This report covers processors and the surrounding ecosystem for artificial intelligence and machine learning at the edge, focusing on embedded systems ranging from TinyML to those capable of hundreds of TOPS.
The report delves into the evolution of edge AI from a niche technology to a mainstream powerhouse catalyzing change across autonomous vehicles, IoT, healthcare, and more. From real-time decision-making in autonomous vehicles to immediate patient monitoring in healthcare, edge AI is setting new standards for safety, efficiency, and performance.
Ceva’s NeuPro-Nano licensable neural processing unit (NPU) targets processors that run TinyML workloads, offering up to 200 billion operations per second (GOPS) for power-constrained edge IoT devices.
The wireless device market is evolving rapidly, driven by advances in multi-protocol connectivity technologies like Wi-Fi, Bluetooth®, 802.15.4, and UWB. These innovations are enabling groundbreaking applications across consumer, enterprise, and industrial sectors—enhancing performance, reducing power consumption, and delivering smarter, faster devices with unprecedented features like positioning and sensing.
By leveraging Edge AI processors, a software ISP based on neural network technology can process and optimize video in real-time, surpassing human vision capabilities in these challenging conditions.
Join Ceva and Visionary.ai experts to learn about the power of Edge AI for computer vision.
Ceva-Waves™ Links - Family of innovative multi-protocol wireless platform IPs integrating Wi-Fi, Bluetooth, IEEE 802.15.4 (for Thread/ Matter and Zigbee) and Ultra-Wideband (UWB), to accelerate the development of connected Smart Edge SoCs.
Ceva-RealSpace® Elevate – Complete Spatial Audio Software Solution for Android, PC, and Gaming systems
Ceva-RealSpace Elevate is a full spatial audio solution with immersive spatial audio rendering on a mobile phone, PC, or gaming system with or without precise head tracking on an accompanying headset.
This means a great experience isn’t limited by bandwidth or tied to a particular content provider or codec.
Ceva-RealSpace® Personal – Complete Spatial Audio Software Solution for Personal Audio
Ceva-RealSpace Personal is a fully embedded solution with precise head tracking and immersive spatial audio rendering on a TWS earbud, headphone, or XR headset.
This means a great experience isn’t tied to a particular device ecosystem, content provider, or codec.
Ceva-RealSpace® Ambient – Complete Spatial Audio Software Solution for Speakers
Ceva-RealSpace Ambient is a fully embedded solution bringing immersive spatial audio to Bluetooth speakers, smart speakers, soundbars, and cars. A wider, more expansive sound field means a great listening experience is no longer just for headphones.
Ceva is a leading provider of Bluetooth semiconductor IP and software for integration into SoC. We design a complete portfolio, spanning from the PHY all the way up the host software protocol stack and profiles, addressing both Bluetooth low energy and Bluetooth dual mode flavors. The Ceva Bluetooth IP is used in many devices such as smartphones, tablets, headsets, earbuds, hearing aid devices, speakers, medical devices, smart home devices, industrial IoT, and many others including applications such as asset tracking and indoor navigation.
Ceva has updated its licensable NeuPro-M design to better handle Transformers—the neural networks underpinning ChatGPT and Dall-E AI software and now finding application at the edge in computer vision. Architectural changes to Ceva-NeuPro-M improve its power efficiency and can increase its performance sevenfold on models that can take advantage of the new features. Critical among these is the addition of BF16 and FP8 support to more Ceva-NeuPro-M function units and improved handling of sparse data.
Driven by the recent uptake by major smartphone brands, UWB wireless technology is experiencing rapid growth in market interest and applications. In the automotive sector, UWB has become the standard for Digital Keys in the premium segment. Installing UWB anchor points in modern vehicles provides a cost-effective platform for advanced in-cabin radar, enabling features like Child Presence Detection (CPD).
UWB Radar for automotive demo video showcases the functionality of Ceva's UWB sensor, which serves as both a child presence detector and a gesture recognition system.
It demonstrates how the same UWB IP, used for Digital Key 3.0 anchors, enables these features
Any automated machine needs to have a strong understanding of where it is in order to complete its task. A robot arm needs to be able to weld one section of a car to another, a 3D printer needs to print fine details with precision layer by layer, and consumer robots need to navigate an everchanging space to clean.
Today, we want to focus on the last case – consumer robots moving along relatively flat surfaces to clean. These robots determine location using a technique called SLAM (Simultaneous Localization and Mapping). Data from the wheel encoders and inertial measurement unit (IMU) - and optionally optical flow/camera/LIDAR - all contribute to determining the movement of the robot. This is the localization part. Mapping is what keeps the robot on an efficient path that avoids obstacles and avoids repeating areas.
Within this set of robots, we’re going to focus on precise localization using a price conscious set of sensors and the process of optimizing performance. Data from the wheel encoder, optical flow sensor, and IMU can be combined for odometry (the change in location over time). When these calculations are computed without the aid of wireless beacons or LIDAR to get an exact location, this is also known as Robot Dead Reckoning. This paper will go over in detail the comparison and analysis of dead reckoning algorithms. Whereas a previous white paper focused on the basics
of robotics sensors and touched on testing, this one will focus in depth on the testing process. We will discuss how our data was collected, what we are comparing against, and how we analyzed the comparison data along with the presenting the data itself.
Fueled by the recent adoption by leading smartphone brands, the UWB wireless technology is enjoying an explosive growth in market interest and applications. In automotive, UWB is already the de-facto choice for Digital Keys in the premium segment. The deployment of UWB anchor points in modern cars creates a cost-effective platform for offering advanced in-cabin radar, enabling features such as Child Presence Detection (CPD).
As consumers demand more immersive and seamless experiences, Spatial Audio has become a key technology in a broad range of applications. Whether it’s pinpointing an approaching enemy in a game, feeling the roar of the crowd at a stadium or concert, or reducing fatigue from video calls, the benefits of having a more realistic experience are fueling the growth of the wireless audio market. Ceva-RealSpace® is a complete spatial audio solution, combining proven best-in-class 3D rendering with precise head tracking to bring total immersion to any audio content in an easy-to-integrate package. Get out of your head and into the action with RealSpace®
"Ceva’s XC20 vector-DSP architecture implements a new simultaneous multithreading technique, allocating the resources of a single vector unit to two threads instruction by instruction.
