The Road to embedded world: EMASS ECS-DoT SoC Powers Real-Time Edge Inference
February 20, 2026
Blog
During its exhibition at embedded world Germany, EMASS (Hall 4, Booth 4-483) will have application specific demonstrations with its ECS-DoT system-on-chip (SoC), a milliwatt-class, on-device AI platform designed for always-on intelligence, ultra-low latency, and reduced power consumption. Demos will highlight how ECS-DoT enables continuous sensing, inference, and decision-making directly on-device, without reliance on the cloud or regular battery replacement.
The ECS-DoT SoC features a 32-bit RISC-V core that delivers up to 30 GOPS at just 5mW. An integrated built in decompression engine supports sub-2-bit AI models (-1.3 bits/weight), maximizing on-chip model capacity. ECS-DoT includes standard 1/0 (12C, 12S, UART, CPI, QSPI, 32 GPIOs) and supports bare-metal or RTOS development with a full SDK and evaluation platform that is compatible with Caffe, PyTorch, TensorFlow Lite and Edgelmpulse.
Ideal applications include wearables, AR glasses, drones, IoT devices, hearables, AI cameras, and smart remotes.
ew Live Demonstrations:
Predictive Maintenance with Semtech LoRaWAN
- This demo explains how ECS-DoT enables real-time equipment health monitoring at the edge as sensor data is processed on-device and transmitted wirelessly via Semtech LoRaWAN, enabling scalable industrial monitoring without cloud dependency.
Audio Security Detection with Semtech LoRaWAN
- The on-device AI consistently listens to critical acoustic events, such as glass break or intrusion signatures in busy environments. ECS-DoT performs real-time detection at sub-milliwatt power levels, extending battery life and low-latency response.
Bone-Conduction Audio & Voice Detection Using
- Demonstration includes a hearable application using an Inertial Measurement Unit (IMU) to facilitate bone-conduction audio sensing. ECS-DoT perceives voice activity and performs keyword spotting by analyzing jaw movement through the IMU, reducing reliance on always-on microphones and supporting private, low-power voice interaction in compact wearable devices.
Vision-Based Smart Meter Retrofit
- Showcased is a vision-based edge AI application with ECS-DoT processing visual data from a camera focused on analog water or gas meters, interpreting dial movement directly on-device. A change in usage patterns, such as rapidly increasing consumption, can be detected locally and transmitted wirelessly, enabling cost-effective smart metering without infrastructure replacement.
For more information, visit https://nanoveu.com/emass/.
Learn about all the latest news and articles about embedded world.
Secure your free ticket, click here.
Exhibiting at embedded world? Download our Marketer's Guide to embedded world.
Chad Cox. Production Editor, Embedded Computing Design, has responsibilities that include handling the news cycle, newsletters, social media, and advertising. Chad graduated from the University of Cincinnati with a B.A. in Cultural and Analytical Literature.
