by Chris Knorowski, September 24, 2018
"SensiML brings real-time event detection to the sensor endpoint with a platform that is accessible to any application developer. Algorithms are created by feeding labeled sets into a cloud-based analytic engine, where SensiML’s AI routines create an optimized device-ready algorithm that balances the desired accuracy with the resource constraints of the target hardware. These algorithms are automatically compiled to optimized machine code that can run in real time on the target embedded platform. SensiML’s platform brings the firmware and data science expertise so that you can go from PoC to production rapidly and with confidence."
QuickLogic's QuickAI™ and EOS™ S3AI platforms are both supported within SensiML Analytics Studio. Unique ultra-low power AI platforms with heterogeneous processing cores, the QuickLogic family of endpoint AI platforms allow SensiML algorithms to be implemented with hardware optimization for FPGA, DSP, and neuromorphic memory.
QuickLogic’s QuickAI platform for endpoint artificial intelligence (AI) applications provides an all-inclusive low power solution and development environment to economically incorporate the benefits of AI in endpoint applications. Full details can be found at the Quicklogic QuickAI product site.
Nordic Thingy:52® is natively supported within SensiML Analytics Studio
The Nordic Thingy:52® is a compact, power-optimized, multi-sensor development kit. It is an easy-to-use development platform, designed to help you build IoT prototypes and demos, without the need to build hardware or write firmware.
Intel® Atom™ (and other x86 compatible processors) are supported within SensiML Analytics Studio
The Intel® Atom™ processor line is built for embedded applications. Power mobile, portable, and small-scale devices of every kind. Choose from processor/chipset combinations and system-on-chip configurations that deliver excellent performance per watt, rich graphics, and I/O integration.
is supported within SensiML Analytics Studio
The Raspberry Pi 3 Model B+ is the latest product in the Raspberry Pi 3 range, boasting a 64-bit quad core processor running at 1.4GHz, dual-band 2.4GHz and 5GHz wireless LAN, Bluetooth 4.2/BLE, faster Ethernet, and PoE capability via a separate PoE HAT.
by Kevin Morris, February 21, 2019
"We’ve written a lot about AI in the cloud, and we’ve discussed data center solutions such as GPUs, high-end FPGAs, and dedicated AI chips such as Intel’s Nervana. For many applications, training and inferencing CNNs and other AI-based systems in cloud data centers is the only way to get the compute power required to crunch the vast data sets and complex models. But, for perhaps an even larger set of applications, cloud-based inference is not practical. We may need latency that cannot be achieved by shipping data upstream to be analyzed. We may not have the ability to maintain a full-time network connection. We may have any of a number of other factors that preclude sending data to a data center and waiting for results to return.
For these applications, the only practical solution is to do the AI inferencing at the edge, right where the data is collected. But AI at the edge brings a host of challenges. The computation required for inferencing of even modest-sized models is enormous. Conventional applications processors cannot come close to the performance required, and their architecture is far from ideal for neural network inferencing tasks. GPUs are power hungry and expensive. Most edge devices are heavily constrained on cost, power, and form factor. Throwing in a real-time latency requirement brings the problem almost to the realm of “unsolvable” with current technology."
by Chris Rogers, June 28, 2018
"While much has been made of AI and ML analytics in the cloud and in edge computing, the overlooked opportunity for analytics contributions at the extreme edge (i.e. the sensor processor itself) remains largely untapped. This presentation - delivered at Sensors Expo 2018 in San Jose, CA - delves into the current and forthcoming advancements in extreme edge processing as part of an overall IoT network solution and the benefits and challenges of a more active role for sensor analytics processing."
by Kevin Morris, May 16, 2018
"Computation is entering an era of unprecedented heterogeneous distribution. The diverse demands of IoT applications require everything from heavy-iron, deep-learning data-crunching to ultra-low-latency snap recognition and judgment. Our IoT devices and systems must be simultaneously aware and responsive to their own local context and able to harness the power of massive compute resources for more global issues. A self-driving vehicle can’t afford to send gobs of raw sensor data upstream to the cloud and then wait for an answer on target identification to return before deciding whether to brake or swerve. It needs to decide immediately whether or not there’s a human in the crosswalk, but it can wait awhile before rendering an AI judgment on whether the pedestrian’s attire was fashionable....
Life would be easy if every engineering team included data scientists who could design the training regimens working hand in hand with hardware experts who could partition the problem between conventional software, programmable hardware, and specialized neural network configuration. But life is not easy. Most projects don’t have access to the wide range of skills and expertise required to optimally engineer an AI endpoint for their IoT design. To make that happen, we need an ecosystem with plug-and-play hardware, software, and AI components and IP that will allow an average engineering project to take advantage of endpoint AI. This month, QuickLogic and several partners are introducing just such an ecosystem.... The collaboration with specialized AI players like General Vision, Nepes, and SensiML creates a robust development platform that should eliminate much of the friction for design teams wanting to take advantage of AI technology at the IoT edge."
by Marcellino Gemelli, October 13, 2017
"Counter intuitively, it is often more efficient to simply leave a sensor on permanently, waiting to identify useful information, e.g. an accelerometer in a step counter application. Our sensor system must intelligently determine which data is worth transferring to the cloud and thereby efficiently utilize the available bandwidth and power. The key is for local on-sensor processing to discard most of this superfluous data autonomously and thus save valuable system driver capacities."
Industrial Machinery Monitoring Demo
Process Overview - Building a Smart Endpoint Sensor
Industrial Wearable Demo
Educational Wearable Demo
Mando-Hella Electronics Rapidly Develops Smart Sensor Applications for New Markets with SensiML Analytics Toolkit
"With the combination of hardware know-how and SensiML Analytics Toolkit for automating the development of smart sensor algorithms, MHE has developed no fewer than five completely unique intelligent products in less than 9 months’ time. The significantly improvement productivity allows MHE to quickly iterate on product features and applications and ensures they have a competitive advantage in expanding their existing business into new markets."