Vector-unit utilization increases, boosting the licensable DSP’s area efficiency. Combined with other updates, Ceva reports the first XC20-based core, the dual-thread XC22, is more efficient than the single-thread XC4500 and the dual-thread XC16. "
Read the full report about our solution for intense 5G and 5G-Advanced workloads.
Ceva-XC22 is the first member of fifth generation Ceva-XC20 vector processor architecture from Ceva, designed to address 5G-Advanced compute challenges, scalable vector multi-thread framework, targeted for advanced 5G RAN and high end UE terminals, from Base stations and Massive MIMO Radios to next generation CPE and FWA terminals.
Ceva’s PentaG2-Lite platform IP is enabling 5G connectivity in power-constraint IoT and consumer devices. 5G for IoT is the promising next phase for utilizing the real potential of 5G networks, opening up a multitude of new vertical use cases.
Already supplying two generations of 5G basebands for user equipment, Ceva now offers one for the radio access network (RAN).
Like PentaG2 for smartphones and low-data-rate cellular devices, PentaG-RAN integrates DSP cores with configurable accelerators.
In addition, through the Intrinsix business it acquired last year, Ceva helps cellular-infrastructure customers design their base-station chips.
Baseband Platform IP for 5G RAN ASICs
PentaG-RAN is industry's most comprehensive and integrated baseband platform IP targeting base stations, radio devices, Open-RAN and Massive MIMO devices.
The Ceva-NeuPro Studio is a comprehensive graph compiler that simplifies the development and deployment of deep learning systems for mass-market embedded devices. CDNN incorporates a broad range of network optimizations, advanced quantization algorithms, data flow management and fully-optimized CNN, RNN, and other type networks into a holistic solution that enables cloud-trained AI models to be deployed on edge devices for inference processing, using low compute and memory resources. CDNN enables heterogeneous computing and is flexible to split a network between multiple compute engines such as SensPro or NeuPro processors and custom AI Engines, to ensure superior performance with minimal power consumption. In this video, we are showing how the CDNN Graph Compiler and GUI enable users to quickly configure the CDNN tool and easily analyze their neural networks performance on any of Ceva’s AI processors. The example we are using is inferencing of the ssd_mobilenet network on the SP500 DSP, both natively (DSP only), and also with a hardware accelerator connected via CDNN-Invite API for higher performance.
Ceva's low power Ultra-Wide Band or UWB modem IP, that brings secured ranging and Angle of Arrival measurements for various applications. The Ceva-Waves UWB platform IP consists of a power-optimized hardware PHY along with a flexible, low latency hardware and software MAC layer and
Building on its vast experience with Bluetooth and Audio DSP, Ceva developed Bluebud, a turnkey hardware and software integrated IP platform for TWS earbuds, smartwatches, wearables and other IoT devices that require audio, voice, and sensor processing.
Ceva-NeuPro-M NPU architecture offers scalable performance of 4 to 1200 TOPS with highest power efficiency for Generative and Classic Edge AI workloads.
5G RedCap is an exciting new 3GPP feature, soon to be introduced in Rel. 17 of the standard, targeting reduced capability use cases for industrial, wearables, and IoT in general. We will discuss the new standard, the market potential, and Ceva’s new PentaG2-Lite comprehensive baseband modem solution.
PentaG2™ – Next Generation 5G NR IP Platform. Get the introduction to the 5G RedCap standard, learn about 5G RedCap use cases and market potential, and Ceva PentaG2-Lite architecture
Our FSP201 provides superior planar heading and 6 axis IMU performance ideal for high-volume, cost-conscious consumer robots, XR, spatial audio, and other motion-based devices. The FSP201 is an application-specific standard product (ASSP) integrating Hillcrest’s high-performance sensor hub software into a low-power 32-bit ARM Cortex M23 MCU.
Join Ceva and VisiSonics experts to learn about:
What is Spatial/3D Audio
How does Spatial/3D Audio work
What are the technical challenges to implementing an effective Spatial/3D Audio system
Why is head tracking so important to Spatial/3D Audio
How can the technical challenges be addressed to provide a robust, complete solution
For its second-generation 5G baseband, Ceva is going down, not up. The original PentaG design already scales to 10Gbps, the maximum speed we expect 5G to reach.
So for PentaG2, the company focused on reducing energy per bit, limiting heat and stretching battery life for its smartphone customers. The new design is also suitable for use beyond phones, in products such as laptop PCs, tablets, cellular hot spots, and automotive telematics. It can scale down to a “lite” configuration that fits into low-data-rate devices such as asset trackers, smart meters, and remote sensors.
Watch Sanjay Gangal of EDACafe interviews Gil Abraham from Ceva as they dive into the innovations powering Ceva’s NeuPro-M, an heterogeneous and secure AI/ML processor architecture for smart edge devices, which can provide a scalable 8 to 1200 TOPS
Ceva-NeuPro-M deep-learning accelerator (DLA) offers 16x faster inferencing than the previous generation thanks to its new architecture.
The licensable core showcases Winograd-convolution and unstructured-sparsity engines, which boost performance and reduce power consumption.
These accelerators also keep multiply-accumulate (MAC) units busy, increasing utilization.
Ceva-PentaG2 Lite is indestry's first comprehensive and integrative baseband IP Platform meeting challenging tight power envelope targets for next generation cellular IoT devices, supporting LTE Cat.1 and future 5G RedCap devices
Ceva-PentaG2 Max is Indestriy's comprehensive IP Platform capable of meeting the extreme performance, low latency, and strict power budget for 3GPP 5G UE broadband eMBB and URLLC devices
Imagine how will our lives look in a 5G world!
• We will have fast cellular connectivity, everywhere!
• At home, 5G removes the limits on the number of devices
• Allowing for countless wireless connections
• Changing what we can do as we commute to and from work, or travel around the world
• Creating new possibilities at the office, the factory, the hospital— wherever we work
• Connecting it all together
Bringing the 5G dream to life: with Ceva PentaG for UE applications and Ceva-XC16 for infrastructure applications.
Create smart devices that understand the world around us, and seamlessly communicate.
