Збирач потоків

Courtesy: Rhode and Schwarz

Quality Evaluation of Speech Coding Technologies

A comprehensive quality test was conducted to evaluate the perceived quality of various speech coding technologies under realistic conditions. The study compared current mobile network codecs with traditional low-bitrate codecs and emerging AI-based ultra-low bitrate speech coding solutions.

In the test, a set of German speech samples spoken by various speakers was processed through each codec type. A controlled listening experiment was applied to assess overall speech quality with respect to the naturalness of reproduced speech, combined with typical transmission impairments such as packet loss and bandwidth constraints. The evaluation aimed to reflect real-world usage scenarios, including mobile calls, popular IP-based voice services, and speech transmission over satellite links.

To achieve statistically meaningful results, a formal listening test was conducted in a standardised acoustic environment following the ITU-T P.800 methodology using the Absolute Category Rating (ACR) approach. A total of 32 participants – men and women from various age groups – were invited to rate the speech samples. The test ensured balanced demographic representation and controlled conditions to obtain reliable subjective quality scores. Participants evaluated multiple samples per codec type, and the results were statistically analysed to identify significant differences in perceived quality.

Key categories included:

• Modern Mobile Codecs: Including EVS and AMR-WB, which are widely deployed in LTE and 5G networks. Additionally, OPUS (used in WhatsApp) and Satin (used in MS Teams) were considered under real transmission conditions. These codecs offer high fidelity and robustness, especially under variable network conditions.
• Legacy Low-Bitrate Codecs: Such as MELP and LPC-10, and the amateur radio codec Codec2, representing earlier generations of strong speech compression. These codecs were originally designed for extremely bandwidth-constrained environments and are still used in specialised applications.
• Ultra-Low Bitrate AI-Based Codecs: Leveraging deep learning models for end-to-end speech representation and reconstruction. The tested codecs operate in the bitrate range of approximately 600 bit/s to 3 kbit/s. For comparison, 600 bit/s is only one hundredth of the well-known ISDN transmission rate (64 kbit/s) and just one fortieth of the bitrate typically used in VoLTE (24 kbit/s).

Ultra-low bitrate codecs are of particular interest for use in satellite-based communication systems (e.g., Non-Terrestrial Networks, NTN) in Direct-to-Cell or Direct-to-Device mode (smartphones receive signals directly from satellites), where bandwidth is highly constrained, and latency is critical. They are also relevant in military and tactical communication scenarios, where efficient spectrum usage and resilience to transmission errors are essential.

Performance of AI-Based Codecs

The new AI-based codecs support 8 kHz wideband and 12 kHz super-wideband audio and demonstrate a significant leap in perceived speech quality and naturalness compared to classical low-bitrate codecs. Some AI-based solutions approached the performance level of high-quality codecs such as AMR-WB and EVS, making them promising candidates for future communication systems under strong bitrate constraints or high network load situations. The computational complexity of these codecs was not investigated in this study; however, some implementations introduce only a short delay that is acceptable for use in real-time communication.

These codecs deliver speech that sounds natural and pleasant to the listener without question. However, they do not always reproduce all speaker-specific characteristics with full accuracy. For example, pitch and intonation may be slightly altered, and in some cases, initial phonemes or consonants may be replaced or smoothed. While this may be acceptable for everyday conversation, it can limit their applicability in scenarios requiring speaker identification, authentication, or mission-critical communication.

The following table shows some representative results of the listening experiment; the Mean Opinion Score (MOS) rates the subjectively perceived quality on a scale from 1 (bad) to 5 (excellent):

The detailed results of this evaluation, including statistical analysis, codec performance rankings, and listener feedback, are presented at the ITU-T SG12 meeting in September 2025. These insights are expected to contribute to ongoing discussions around codec standardisation, the definition of “quality,” and its automated prediction, particularly in the context of future mobile and satellite communication systems.

The post How good are ultra-low bitrate speech codecs? appeared first on ELE Times.

NXP Semiconductors introduced its NXP CoreRide Z248 zonal reference system – the semiconductor industry’s first pre-validated, design-ready zonal foundation that combines advanced 48 V energy distribution, deterministic data handling, functional safety, and real-time responsiveness. The hardware-software foundation is designed to optimise system performance, reduce system integration effort, shorten development cycles, and allow OEMs and Tier 1s to focus investment where it matters most. It sets a new benchmark for accelerating the journey from zonal architecture concepts to production‑ready implementations.

