics continued its focus on industrial and automotive applications, with technologies for machine condition monitoring and predictive maintenance using artificial intelligence (AI)-enhanced technologies running on STM32 microcontrollers, a test bench using AI to identify defects on a railway, 48-V automotive subsystems, 77-GHz radar, secure and tamper-resistant telematics, and automotive subsystem modules including LED lighting.
In advanced driver assistance systems (ADAS) and secure infotainment and telematics, it had its in-cabin imaging for driver monitoring, radar sensing, precise positioning, V2X communication, and over-the-air (OTA) system updating. For electric vehicles (EVs) and hybrids, silicon-carbide–based solutions for EV traction inverters, battery management systems (BMS), on-board chargers, and acoustic vehicle alert systems (AVAS) were all a key focus.
NXP Semiconductors featured its edge-based ML solutions for Industry 4.0 as well as its electric vehicle battery management and power inverter platform, plus its new security MCUs. The ML applications included facial recognition for access control, object recognition for operator safety, local voice control commands, and artificial intelligence (AI) derived anomaly detection for predicting and preventing failures of industrial systems. Demos included edge computing, neural-net object detection with NXP’s i.MX 8QuadMax, and i.MX 8QuadXPlus applications processors consuming data from a camera and applying a convolutional neural net (CNN) engine to classify object types.
Infineon showed its electromobility solutions, including the all-electric concept microbus in its partnership with Volkswagen, and IONITY charging stations for rapid charging to enable several hundreds of kilometers. The company also debuted its CoolGaN product portfolio for use in data transmission (telecom) and servers, enabling savings in energy and, hence, CO2.
For autonomous carrier drones, Infineon showed its AURIX microcontrollers as flight controllers and CoolGaN solutions for engine control. Also introduced was the LITIX Basic+ LED driver family addressing the American market and German car manufacturers who are now demanding LED drivers that can detect single LED short circuits.
Renesas was focused on edge AI and higher intelligence at the endpoint, showing examples of AI inference in real time at the endpoints. In its embedded AI demos, it was racing cloud AI, edge AI, and embedded AI against each other to compare speed and performance; another demo showcased the single RX microcontroller (MCU) simultaneously running sensorless vector motor control and executing predictive motor failure analysis; a third demo highlighted image recognition using a RZ/A2M microprocessor (MPU), which incorporates Renesas’ dynamic reconfigurable processor (DRP) that executes both machine vision and AI in real time.
>> Continue reading page two of this article on our sister site, EE Times: "Highlights of electronica 2018."