Ceva-Waves™ UWB is a low power ultra-wideband (UWB) MAC and PHY platform IP based on 802.15.4 HRP and the FiRa Consortium requirements. It delivers secure, centimerter-level accuracy and robust location information through Time-of-Flight (ToF) ranging and Angle-of-Arrival (AoA) processing.
Robotic systems are a complex beast, requiring the combination of carefully picked mechanical, electrical, and programmed parts.
Before a line of code is written, you have to pore over spec sheets to determine which mechanical components stand a chance of achieving your end goal. Every robot also requires sensors to understand the world around them, then rely on mapping algorithms to move where we want them to, not just where the code tells them (old programming joke).
After you have something that proves your concept, you need to test more to fine tune the idea over time. Testing is another sophisticated and arduous process, to measure and analyze what the robot is seeing relative to what you see and want.
If you're tired of reading this already, that's understandable. Even with this high-level summary of the process, there are still so many combinations of places where choices need to be made and things can go wrong.
Join Ceva and PixArt experts to learn about:
The importance of great dead reckoning for autonomous robots
Obstacles to accurate and robust dead reckoning
How dynamic calibration and sensor fusion can dramatically improve accuracy and robustness
Verifying performance in the real world
More about the PAA5101 sensor and its capabilities
Bluebud: DSP-enabled Bluetooth and audio IP platform for TWS earbuds and other wireless audio devices
Bluebud™ is a turnkey, feature-rich hardware and software platform to streamline the development of SoCs for True Wireless Stereo (TWS) earbuds, wireless speakers, and wearables. Self contained, fully integrated with RivieraWaves Bluetooth baseband controller, Ceva-BX1 Digital Signal Controller and a comprehensive software suite.
Read more...
Join Ceva and experts to learn about:
State of true wireless stereo (TWS)
Factors affecting the user experience with TWS
Challenges in current approaches
How sensor fusion helps solve context awareness
What we see in the future of TWS
Join Ceva and VisiSonics experts to learn about:
Multichannel Audio and Ambisonics
3D Audio Applications and markets
HRTF and Head-tracking for 3D Audio
3D Audio technology has been around for a while, but its recent advent, presented by tech giants has opened an interest to push its boundaries with innovative use cases.
3D audio will see rapid adoption in applications such as gaming, video and music playback by generating an immersive real life audio environment through which the listener can distinguish the distance of the sound.
While VR and AR headsets were the first natural candidates to deploy 3D Audio as they already have high-quality head tracking data for graphics rendering - same data is now used for positioning audio.
3D audio can truly immerse a user in a new space. Sounds from a real-world source reach each ear at a different time, and your brain interprets these differences in time to determine directionality. Recreating these audio cues for a truly 3D experience requires accurate, low-latency head tracking
Ceva's sound enhancement algortithms engineer, Ofer Shahen Tov, demonstrate the company's new software package including the ClearVox™ package for Front-end voice pickup with Noise reduction and with Acoustic echo cancellation. Then he demonstrates the WhisPro™ Speech recognition SDKit.
Applications Include Smart IoT Devices, Drones, and Small Robots
By Mike Demler (April 20, 2020)
...................................................................................................................
Ceva-SensPro gives designers a versatile and highly configurable set of licensable intellectual-property (IP) cores for scalar and vector signal processing as well as neural-network acceleration. The company designed Ceva-SensPro, which it introduced at the recent Linley Spring Processor Conference, as a powerful sensor hub that combines technology from its Ceva-BX2 DSP, Ceva-XM6 computer-vision engine, and Ceva-NeuPro deep-learning accelerators (DLAs). Although the new offering’s neural-network capacity is less than Ceva-NeuPro’s, its AI/sensor-fusion capabilities are a good fit for a wide range of products, from IoT devices and smartphones to ADASs, augmented/virtual-reality (AR/VR) headsets, drones, and robots.
Ceva-SensPro™ is a family of vector DSP cores designed for computer vision algorithms, AI foundation operations, and AI pre & post-processing capabilities in an NPU companion configuration, with excellent balance of Performance/Watt and Performance/area.
The XC16 DSP is the first product to implement Ceva’s fourth-generation XC architecture, which includes a new dynamically configurable multicore/multithread feature.
The design comprises two eight-way VLIW cores—scalar processors (SPs), in the company’s parlance—along with four vector compute units (VCUs). During run time, it allows software to configure the SPs and VCUs as two symmetric dual-vector (DV) processors or allocate all four VCUs to one SP to form a quad-vector (QV) processor, with the second SP handling control operations.
Enabling next generation 5G RAN - world’s strongest vector DSP for cellular baseband applications.
Ceva-XC16 is sixth generation vector processor IP from Ceva, designed to bring multi-gigabit high-end communication and cellular capabilities to base-stations, small-cells, gateways, access points and CPEs.
A modern, high performance Audio DSP, optimized for far-field noise reduction and Artificial Intelligence speech recognition.
Ceva-BX2 designed for intensive Audio applications such as, Voice assistants with multi-microphone beamforming and noise reduction, Speech recognition engines (ASR) for voice trigger and commands, Object based audio rendering and 3D Audio processing, Natural Language Processing (NLP) at the edge, Sound sensing and analytics using neural networks...
A modern Audio DSP, designed for battery operated, high-performance, audio and voice applications.
Ceva-BX1 is an ultra-low-power Audio DSP, designed for Bluetooth earbuds and headsets using audio over BLE/BTDM, Voice controlled IoT with always-on speech recognition, Sound sensing in wearables, Audio codecs and high quality Audio post processing.
As the IoT industry continues to grow and more devices become connected, the amount of data being collected and analyzed grows with it. Smart phones, robot vacuums, wearables, drones, video game controllers, and even automobiles use sensors to give them context in the world they operate. These applications in particular revolve around moving components. And a key part to their proper understanding is utilizing the data they collect effectively...
Robotics is a field that is growing rapidly. In particular, consumer ground-roving robots are becoming mainstream now that they commonly incorporate more intelligent navigation to operate more autonomously. Inertial measurement units (IMUs) are sensors that are essential for achieving precise navigation. IMUs can be used as a primary sensor for navigation or as a complementary sensor that helps VSLAM systems achieve robust performance. But making use of IMUs requires deep understanding of their idiosyncrasies and modeling to assess their impact. In this webinar, engineers will learn about the challenges when working with IMUs, how IMUs are applied in different robotics applications, and what is necessary to test IMU-based robots to achieve great performance.
