Physical artificial intelligence (PAI) enables machines to perceive, reason, and act within the real world, bridging the gap between digital AI (DAI), sometimes called virtual AI, and physical action. PAI often leverages spatial artificial intelligence (SAI) technology. PAI applications span numerous industries, from basic automation to autonomous vehicles and complex surgical procedures. PAI applications represent […]
Artificial intelligence/ML
Multimodal AI workloads run on single processor
The MIPS S8200 is a RISC-V neural processing unit designed to run transformer-based and agentic AI models directly on autonomous edge platforms. Developed with a software-first architecture, it supports early model optimization on virtual platforms and targets embedded transportation, robotics, and mission-critical systems requiring multimodal AI processing. The device is sampling now, with silicon reference […]
What is physical artificial intelligence and why is it important?
Physical artificial intelligence (PAI) refers to AI systems that can perceive, understand, reason about, and interact with the physical world in real time through sensors and actuators. Unlike digital AI (DAI), which operates in virtual domains, PAI powers tangible actions in dynamic, real-world environments. PAI is used in closed-loop systems where the AI model not […]
Cloud connectivity for edge AI: bridging the demo-to-deployment gap
Most edge AI demonstrations operate flawlessly in controlled environments with stable networks, predictable traffic, and carefully managed credentials. In contrast, many production deployments fail under real-world conditions. This technical article outlines the best practices required for reliable edge AI deployment. It covers bandwidth planning for peak conditions and buffering strategies that maintain stability during degraded […]
Expanded HMI models for defect detection
Lattice Semiconductor announced the latest release of the Lattice sensAI solution stack, delivering expanded model support, enhanced AI performance, and greater deployment flexibility for a wider range of edge applications. With new purpose-built models, upgraded tools, and improved compatibility, Lattice sensAI empowers developers to achieve higher performance, lower power consumption, and faster integration of AI […]
Evaluation boards and SDK available for developers
RRenesas Electronics Corporation has released silicon samples, full evaluation boards, and the RoX Whitebox SDK for the R-Car X5H, a 3nm, multi-domain automotive SoC designed to consolidate ADAS, IVI, and gateway workloads on a single device. The R-Car X5H delivers up to 400 TOPS AI performance, 4 TFLOPS-class GPU capability, and over 1,000k DMIPS from 32 […]
42 V I2C power monitors target portable device telemetry
Microchip Technology has introduced the PAC1711 and PAC1811 digital power monitors for battery-operated and energy-constrained systems that need continuous power measurement with minimal measurement overhead. Microchip states the devices consume about half the power of comparable solutions under typical operating conditions at a 1024 samples-per-second rate, while providing alert functions to reduce how often a […]
Embedded IDE adds in-editor AI for code review
MIKROE has released an updated version of its NECTO multi-architectural IDE that adds integrated AI assistance inside the code editor. The new capability is intended to support embedded developers during implementation and review by enabling context-aware help while working directly with target boards and microcontrollers. Within the editor, the NECTO AI Code Assistant can repair […]
Compact module targets edge systems requiring on-device AI processing
congatec has introduced its first COM-HPC Mini computer-on-module based on Qualcomm Dragonwing IQ-X Series processors. The conga-HPC/mIQ-X uses the Qualcomm Oryon CPU and an integrated Hexagon NPU to support local AI workloads, including machine learning inference and large language model execution. The module delivers up to 45 TOPS of AI performance and is intended for […]
AI chatbot supports PLC selection and configuration tasks
TRi PLC has released an AI-based chatbot designed to provide faster access to product information and technical support resources. The system uses natural language processing and a structured product knowledge base to interpret user questions and return detailed, context-relevant information. The chatbot assists engineers and system designers in identifying appropriate PLC models by comparing application […]
