SensiML, an Oregon-based developer of AI tools for building intelligent IoT endpoints, has announced a free trial version of its Analytics Toolkit and the introduction of its Data Depot sample dataset repository.
 
The company offers an IoT AI software programme that lets developers build intelligent endpoints quickly and easily, a claimed up to five times faster than hand-coded methods. Now with the free trial and Data Depot dataset repository, developers can preview the toolkit and learn how SensiML can accelerate IoT algorithm development prior to purchasing.
 
The Analytics Toolkit is said to deliver benefits for the AI endpoint product, as well as the development process for that product. Data science is built into the toolkit, making it usable for mainstream developers; there is no need for a team of AI experts.
 
The toolkit generates code that runs on MCUs rather than CPUs, enabling the delivery of practical AI on embedded wireless IoT devices. It also increases power efficiency, with built-in automatically generated optimisations for MCUs, DSPs and FPGAs.
 
All phases of design development are supported, including raw signal capture, data insight and labelling, algorithm generation, firmware code generation, and test validation and support. By automating the implementation of AI-based designs, the toolkit reduces the time and resources typically needed, from five engineer years to just four engineer months.
 
Applications for the toolkit include a broad array of time-series sensor intelligence usages across consumer and industrial IoT devices. Examples of such applications include predictive maintenance, safety wearables, smart appliances, fitness wearables and industrial monitoring. Sample application datasets and documentation are now available for exploration and use in the Data Depot dataset repository, which provides a curated set of labelled datasets for a growing list of applications.
 
The toolkit itself is made up of three components: data capture lab, analytics studio and test app. Together, these components provide complete end-to-end development. Multiple hardware evaluation platforms are supported as well, including the QuickLogic Merced and Chilkat platforms, Raspberry Pi, ST Sensor Tile, and Nordic Thingy to speed evaluation and subsequent design development.
 
The free trial version includes:

• Data collection and labelling using sample datasets;
• Automated ML algorithm creation;
• Algorithm performance visualisation;
• Auto-generation of optimised device code;
• Embedded binary executable output with limited number of classification results; and
• Device test and validation application.

“We wanted to make it as easy as possible for potential customers to learn about adding AI to their endpoint IoT applications,” said Chris Rogers, CEO at SensiML. “With this trial version and the SensiML Data Depot, users can quickly explore the design flow for AI-based IoT applications and gain a deeper understanding of how the toolkit can speed their own application development, without requiring large teams of data scientists or engineers.”

SensiML, a subsidiary of QuickLogic, offers software that enables low-power IoT endpoints that implement AI to transform raw sensor data into meaningful insight at the device itself. The toolkit supports Arm Cortex-M class and higher microcontroller cores, Intel x86 instruction set processors, and heterogeneous core QuickLogic SoCs and QuickAI platforms with FPGA optimisations.