Join Ceva's experts to learn about:
Consumer Robot Vacuum Industry Overview
Common home robot navigation paradigms
Sensor idiosyncrasies and the importance of qualifying sensors
How to maintain accuracy over time, temperature, and environment
How to create robust algorithms and verify performance
Target Audience
Design, system, and product engineers looking to enhance their understanding of IMUs, sensor fusion, and how they relate to SLAM. Robot OEMs interested in navigation for their ground-roving robots.
Speakers
Charles Pao
Former Sr. Marketing Specialist, Sensor fusion BU, Ceva
Steve Scheirey
Sr. Director, Software, Sensor Fusion BU, Ceva
Ceva provides embedded solutions for low-power Bluetooth and a wide range of audio capabilities to build your advanced, ultra-low power wireless earbuds and hearing aid devices. This video will show you the advantages of the new BT LE Audio that incorporates Multi-Stream Audio, and how you can apply it to your true wireless earbuds. Create smart devices that understand the world around us and seamlessly communicate.
Introduction - Making Sense of the Sensors
Everywhere Around Us Embedding sensor technology into products has been a very strong trend in recent years.
From smartphones, fitness bands, gaming controllers, smartwatches, head-mounted displays the list of consumer devices with sensors is rapidly expanding. Traditionally we have seen sensors used to track location and movement (e.g. GPS, accelerometers, gyroscopes, and magnetometers), but this too is quickly shifting to a raft of new advanced sensors such as biomedical, audio and visual.
The versatile and broad reaching MotionEngine software system is the product of over 15 years of research and development. Our deep understanding of sensors is at the core of why MotionEngine delivers best-inclass performance when used with consumer grade sensors. This is achieved through innovations like dynamic sensor calibration, high accuracy 6-axis and 9-axis sensor fusion, precise heading and application specific features such as cursor control, gesture recognition, activity tracking, context awareness, and AR/VR stabilization.
The Ceva-Toolbox™ provides a software development framework for Ceva DSP cores, and includes a complete set of development, debug, and optimization tools for each of Ceva’s DSP product families. All of these tools support the customer through the entire embedded application development flow.
IP Vendors Roll Out Multicore Deep-Learning Accelerators
By Mike Demler (October 7, 2019)
The rapid adoption of machine learning is driving IP vendors to compete by scaling up performance in each new generation of licensable deep-learning accelerators (DLAs).
Last year, DLAs integrating up to 4,096 multiply-accumulate (MAC) units per core set the standard for licensable inference engines.
But advanced driver assistance systems (ADASs) and other high-performance edge devices demand even greater performance, so with their latest products, Ceva and Synopsys aim to shatter that mark.
Read More...
Samsung Electronics is a key household name when it comes to all things electronic. They also offer a full suite of products for consumers and businesses developed around the concept of Connected Living. From phones, computers, and displays to Smart Home refrigerators, wall ovens, and washing machines all are connected through the Samsung SmartThings Hub.
Adapting your products to customer needs is critical, and robotic vacuum cleaners have made serious headway in recent years when it comes to improving the user experience. But challenges still exist: robots that get stuck on chair legs, lost in corners, or take inefficient routes, all of which lengthen cleaning times and drive up battery consumption (driving down user satisfaction).
Motion sensors play a key role in improving robotic vacuum cleaner performance. Read our Ebook, “Unlocking the Robotic Cleaner of Tomorrow” to learn:
The top challenges that persist in the robotic vacuum industry
A few key trends shaping the market
How to choose the right sensor set (to mag, or not to mag...?)
How sensor fusion ensures peak performance
In the not-so-distant future, handheld motion control will play a crucial role in how people interact with their smart devices. These technologies are already instrumental in gaming, smart TVs, and VR applications, and are being experimented with in other use cases ranging from consumer electronics to elder care and more. Creating a more natural user interface using handheld motion control can drive product adoption, user satisfaction and customer loyalty. It can unlock a whole new set of features that will impact how consumers use the technology they love.
Enabling SDR based, multigigabit modems for 5G, LTE and Wi-Fi infrastructure
Ceva-XC12 is fifth generation vector processor IP from Ceva, designed to bring multi-gigabit high-end communication and cellular capabilities to base-stations, gateways, access points and CPE.
Read more...
The World’s Most Advanced Communication DSP for Multi-Gigabit Connectivity
Industry’s first DSP capable of meeting the extreme performance requirements critical to the success of 5G, Gigabit LTE, MU-MIMO Wi-Fi and other multi-gigabit modems
Designed leveraging our unique expertise in DSP and from working with world-leading wireless vendors
Can be custom configured and scaled to address a wide range of applications
Cloud RAN, macro cells, small cells, advanced and centralized access points, smartphones and other terminals
Read more...
A common method for creating a 6 degree of freedom (DoF) device is to track a single marker with a camera to measure linear position, and use an IMU to measure angular position (orientation). This is often used for AR and VR controllers and head tracking. One key challenge is aligning the coordinate frames of the camera and the IMU, and keeping them aligned over time. This paper presents a method of frame alignment that doesn’t require user intervention, and properly handles integration errors from the gyroscope as well as the very large errors encountered when performing double integration on the linear acceleration of the IMU’s accelerometer to compute linear position. The paper also compares two options for solving for the alignment, either a recursive least-squares (RLS) approach or as a solution to Wahba’s problem.
High level programmable, modern processor architecture for a broad range of signal processing and control workloads
Ceva-BX2 is a multipurpose hybrid DSP and Controller designed for the inherent low power requirements of DSP kernels with high-level programming and compact code size requirements of a large control code base.
High level programmable, modern processor architecture for a broad range of signal processing and control workloads
Ceva-BX1 is a multipurpose hybrid DSP and Controller designed for the inherent low power requirements of DSP kernels with high-level programming and compact code size requirements of a large control code base.
By Mike Demler (February 4, 2019).
