Renesas & Syntiant Launch Low‑Power Voice‑Controlled Vision AI for Embedded IoT Systems

Renesas Electronics Corporation and Syntiant Corp. have partnered to create a low‑power, voice‑controlled multimodal AI platform that delivers contactless image processing for embedded vision‑AI IoT and edge applications.

The integration of Renesas’ RZ/V series vision‑AI microprocessor unit (MPU) with Syntiant’s ultra‑low‑power NDP120 neural decision processor delivers advanced voice and image processing for use cases ranging from self‑checkout kiosks and security cameras to video‑conferencing systems and smart appliances such as robotic vacuums.

The joint solution features an always‑on mode that allows rapid voice‑triggered wake‑up from standby, enabling object and facial recognition, as well as other vision‑based tasks essential for security cameras and similar systems. User‑defined voice cues initiate activation and operation, while the vision AI tracks operator behaviour and can alert operators when suspicious activity is detected.

By separating voice recognition onto a dedicated, power‑efficient chip, the architecture reduces standby power consumption and accelerates development. Developers can program voice and vision components independently, simplifying integration.

The Renesas RZ/V series MPU incorporates the DRP‑AI (dynamically reconfigurable processor‑AI) accelerator, combining high‑precision inference with power efficiency that Renesas claims is among the best in the industry. This performance eliminates the need for heat sinks or cooling fans, lowering BOM cost and enabling vision‑AI integration across a wide range of embedded applications. DRP‑AI is built from AI‑MAC and DRP cores, which efficiently handle convolutional and fully‑connected layers by optimizing data flow with internal switches. The DRP can dynamically reconfigure to process complex tasks such as image pre‑processing and AI model pooling layers quickly. A DRP‑AI translator automatically maps AI model operations to the appropriate hardware, allowing developers to use DRP‑AI without detailed hardware knowledge. Multiple executables can reside in external memory, enabling dynamic switching between AI models at runtime, and the translator can be updated to support new models without hardware changes.

The Syntiant NDP120 chip incorporates sophisticated AI capabilities that support high‑precision, hands‑free voice functions, including speaker recognition, keyword detection, multiple wake words, and local command recognition. Paired with the Syntiant Core 2 neural‑network inference engine, the NDP120 can run multiple applications simultaneously while consuming as little as 1 mW of battery power.

Speaking for Renesas, Hiroto Nitta, Senior Vice President and Head of SoC Business in the IoT and Infrastructure unit, said, “We anticipate that demand for multimodal systems that use multiple streams of input information – both image and voice – will increase moving forward as a way to improve both ease of use and safety.” He added, “Through the collaboration between Renesas, a leader in low‑power image AI technology, and Syntiant, a leader in voice AI technology, we will accelerate the adoption of low‑power, ultra‑small smart voice AI technology in embedded systems and deliver new combined solutions to customers globally.”

Syntiant CEO Kurt Busch noted, “Voice‑based user interfaces will make it possible for customers to deliver new user experiences that bring the next generation of innovative ideas from concept to reality. We’ve already shipped more than 15 million of our deep‑learning NDPs globally to enable always‑on voice in a wide variety of consumer and industrial IoT applications. Our collaboration with Renesas delivers a powerful, low‑power voice and image solution that is certain to accelerate traction among a global customer base in a variety of devices and use cases.”

The new voice‑controlled multimodal AI solution draws on multiple compatible devices from Renesas’ broader portfolio, providing customers with an elevated prototyping platform for faster time to market and reduced risk. As part of Renesas’ “winning combinations,” which pair compelling analog, power, and embedded processing products, the solution helps customers accelerate design and launch new products more quickly.

A reference design for the multimodal AI solution, including circuit diagrams and BOM lists, is now available.