Built on NXP’s S32K5 microcontroller series, its integrated advanced MRAM technology unlocks ultra-fast, ultra-frequent over-the-air updates throughout the entire vehicle lifecycle. At the software level, the Z248 integrates a comprehensive pre-validated software stack that streamlines complex development of smart data energy network (SDEN) functionalities such as impedance, power and protection monitoring, intelligent data routing, AI‑enabled virtual sensing, diagnostics, and audio.

With its built-in, validated remote protocol stack (RCP), it supports the up-integration of end node functions and ECU consolidation to enable new cost-optimised vehicle architectures. It also addresses key challenges of 48 V zonal systems by managing energy conversion, distribution, and protection within a single, integrated architecture.

The Z248 is rigorously validated through thousands of system-level tests demonstrating outstanding low-power modes, fast boot and fast wake-up response. It is supported by a modern, collaborative continuous integration, continuous testing and continuous delivery (CI/CT/CD) development environment that allows significantly faster test loops with OEMs and tier 1s, shortening validation cycles.

Why it matters: Automakers are being asked to move faster, scale broader, and spend smarter – even as safe zonal consolidation, hybrid power systems, and AI-enabled features dramatically increase architecture complexity. NXP’s new CoreRide zonal reference system brings scalability to this rising architectural complexity. It reduces risk by helping OEMs and tier 1s accelerate development into production, and it eases the switch from legacy platforms and lower total cost of ownership – freeing them from complex integration to put them on a path to production.

“As new E/E architectures redefine vehicle design, our focus is simple: give the automotive ecosystem the foundation to move faster and differentiate with confidence,” said Sébastien Clamagirand, SVP and General Manager, Automotive Systems & Platforms at NXP Semiconductors. “The NXP CoreRide zonal reference system Z248 delivers a performance-optimised, scalable 48 V foundation that intelligently fuses power, data and software, while dramatically simplifying system integration, reducing time to market, and enabling OEMs to focus on vehicle differentiation and long‑term value creation.”

More details: The Z248 zonal reference system is delivered with a complete Board Support Package (BSP) with pre-integrated software from the NXP CoreRide partner ecosystem, including GLIWA’s performance monitoring suite, Green Hills’ software compiler and Vector’s embedded software and tools. The full package undergoes extensive validation to help ensure optimised performance, while continuously improving processing efficiency and power consumption based on the primary use cases of a zonal ECU.

It’s a scalable, safe and secure hardware-software stack that adapts easily to different variants of SDV E/E architectures and integrates naturally with NXP’s broader system offering. It leverages technologies across computing, networking, power management and 48 V energy distribution, including NXP’s S32K566 zonal microcontroller featuring on-chip MRAM that significantly accelerates ECU programming times, both in factory settings and during over-the-air (OTA) updates.

The reference system also integrates 48‑volt‑capable power components such as eFuse, PMIC and DC‑DC converters, robust in‑vehicle networking through Ethernet PHY and CAN transceivers, and built-in audio support. In addition, it introduces a new concept for zonal I/O extension. Designed for broad applicability with housing and a wiring loom, the new NXP CoreRide Z248 zonal reference system can be deployed across ICE, hybrid and BEV platforms, supporting the industry’s move toward zonal processing and ECU consolidation.

Ecosystem Voices

Peter Gliwa, CEO and Founder of GLIWA

“NXP understood that the eco-system, the tooling around a new platform, is essential for its success. With our Analysis Suite T1 built into the NXP CoreRide Z248 zonal reference system, high efficiency, proper timing analysis and timing verification are very well addressed.”

Dan Mender, Vice President of Business Development at Green Hills Software

“Green Hills is proud to play a central role in NXP’s transformative reference solution strategy, which simplifies and accelerates production-focused automotive ECU development through pre-integrated hardware and software optimised for zonal automotive architectures. By leveraging Green Hills’ integrated software solutions, customers can develop high-quality, safety-critical applications with a minimal footprint and optimal performance, while significantly reducing time to deployment.”

Sam Yeh, Chairman of Inventec

“In response to the automotive E/E architecture trend toward zonal and centralised designs, Inventec is collaborating with NXP Semiconductors to support the advancement of next-generation zonal architectures. Through this collaboration, Inventec can provide hardware design and JDM support to OEMs as part of NXP’s zonal E/E architecture initiatives.”