Ceva’s new BX DSP cores offer upgrades to the first-generation X-series models, which employ a hybrid architec-ture combining multicore scalar CPUs for control functions with a highly configurable five-way SIMD/VLIW DSP en-gine for signal processing. The new intellectual property (IP) core tightens security by using a TrustZone-like memory subsystem to isolate secure applications from nonsecure ones. Architectural enhancements double the fixed-point DSP performance compared with the previous generation, and floating-point DSP operations receive up to an 8x boost....
This product brief provides a high level description of the FSP200 Development Kit, the FSP200 BMI055 Reference Module, and the FSP200 Serial/USB Adapter.
A recent market study from the Bluetooth SIG shows that shipments of Bluetooth audio devices are growing steadily, and set to exceed 1.2 billion units annually by 2022. This market is dominated by headsets shipments, with wireless earbuds as the hottest and fastest growing segment today. To facilitate wireless earbuds, true wireless stereo is a mandatory technology. This webinar presents the various solutions available on the market for true wireless stereo, with particular focus on the Ceva innovative solution.
Introduction – Sensor Hubs and Beyond
While sensor fusion based around inertial motion sensors has been around for some time, recent developments in mobile applications have cast a spotlight on the algorithms, processors and infrastructure for sensor fusion on consumer and industrial devices.
Summary
The headset market has seen major advances in recent years. Headsets are becoming truly wireless, smarter, smaller, and provide a heightened user experience. Enriched headset features include integrated voice assistant, advanced audio analytics, and above all enhanced sound experience powered by Active Noise Control (ANC). Flip the switch and you are immersed in Silence. Turn on your favorite tune or podcast and transplant yourself from a loud subway to your favorite concert hall.
While the early research on ANC dates back to the beginning of the century, it is only recently that new advancements in silicon and audio processing have enabled ANC technology to be implemented by software in battery-powered wireless headsets.
Ceva and Silentium have partnered to drive these trends with DSP and software ANC technologies which can be easily and quickly configured and ported to SoC designs. In this webinar, we depict a holistic wireless sound solution based on Ceva’s audio DSP and Silentium’s software-based ANC application targeting cost and power sensitive wireless headsets and earbuds.
Join Ceva and Silentium experts to learn about:
Wireless headset market trends
Ultra-low-power sound DSP platform
Active Noise Control digital technology
System design considerations for ANC enabled headsets
Target Audience
Audio system engineers, SoC designers and marketing managers targeting advanced headset design, and headset OEMs looking to ANC-enable their products
Speakers
Youval Nachum
Former Senior Product Marketing Manager, Ceva, Inc.
Igor Kofman
Director Product Management & Business Development, North America, Silentium Ltd.
Dragonfly NB2 Implements Cat-NB2 Specification in 3GPP Release 14
By Tom R. Halfhill (July 9, 2018)
New ways of enabling machines to wirelessly talk with each other are propagating as fast as blockchain startups. Ceva is staying abreast by introducing Dragonfly NB2, its second-generation platform for Narrowband IoT (NB-IoT) communications on cellular networks. Whereas the first-generation Dragonfly NB1 supported the Cat-NB1 specifica-tion in 3GPP Release 13, the new product supports the en-hanced Cat-NB2 spec in Release 14. In fact, it appears to be the world’s first implementation of this protocol.
The IEEE’s next-generation 802.11ax Wi-Fi specifications are more than a year from ratification, but (as usual) several wireless-chip vendors have already announced pre-standard products. To enable them, DSP-intellectual-property (DSP-IP) specialist Ceva is offering 802.11ax additions to its RivieraWaves Wi-Fi family. Its new RW-AX-series comprises three models targeting a range of products that spans IoT/wearables, mobile devices, access points, and gateways. The company plans to release the RW-AX IP for general licensing by the end of this quarter.
Ceva demonstrates the 1st High Quality True Stereo Audio streaming over Bluetooth low energy (BLE).
It is powered by Ceva RivieraWaves main features:
Ceva Rivierawaves BLE5 baseband controller IP core and software protocol stack
RISC-V based fully integrated platform
Ceva-TeakLite-4 low power DSP used for high quality stereo AAC-ELD 24KHz decoding
Ideal for headsets and hearables with 18uW/MHz @ 28nm power consumption
Compliant with Apple MFi LEA
The Linley Group.
Neural-Network Engine Optimizes for Variable Channel Conditions.
By Mike Demler (March 12, 2018).
Ceva’s new PentaG DSP comprises a set of configurable processing blocks that allow designers to build digital basebands implementing the recently ratified 3GPP 5G New Radio (5G-NR) specification...
Introduction
Scope
This document describes the narrowband IoT (NB-IoT) radio access technology, and focuses mainly on the primary synchronization process of the NB-IoT in the downlink, which is activated when the user equipment (UE) attempts to receive data from the base station (BS) when it leaves its dormant state and wakes up.
When the UE wakes up, it must first synchronize with the time, frequency, and base station parameters (for example, cell ID or radio network temporary identifier (RNTI)) to be able to receive data. In this document, we mainly focus on the UE's initial cell search, in which the range of the carrier frequency uncertainty is very high. This process is one of the most power-consuming processes in NB-IoT because it requires the UE to continuously activate an RX radio frequency (RF) chain and to constantly process input samples until the synchronization stage is complete...
Enabling high-performance sound processing applications requiring multi-microphone speech processing and high-quality audio playback.
Ceva-X2 is a unified DSP/controller, designed for speech recognition, deep learning sound analytics, and high-end audio processing...
NeuPro AI processors offer a considerable step-up in performance, ranging from 2 Tera Ops Per Second (TOPS) for the entry-level processor and 12.5 TOPS for the most advanced configuration and are optimized for scalable power consumption, performance, and area (PPA) requirement.
Introduction
Challenges toward 5G NR Release 15
Compared to previous LTE releases, 5G NR introduces a significant increase in the throughput, latency, and number of users, as well as in the connection density.
Because 5G NR is targeted at supporting the entire frequency band, from low band in UHF up to 100 GHz, a number of frequency-related signal-propagation challenges must be addressed...
