While major microcontroller (MCU) suppliers like Infineon, Renesas, and STMicroelectronics have been incorporating artificial intelligence (AI) capabilities into their chips to facilitate applications at the edge, two smaller outfits have joined hands to offer a sparse AI MCU. Femtosense integrated its Sparse Processing Unit 001 (SPU-001), a neural processing unit (NPU), with ABOV Semiconductor’s MCU.
The outcome, AI-ADAM-100, is an MCU built on sparse AI technology to enable on-device AI features such as voice-based control in home appliances and other products. It cleans up voice/audio data before it is sent to the cloud, improving reliability and accuracy and reducing the volume of data sent to the cloud. AI-ADAM-100 enables designers to implement voice language interfaces at the edge, even for devices that are not connected to the cloud.
AI-ADAM-100 allows manufacturers to implement voice language interfaces at the edge, even for devices that are not connected to the cloud.
“With sparsity integrated throughout the AI development stack, the AI-ADAM-100 is the first device on the market to fully unlock the advantages of sparse AI,” said Sam Fok, CEO for Femtosense. The sparse AI MCU is being targeted at home appliances as well as small form factor, battery-operated devices such as high-fidelity hearing aids, industrial headsets, and consumer earbuds.
But how does sparse AI work? It reduces the cost of AI inferencing by zeroing out irrelevant portions of an algorithm and then only allocating hardware memory and compute resources to the remaining nonzero, relevant portions of the algorithm. As a result, a system that stores and computes only nonzero weights can deliver up to a 10x improvement in speed, efficiency, and memory footprint.
AI-ADAM-100 enables home appliance manufacturers to implement sophisticated wake-up and control functionality. This allows other system controllers and connectivity modules to drop into sleep mode and consume substantially less power when a user is not interacting with the system. Moreover, ABOV has verified AI-ADAM-100’s voice command recognition performance under multiple noise conditions.
AI-ADAM-100 adds a new flavor to the ongoing marriage of AI and MCUs. It brings language processing and voice cleanup capabilities to an MCU and works with a manufacturer’s own AI models, whether dense or sparse.
Related Content
- Smarter MCUs Keep AI at the Edge
- AI and the IoT merge to form the AIoT
- How to implement AI of Things (AIoT) on MCUs
- Add voice on a microcontroller without having to code
- The MCU pairing streamlines voice control interface design
googletag.cmd.push(function() { googletag.display(‘div-gpt-ad-native’); });
–>
The post Sparse AI MCU facilitates voice processing and cleanup at edge appeared first on EDN.