Jochen Rein, Senior Vice President, Business Unit Software Platform at Vector

“The combination of the NXP CoreRide platform and Vector’s software foundation provides a robust basis for next‑generation zonal architectures. We enable our joint customers to reduce their time- to-market due to a pre-integrated and highly optimised software stack.” Vector contributes as an NXP CoreRide partner, providing pre‑integrated software and tools that help streamline development and ensure smooth integration within the zonal ECU architecture.”

The post NXP CoreRide Puts Automakers on Fast Path to 48 V Scalable Zonal Architectures appeared first on ELE Times.

The post Microchip Helps Manufacturers Meet Cybersecurity Regulations, Expands Security Services in the Trust Platform appeared first on ELE Times.

Fuel cell sensors are electrochemical devices designed for precise measurement. In measurement applications, they have become the gold standard for breath alcohol concentration detection, valued for their ethanol specificity, stability, and courtroom-grade accuracy. Compact and low power, they form the backbone of law enforcement breathalyzers, workplace safety programs, and consumer devices, consistently outperforming semiconductor and infrared (IR) alternatives.

Their proven reliability in complex breath matrices has made them indispensable for safety and compliance, while ongoing innovation is extending their reach into broader analytical domains. And while fuel cells generate clean energy, fuel cell sensors generate precise measurements—a distinction that defines their unique role in modern technology.

Applications and history

Before we get into the basics of how fuel cell sensors work, it’s worth noting their application landscape. While research has explored microbial fuel cell biosensors for environmental monitoring and niche industrial uses, the overwhelming commercial reality today is breath alcohol concentration (BAC) measurement.

Fuel cell sensors have become synonymous with BAC detection because of their unmatched ethanol specificity, stability, and courtroom-grade accuracy. Although BAC formally refers to blood alcohol concentration, in practice it is estimated through breath alcohol analysis. This singular focus has defined their role in law enforcement, workplace safety, and consumer devices, making BAC not just their flagship application but essentially their identity in the marketplace.

Technology itself traces its roots to the 1960s, when early electrochemical cells were adapted to detect ethanol in breath samples. By the late 1970s and early 1980s, law enforcement agencies began adopting fuel cell-based breathalyzers, recognizing their superior specificity compared to semiconductor sensors.

Over time, improvements in miniaturization, catalyst stability, and calibration protocols transformed them from bulky instruments into compact, portable devices. This evolution cemented fuel cell sensors as the trusted backbone of alcohol detection, setting the stage for their enduring role in safety and compliance.

Figure 1 A compact breathalyzer with a fuel cell breath alcohol sensor—Alcotest 4000—simplifies portable BAC measurement. Source: Dräger

As a quick aside, while fuel cells rely on chemical reactions, IR spectroscopy uses light to identify alcohol’s unique spectral fingerprint. By directing an IR beam through a breath sample, the instrument measures the specific wavelengths absorbed by ethanol molecules.

This physics-based method is non-destructive and highly precise, enabling real-time detection of “mouth alcohol” that could otherwise distort results. Because of their sophistication, accuracy, and long-term stability, IR units are reserved as definitive, desktop-based instruments in police stations, providing the courtroom-grade evidence required for testimony.

Fuel cell breath alcohol sensors

Now is the time for a gentle dive into a bit of theory and practice. At their core, these sensors operate on an electrochemical principle: ethanol molecules in exhaled breath are oxidized at a platinum electrode, producing an electrical current directly proportional to concentration. This reaction is simple yet elegant, converting chemical energy into a measurable signal that reflects blood alcohol concentration (BAC).

In practice, this design delivers a combination of portability, stability, and specificity that has made fuel cell sensors the dominant choice for breath alcohol testing. Unlike semiconductor sensors, which can be affected by other volatile compounds, fuel cells respond almost exclusively to ethanol.

Their compact form factor allows integration into handheld devices, while their long-term consistency ensures reliable results in roadside, workplace, and consumer contexts. This balance of theory and application explains why fuel cell sensors remain the benchmark technology for BAC measurement today.

In a nutshell, a fuel cell breath alcohol sensor is essentially a pair of platinum electrodes immersed in a dilute acid electrolyte. When a trace amount of ethanol from exhaled breath reaches the electrodes, it undergoes oxidation, releasing electrons that flow as current. The magnitude of this current is directly proportional to ethanol concentration, providing a simple yet highly reliable way to quantify blood alcohol concentration.