Liran Bar, Director of Product Marketing at Ceva, demonstrates the company's latest embedded vision technologies and products at the January 2018 Consumer Electronics Show. Specifically, Bar demonstrates computer and vision technology in the form of the latest iCatchtek V50 chip, powered by Ceva's vision processor and running real-time multi-exposure HDR (high dynamic range) functions.
Liran Bar, Director of Product Marketing at Ceva, demonstrates the company's latest embedded vision technologies and products at the January 2018 Consumer Electronics Show. Specifically, Bar demonstrates computer and vision technology in the form of the Wunder360 C1, a commercially available 360 degree device that incorporate Rockchip's RV1108 chip powered by the Ceva-XM4 vision processor. Real-time 3K video and 4K still image stitching is performed on the device itself, specifically running on the Ceva-XM4. The output stream is saved on the internal SD memory card. Unlike with other solutions, there is no need to do additional processing on a PC or smartphone.
Liran Bar, Director of Product Marketing at Ceva, demonstrates the company's latest embedded vision technologies and products at the January 2018 Consumer Electronics Show. Specifically, Bar demonstrates Brodman17's real-time face detection and tracking neural network running on a FPGA-based Ceva-XM4 vision processor at 60 MHz. A production ASIC-based implementation will run above 800 MHz. The neural network porting work to the Ceva-XM4 was performed by Brodman17's engineers, leveraging Ceva's award winning CDNN SW framework, and was completed in a very short time.
Liran Bar, Director of Product Marketing at CEVA, demonstrates the company's latest embedded vision technologies and products at the January 2018 Consumer Electronics Show. Specifically, Bar demonstrates state-of-the-art deep learning technology in the form of a real-time Tiny Yolo neural network running on Inuitive's latest NU4000 SoC powered by the CEVA-XM4 vision processor. The Tiny Yolo porting work to the CEVA-XM4 was performed by Inuitive's engineers, leveraging CEVA’s award winning CDNN software framework, and was completed in a very short time.
Liran Bar, Director of Product Marketing at Ceva, demonstrates the company's latest embedded vision technologies and products at the January 2018 Consumer Electronics Show. Specifically, Bar demonstrates image enhancement technology as exemplified by commercially available smartphones that incorporate a standalone pre-ISP chip to enhance the overall image quality. The fundamental OEM motivation is to reduce the overall smartphone cost by using a mid-range application processor instead of a high-end one. OEMs supplement the mid-range application processor by incorporating Rockchip's RK1608 standalone pre-ISP chip, in order to optimize image quality. In this demonstration, Ceva is showing the ASUS Zenfone 3 Zoom smartphone's high image quality under 5 LUX low-light environmental conditions, comparing it to other available smartphones in the market.
Jeff VanWashenova, Director of Automotive Segment Marketing at Ceva, and Young-Jun Yoo, Director of Strategic Marketing at NextChip, demonstrate the companies latest embedded vision technologies and products at the January 2018 Consumer Electronics Show. Specifically, VanWashenova and Yoo demonstrate NextChip’s APACHE4 SoC, an entry-level ADAS solution that provides vehicle, pedestrian, moving object , and lane detection based on NextChip’s efficient detection engines. The APACHE4 also leverages Ceva's XM4 imaging and computer vision processor IP to add additional capability, flexibility and customer differentiation.
As the automotive market experiences accelerated growth and rapid adoption of vision applications such as Camera Monitoring Systems, Smart Rear Cameras, and Driver Monitoring Systems, there is a need for solutions that are both efficient and cost effective to address these applications in high volumes. In addition, these solutions must also allow for Tier-1s to both differentiate and meet the growing demands in performance from today’s OEMs.
NextChip’s APACHE4 is a vision-based pre-processor SoC targeting next-generation ADAS systems that uses a dedicated sub-system of image processing accelerators and optimized software. The APACHE4 incorporates dedicated detection engines that include pedestrian detection, vehicle detection, lane detection and moving object detection and have incorporated Ceva’s programmable vision platform into the APACHE4 alongside its differentiated image processing accelerators to enable advanced and affordable ADAS applications.
This application note provides customers that are designing with Ceva’s BNO070 Sensor Hub device design details that will allow that design to migrate to the future BNO085 device.
While the BNO085 and BNO070 use the same physical device, there are software differences that provide new features and make subtle changes to the I/O provisioning of the device.
Ceva-XM および NeuPro AI プロセッサー用ニューラルネットワーク推論コンパイラグラフ
Ceva ディープニューラルネットワーク通称 Ceva-NeuPro Studio は、マスマーケット用組込みデバイス向けディープラーニングシステムの開発と導入をサポートする包括的なコンパイラグラフです。CDNN コンパイラグラフには、Ceva オプティマイザーと Ceva-NeuPro Studio ランタイムフレームワークが含まれます。
This webinar covers the current state and future possibilities of voice interfaces. It surveys the technologies that have enabled current proliferation of voice interfaces but also takes a critical look at the faults and drawbacks of current implementations. Finally, it explores the existing, emerging and future technologies that will eventually generate a revolution in the way we interact with machines.
Turning yesterday’s sci-fi into today’s reality, voice interfaces are gaining traction but still haven’t reached their peak. Enabling technologies are all around and can offer smarter and more efficient applications with more natural and intuitive interfaces. An always-on voice interface with human-like intelligence, capable of understanding intonations and inflections, responding to context and anticipating our needs and desires may be much closer than most people think.
Join Ceva experts to learn about:
Natural Human-machine interfaces (HMI) of the future
Far-field voice pickup and its enabling technologies
Under the hood of smart speakers
Human-like virtual assistants
Target Audience
Audio system engineers targeting natural human-machine interfaces, and marketing managers looking to voice-enable their smart home, mobile, and automotive products.
Speakers
Eran Belaish
Product Marketing Manager, Audio/Voice, Ceva
Robert Schrager
Director of sales and marketing, Alango Technologies
High-Performance DSP Targets Wireless Infrastructure,
By David Kanter (April 3, 2017)
The move to 5G and the arrival of myriad new networked
devices will increase demand for bandwidth. To
meet this demand, infrastructure will employ more antennas
and higher bit rates, which all require greater computational
throughput. At the same time, latency requirements
will tighten to enable heterogeneous networks, such
as a combination of licensed and unlicensed spectrum as
well as responsive communication for automation.
