SynSense launches the Xylo™IMU HDK, empowering ultra-low-power applications for smart wearables and industrial monitoring devices
Xylo™ is the world’s leading programmable, ultra-low-power neuromorphic chip for low-dimensional signal processing, enabling always-on real-time temporal signal processing and recognition applications, at ultra-low-power (<500μW). Coupled with a range of efficient direct sensor interfaces, the Xylo™ family is highly suited for intelligent processing of a variety of sensor signals, including MEMS microphones, temperature sensors, pressure sensors, vibration sensors, acceleration sensors, gyroscopes, EMG, ECG, and more.
Xylo™IMU is the latest addition to the Xylo™ family, joining the Xylo™Audio chip. A fully digital AI chip consisting of a dedicated IMU sensor data acquisition and conversion block (supporting the MC3632 IMU sensor) and a low-power inference core (Xylo™ Core), Xylo™ relies on spiking neural networks(SNNs) for real-time ultra-low-power sensory inference.
Xylo™’s IMU interface reads from an external sensor via an SPI digital bus, converts IMU sensory data to events, and sends the events to the Xylo™ Core for AI inference. The Xylo™ Core is a highly configurable SNN processing core, supporting a wide range of flexible network architectures.
- Contains one Xylo™IMU inference processor
- 3-axis accelerometer IMU on board (MC3632), sampled at 400Hz
- USB3.0 bus for power and interfacing with a host PC
- Rockpool High-level Python API for application development, HDK interfacing and deployment : Quick Start
- On board power monitoring of real-time Xylo™IMU power consumption during inference
- Selectable IMU data source: on–board IMU sensor or streaming data from a host PC
The Xylo™IMU Dev Kit is powered by the SynSense Xylo™IMU chip, which brings the flexibility of highly configurable SNN processing to microwatt energy budgets. Powered by SNN inference algorithms, Xylo™IMU supports a range of battery-powered motion processing applications.
- Personalized recognition and feedback of human posture and behavior, based on motion
Human-computer interaction devices
- Vibration / tactile interaction for smart watches
- Movement-based interaction for wearables
Industrial monitoring devices
- Preventative maintenance by detecting imminent failure, continuously at low power
- Vibration-based analysis of machine operation
- Recognize behavior of pets and livestock from movement
SynSense is a world-leading neuromorphic engineering company. SynSense provides custom-tailored, ultra-low-power silicon design solutions for industrial and consumer machine-learning inference applications. As a “full-stack” neuromorphic engineering company, SynSense delivers complete solutions, including custom IP, hardware, and software configurations to meet specific application needs. SynSense was founded in March 2017 in Zürich Switzerland, based on ground-breaking advances in neuromorphic computing hardware developed at the Institute of Neuroinformatics of the University of Zurich and the ETH Zurich.
Dylan Muir, firstname.lastname@example.org
Nancy Huang, email@example.com, 028-62073881