And fundamentally, the fuel cell breath alcohol sensor consists of a porous, chemically inert layer coated on both sides with finely divided platinum black. The porous layer is impregnated with an acidic electrolyte solution, and platinum wire connections are attached to the platinum black surfaces. The assembly is mounted in a plastic case with a gas inlet for introducing a breath sample. While manufacturers add proprietary refinements to this design, the basic configuration is shown in Figure 2.

Figure 2 Drawing illustrates the basic construction of a fuel cell breath alcohol sensor. Source: Author

Hands-on with fuel cell alcohol detection

For those eager to explore fuel cell alcohol sensors, the FS00702 electrochemical ethanol content module offers a robust solution. This fuel cell–type sensor operates through oxidation and reduction reactions at the working and counter electrodes, generating charges that form a measurable current. Current’s magnitude is directly proportional to alcohol concentration, in accordance with Faraday’s law, enabling accurate determination of ethanol levels.

Equipped with a high-stability gas sensor and a high-performance microprocessor, the module supports both UART and analog signal outputs for seamless integration. Its precise automatic calibration and advanced detection systems minimize human interference, ensuring consistent accuracy and reliability in large-scale production environments.

Figure 3 Highlighting FS00702 key specs: enabling makers to detect ethanol with precision, rapid updates, and easy microcontroller integration. Source: Henan Fosen Electronics Technology

As a side note worth mentioning, ethanol is one specific type of alcohol—the compound found in beverages and fuels—whereas “alcohol” broadly refers to a family of related molecules such as methanol, propanol, and isopropanol.

Fuel cell sensors like FS00702 are calibrated for ethanol detection since it’s the relevant analyte for intoxication measurement and fuel monitoring. While the sensor may respond to other alcohols, its accuracy is optimized for ethanol, making precise terminology important in technical contexts.

Practically speaking, sourcing high-quality fuel cell alcohol sensors for hobbyist projects is challenging, since most manufacturers prioritize finished breathalyzer units or bulk industrial modules.

Still, there are accessible alternatives to FS00702 for makers who value the accuracy and specificity of fuel cell technology. The Dart Sensors 2-Electrode fuel cell is considered a gold standard for precision, though it requires a custom amplifier circuit.

Fosensor’s FS00701 provides a smaller footprint than FS00702, ideal for portable builds. Meanwhile, FS00702 itself remains versatile, offering both raw analog output for custom conditioning and a built-in UART option for straightforward microcontroller integration.

Winsen’s ZE321 automotive alcohol module offers a compact design with a convenient UART interface, making it more user-friendly for DIY integration. The ZE321 module operates on the fuel cell electrochemical principle. When the built-in pressure sensor detects exhaled air flowing through the sampling tube at the required rate, the solenoid valve quickly opens to admit a measured volume of breath.

Within the sensor, alcohol and oxygen undergo a redox reaction, generating an electrical current proportional to ethanol concentration. The module’s circuitry measures this current and, after algorithmic processing, outputs an accurate determination of breath alcohol content.

Figure 4 The ZE321 automotive alcohol module monitors exhaled breath flow, samples a fixed volume of gas, and actively detects alcohol content through its fuel cell electrochemical reaction. The onboard circuitry processes the resulting current signal to deliver accurate breath alcohol measurements. Source: Winsen

Accuracy today, innovation ahead

In practical terms, fuel cell–based alcohol testing devices deliver the highest accuracy in measuring breath alcohol content, leaving little room for error. Even so, it’s wise to allow for a small margin of discrepancy. When evaluating any alcohol detection instrument—whether for personal safety, workplace compliance, or automotive use—the sensor type is critical. If precision matters most, fuel cell sensor technology remains the benchmark to aim for.

For makers and engineers, the challenge is clear: fuel cell sensors are not confined to alcohol testing; they are gateways to precision sensing, sustainable energy, and inventive applications across domains. Experiment boldly, share your builds, and push the boundaries of what these devices can achieve. The next breakthrough could start on your workbench.

T. K. Hareendran is a self-taught electronics enthusiast with a strong passion for innovative circuit design and hands-on technology. He develops both experimental and practical electronic projects, documenting and sharing his work to support fellow tinkerers and learners. Beyond the workbench, he dedicates time to technical writing and hardware evaluations to contribute meaningfully to the maker community.