We are the world's leading licensor of signal processing IP, partnering with semiconductor companies and OEMs worldwide to create power-efficient, intelligent and connected devices. We target a range of end markets with our cores and platforms, including mobile, consumer, automotive, industrial and IoT. More than 8 billion Ceva-powered products have shipped to date, including 1 out of every 3 handsets sold worldwide.
By 2020, ABI Research predicts that there will be more than 30 billion connected devices with Wi-Fi and/or Bluetooth.
While most people first think about Bluetooth when talking about connectivity in IoT devices, it is important to note that Wi-Fi is expected to also be massively used, with Wi-Fi connections forecast to overtake Bluetooth!
Smart Home devices will represent more than half of the Wi-Fi enabled IoT devices, and 20% of wearables will integrate Wi-Fi.
To meet these goals, low power, low cost, reliable and secured Wi-Fi solutions are required.
The 3GPP recently completed the standardization of a new LTE-based narrowband technology to support the emerging needs of cellular-IOT devices. Known as Cat-NB1 (formerly NB-IoT), this new low data rate category will enable the development of ultra-low power, ultra-low cost solutions that can leverage the existing LTE infrastructure for a new wave of connected devices. However, in order to efficiently leverage the Cat-NB1 standard, a new approach to cellular SoC design is required, where low data rate transfer in very large time intervals is the goal. As such, solution designers might step out of their “comfort zone” and reevaluate existing modem solutions.
3GPP is currently actively working on 5G New Radio (NR). IMT2020 is defining advanced technology for dramatically increasing network capacity and coverage, in order to answer the ever-increasing demand for higher data rate and traffic with much reduced end to end network latency. Such technology includes wider RF channels in licensed and unlicensed bands up to mmWaves, aggregation of large number of components carriers of various widths and very short TTI.
The need for flexible 5G base station implementations requires novel SDR SoC architectures and SW/HW partitioning strategies optimized to solve the daunting challenges of Beamforming with Massive MIMO systems while maintaining very low latency and blazing fast data rate within a low power budget.
Join Ceva experts to learn about:
Cellular 5G market overview
Introduction to 5G
5G challenges
Impact on SDR architecture and SW/HW partitioning
Ceva's solution for 5G baseband
Target Audience
Communication and system engineers targeting 5G segment
Speakers
Emmanuel Gresset
Business Development Director, Wireless and Wireline Communications, Ceva
Tomer Yablonka
Senior Communication System Architect
The automotive market is seeing accelerated growth and rapid adoption of vision applications that will lead the way to autonomous vehicles. With the complexity of these systems, Tier-1 suppliers, OEMs, and the entire automotive industry are utilizing artificial intelligence and deep learning algorithms to identify objects, determine free space for vehicles and plan the vehicle movement. As companies explore these deep learning algorithms and shift from R&D labs to the realization and deployment of low power embedded solutions, it is important to have a sound foundation in the form of an efficient HW and SW platform that is optimized for CNN workloads and other deep learning approaches.
Join Ceva and AdasWorks experts to hear about:
Challenges of ADAS and vision based autonomous driving
Ceva’s 5th generation deep learning embedded platform based on the Ceva-XM6 vision processor
Implementing low power machine vision solutions using the Ceva Deep Neural Network (CDNN) toolkit
Free space detection utilizing AdasWorks drive 2.0 SW implemented on Ceva's imaging and vision platform
Target Audience
Computer vision engineers, deep learning application designers, project managers, marketing experts and others interested in embedded vision, machine learning, and autonomous driving.
Speakers
Liran Bar
Former Director of Product Marketing, Imaging & Vision, Ceva
Jeff VanWashenova
Former Director of Automotive Segment Marketing, Ceva
Arpad Takacs
Outreach Scientist, AdasWorks
By 2020, ABI Research predicts that there will be more than 45 billion connected devices worldwide. More than half of these devices will incorporate multiple standards in the same device, such as Wi-Fi, 802.15.4g, GNSS and cellular, including the upcoming ultra- low data rate LTE MTC Cat-M. Some of these devices, such as wearables, will only require a battery life of a few days, but others such as asset trackers will demand a battery life of 5-10 years. So how do system designers address these significantly different challenges, while also meeting the stringent cost and power metrics?
HIGH-PERFORMANCE Vector DSP for Soft ware Defined Radio Infrastructure Applications.
Up to 4x performance improvement over existing 4G wireless infrastructure DSP.
DSP Core Handles LTE-Advanced and 802.15.4 Protocols.
By Mike Demler (August 1, 2016).
Ceva’s new X2 DSP core performs physical-layer (PHY) control tasks in complex wireless modems. Intro-duced at the recent Linley Mobile & Wearables Conference, the design targets LTE-Advanced carrier aggregation, which requires digital basebands to simultaneously control PHYs for multiple communications channels and radio access technologies (RATs), such as a combination of small-cell and macrocell connections. The 3GPP’s LTE Release 13 and later require modem designers to support the new protocols, but they must also retain 3G and even 2G functions for fall-back in legacy networks or for voice when voice-over-LTE (VoLTE) is unavailable.
Comprehensive set of Bluetooth Low Energy and Bluetooth dual mode Silicon Intellectual Properties and software for consumer, automotive and IoT devices
Compliant with latest Bluetooth generation, with full features including 2Mbps, Long Range, Direction Finding (AoA/AoD), Isochronous Channels / LE Audio / Auracast, Packet Advertising with Response, and Channel Sounding
Includes baseband controller / link layer, modem, and RF
Full software protocol stack with comprehensive list of profiles for Medical/HealthCare, Sport & Fitness , Audio, Mesh, etc.
Add on for 802.15.4 (for ZigBee, Thread, Matter)
Jeff VanWashenova – Director of Automotive Market Segment.
Automotive Trends Challenges in Deep Learning
Efficient implementation of Neural Networks
Alexnet Example
The evolution of 4G wireless baseband indicates there are two evolving technologies that are currently competing for the leading position; namely LTE and WiMAX. WiMAX has emerged as the technology of choice for computing devices and M2M, as well as for fixed wireless connections in underdeveloped areas. It is promoted by the WiMAX forum, led by Intel and other companies. LTE is evolving as the “GSM” direction towards cellular broadband and is promoted by the 3rd Generation Partnership Project (3GPP) and adopted by baseband OEMs and operators including Qualcomm, ST-Ericsson, Verizon, Vodafone and others.
It is unclear which standard will eventually prevail, and it is highly possible that both standards will coexist and serve different use-cases in different geographies. To ensure that both of these standards are supported in 4G modems, a flexible solution that would allow for specific implementations under both technology roadmaps is required.
ANALYSTS’ CHOICE WINNERS FOR 2015.
Recognizing the Best Chips and Technology of the Past Year.
By The Linley Group (January 18, 2016).
To recognize the top semiconductor offerings of the year, The Linley Group presents its 2015 Analysts’ Choice Awards. These awards span several categories:
embedded processors, mobile processors, PC and server processors, processor-IP (intellectual property) cores, and related technology. We have presented them in Microprocessor Report for several years.
New Ceva-XM4 DSP Core Adds FPUs and 32/64-Bit Vectorsץ
By Tom R. Halfhill (April 27, 2015)ץ
As more machines gain the gift of sight, engineers are rediscovering a principle long known to biologists: vision is equally a sensory perception and a cerebral function. The eyes see, but the brain interprets and reacts. Thus, process-ing power is as vital to computer vision as image capture.
Streamlining the design of cellular-enabled low data rate IoT devices.
Ceva-X1 is a Lightweight Multi-purpose, Multi-mode single-core processor that addresses the severe size, power and cost limitations demanded for deploying the latest LTE Cat-M1 and Cat-NB1 IoT standards.
MOST SIGNIFICANT BITS.
Ceva TEAKLITE-4 ILLUMINATES ROADMAP
J. Scott Gardner (April 16, 2012).
Hot on the heels of introducing its XC4000 family (see MPR 3/5/12, “Ceva Exposes DSP Six Pack”), Ceva has lifted the curtain on another new family of DSP cores. Speaking at the Linley Tech Mobile Conference, the company announced its new TeakLite-4 architecture and provided details about four new cores it plans to launch this year. As the successor to its popular TeakLite-III family, the software-compatible architecture is designed to scale across a broad range of product categories. TeakLite-4 DSPs will primarily target audio and voice processing in everything from low-power mobile devices to high-end home entertainment systems.
By Linley Gwennap (May 5, 2014).
Today’s wearable devices get by with smartphone processors or microcontrollers, but more-successful de-vices will require processors custom-designed for this application. To jumpstart that design process, Ceva an-nounced a collection of useful IP at last week’s Linley Tech Mobile Conference. This integrated platform supports always-on functions and low-power connectivity, two capabilities that are critical in wearable devices.
The platform includes a low-power TeakLite-4 DSP to handle voice functions and sensor management, as Figure 1 shows. It can even run a simple user interface, eliminating the need for a host CPU. The second hardware piece is Ceva’s Bluetooth core, which includes the MAC and baseband for Bluetooth 4.1. Most importantly, the platform comprises a set of pretested software modules that implement these functions, enabling customers to focus on their own value-add software.
This is the first Wi-Fi 802.11ac access point running at speed on FPGA, based on Ceva RivieraWaves Wi-Fi MAC and modem IPs. It can handle up to 10 video streams in parallel.
According to ABI Research, the number of connected devices will reach 48 billion by 2021, a third of which will be Bluetooth wireless technology enabled. The Bluetooth Special Interest Group (SIG) has recently released the highly-anticipated Bluetooth 5, which extends the performance and the scope of Bluetooth low energy. New features include a doubling of speed (from 1Mbps to 2Mbps), as well as a 4X range increase, thus enabling smart home applications. What does it take to design BLE products that are low power and low cost, but also reliable?
Filename: 1000-3924_2.3.8.425_MotionStudio2.zip
Version: 2.3.8.425
FREESPACE® MOTIONSTUDIO 2
MotionStudio 2 is purpose-built for configuring and evaluating advanced Ceva sensor fusion devices, including the BNO08X, FSM30X, and FSP20X. It allows developers to easily perform operations on connected devices. Freespace MotionStudio 2 enables:
Sensor Control and Logging
Visualization of motion with the Virtual Object display
Device firmware management, reorientation, calibration and other critical operations
NOTE: If you are interested in a GUI for the FSM-9, use MotionStudio v3.9.3 instead
Requirements
OS:
Windows 7, Windows 10
Processor:
Intel Core i7 1.8GHz and faster
RAM:
8GB minimum
Enabling Deep Learning and Artificial Intelligence in Low-Power Embedded Devices
Ceva-XM6 is the 5th generation imaging and computer vision processor IP from Ceva, designed to bring deep learning and artificial intelligence capabilities to low-power embedded systems.
Filename: Freespace-MotionStudio-3.9.3.zip
Filesize: 7 MB
Version: 3.9.3
Freespace® MotionStudio is an intuitive toolset that provides an easy-to-use graphical user interface to allow developers to accelerate the evaluation of Freespace motion technologies on the FSM-9. It provides tools for experimenting with FSM-9 devices connected to a PC running Microsoft Windows XP, Vista, and 7. Freespace MotionStudio includes:
Motion Chart: Plot and log motion data.
Mouse Display: Display mouse motion, buttons and scroll wheel data in real time.
Virtual Object: Display a virtual object that moves with the physical Freespace device.
Fitts’ Law Test: Test pointer performance.
Reference Kit Configuration Tools: Additional tools including firmware upgrade, button/LED configuration, and device reorientation.
Designing the cursor for a pointing-based television must consider factors that differ from the implementation of a cursor for a computer screen. Several factors affect cursor design. This document provides an overview of these design factors and recommends specific approaches to cursor implementation that could be made to address them.
As demand for Internet content on television increases, service providers and consumer electronics manufacturers need alternative technologies to the traditional up/down/left/right remote control in order to delight consumers with new interactive experiences on TV. This paper describes why pointing is the right solution for controlling the TV experience, and compares two technologies that enable pointing remote controls: touchpad and motion control. A series of studies compare these two methods of pointing for two common use-cases: navigation and casual gaming. Results show that pointing by motion is superior to pointing with touchpad for both use-cases.