Related Content

• Hydrogen Fuel-Cell Vehicles
• Paper-Based Biofuel Cells Power Disposable Electronics
• xEVs Opening New Opportunities for Sensor Development
• Designing a portable system for in situ failure prediction in fuel cells

The post Fuel cell sensors: From breath to benchmark appeared first on EDN.

Everspin Technologies, a leading developer and manufacturer of magnetoresistive random access memory (MRAM) persistent memory solutions, today announced the UNISYST MRAM family, a new generation of unified memory designed to fundamentally change how embedded systems store and access code and data.

“System designers are running into the physical and performance limits of NOR flash, especially as process nodes move below 40 nanometers and workloads become more demanding,” said Sanjeev Aggarwal, president and CEO of Everspin Technologies. “With UNISYST, we are extending our MRAM roadmap to higher densities while giving customers a practical way to start with PERSYST today and migrate to a code-and-data MRAM architecture as soon as it is available.”

UNISYST is a unified code-and-data MRAM architecture that bridges traditional configuration memory and higher-density persistent storage, extending MRAM into traditional NOR flash applications where superior performance, endurance and reliability are valued. Built as a natural extension of Everspin’s existing PERSYST MRAM platform, UNISYST gives customers a practical, simple migration path from today’s serial MRAM devices to higher-density unified memory without requiring changes to system architecture or software.

Everspin will initially offer the UNISYST family in densities ranging from 128 megabits to 2 gigabits, using a standard xSPI interface operating up to octal SPI at 200MHz. The devices are planned to feature AEC-Q100 Grade 1 qualification and minimum 10-year data retention at extreme temperatures, supporting demanding environments across automotive, aerospace, industrial and edge AI applications.

“As generative AI models move from the cloud to embedded systems, we’re suddenly dealing with assets that are tens or even hundreds of megabytes in size,” said Kwabena W. Agyeman, President and Co-founder of OpenMV. “Storing those models is only part of the challenge — updating them quickly during development and deployment is equally important. High-speed, non-volatile Everspin UNISYST MRAM changes what’s practical for edge AI systems by removing the write bottlenecks associated with traditional flash.”

UNISYST delivers high-bandwidth read and write speeds in a non-volatile memory device, enabling fast boot, rapid updates and predictable performance without the tradeoffs of traditional flash-based designs. By combining high-speed access with persistent storage, UNISYST supports software-defined systems that require frequent reconfiguration while maintaining data integrity across power cycles.

Everspin MRAM has been deployed in mission-critical storage applications for nearly two decades, valued for its endurance and reliability. UNISYST builds on Everspin’s proven MRAM foundation with capabilities designed to support more complex, software-defined systems:

• Code-and-data MRAM architecture designed as a next-generation alternative to other non-volatile memory
• Standard xSPI interface operating up to octal SPI at 200MHz
• Read bandwidth of up to 400 MB/s and write bandwidth of approximately 90 MB/s, over 400 times faster than NOR flash
• Write endurance up to 10 times higher than typical NOR
• AEC-Q100 Grade 1 qualification and minimum 10-year data retention for high-reliability designs

UNISYST is aimed at applications where non-volatile memory must combine high bandwidth, high endurance and predictable behaviour over temperature and time. Target use cases include:

• AI at the edge: Fast AI weight updates, critical storage at the edge, local code-and-data storage for workloads that need fast boot, rapid reconfiguration and non-volatile operation close to the sensor, with the ability to execute in place, removing the need for multiple system memories
• Military and aerospace: Field-programmable gate array (FPGA) configuration and code storage for mission-critical systems, including low-Earth orbit satellites and other platforms that require frequent over-the-air updates
• Automotive: Control, logging and configuration memory in systems that must meet Grade 1 temperature requirements and long-term data retention
• Industrial and casino gaming: High-traffic logging and configuration in environments that demand fast writes, long endurance and persistent storage supporting data logging

The launch of UNISYST represents a platform-level expansion of Everspin’s MRAM portfolio, extending the company’s role from a niche memory supplier to a mainstream memory player serving a multibillion-dollar market. By unifying code storage and data memory, Everspin is addressing the growing demands of software-defined systems that require faster boot times, frequent updates and predictable behaviour over long operating lifetimes.

The post Everspin Launches New Generation of Unified Memory for Embedded Systems appeared first on ELE Times.

Texas Instruments (TI) introduced two new microcontroller (MCU) families with edge artificial intelligence (AI) capabilities, supporting the company’s commitment to enabling edge AI across its entire embedded processing portfolio. The MSPM0G5187 and AM13Ex MCUs integrate TI’s TinyEngine neural processing unit (NPU), a dedicated hardware accelerator for MCUs that optimises deep learning inference operations to reduce latency and improve energy efficiency when processing at the edge.

TI’s embedded processing portfolio is supported by a comprehensive development ecosystem, including the CCStudio integrated development environment (IDE). Its generative AI features allow engineers to use simple language to accelerate code development, system configuration and debugging through industry-standard agents and models paired with TI data. Altogether, TI is accelerating the adoption of edge AI across electronic devices, from real-time monitoring in wearable health monitors and home circuit breakers to physical AI in humanoid robots. These end-to-end innovations are featured in TI’s booth at embedded world 2026, March 10-12, in Nuremberg, Germany.

“TI invented the digital signal processor almost 50 years ago, laying the groundwork for today’s edge AI processing,” said Amichai Ron, senior vice president, Embedded Processing and DLP® Products at TI. “Now TI is leading the next phase of innovation by integrating the TinyEngine NPU across our entire microcontroller portfolio, including general-purpose and high-performance, real-time MCUs. By enabling AI across our software, tools, devices and ecosystem, we are making edge AI accessible and easy to use for every customer and every application.”

“While much of the world has been focused on AI acceleration and NPUs in bigger SoCs, it turns out some of the more interesting and far-reaching applications of AI can be enabled inside smaller chips like microcontrollers,” said Bob O’Donnell, President and Chief Analyst at TECHnalysis Research. “Edge-based applications of AI acceleration can make consumer devices more intelligent and industrial devices more efficient. Plus, if you can combine these chips with software development tools that themselves leverage AI to help build AI features, you bring the power of AI acceleration to a significantly wider audience of engineers and device designers.”

Advanced intelligence at your fingertips

Consumers are always looking for everyday technology to be more intelligent, from fitness wearables to home appliances and electrical systems. However, many engineers believe that AI capabilities are limited to higher-end applications due to high costs, power demands, and coding requirements. TI’s new MSPM0G5187 Arm Cortex-M0+ MSPM0 MCU represents a fundamental shift for embedded designers, who can now bring edge AI to a wide range of simpler, smaller and more cost-effective applications.

With local computation, the TinyEngine NPU executes computations required by neural networks in parallel to the primary CPU running application code. Compared to similar MCUs without an accelerator, this hardware acceleration:

• Minimises the flash memory footprint.
• Lowers latency by up to 90 times per AI inference.
• Reduces energy utilisation by more than 120 times per AI inference.

Such levels of efficiency allow resource-constrained devices – including portable, battery-powered products – to process AI workloads. At under US$1 in 1,000-unit quantities, the MSPM0G5187 MCU reduces system and operating costs by offering an affordable alternative to other MCU or processor architectures.

Real-time control plus AI acceleration for multimotor systems

Motor control applications in appliances, robotics and industrial systems increasingly call for intelligent features such as adaptive control and predictive maintenance, but implementing these capabilities has historically required complex, multi-chip designs. Building on over two decades of motor control leadership through the C2000 real-time MCU portfolio, TI’s new AM13Ex MCUs are the industry’s first to integrate a high-performance Arm Cortex-M33 core, TinyEngine NPU and advanced real-time control architecture into a single chip.

This degree of integration enables designers to implement sophisticated motor control and AI features simultaneously without external components, lowering bill-of-materials costs by up to 30%. Key enhancements include:

• The ability to maintain precise real-time control loops for up to four motors while the TinyEngine NPU runs adaptive control algorithms for load sensing and energy optimisation.
• An integrated trigonometric math accelerator that performs calculations 10 times faster than coordinate rotation digital computer (CORDIC) implementations, delivering more precise, responsive motor-control performance.

Easily train, optimise and deploy AI models

Both MCU families are supported by TI’s CCStudio Edge AI Studio, a free development environment that simplifies model selection, training and deployment across TI’s embedded processing portfolio. This edge AI toolchain gives engineers full flexibility to run AI models on TI MCUs through either hardware or software implementations. Today, there are more than 60 models and application examples available in the tool to help developers start deploying edge AI in any device, with additional tasks and models planned in the future.

The post TI’s microcontroller portfolio and software ecosystem expanded to enable edge AI in every device appeared first on ELE Times.