Sensor fusion IP platform serves consumer electronics
Awareness of context is becoming mainstream feature of consumer electronics. This can take such forms as powering down because an item is in pocket, or powering up because it has been lifted to the ear, or place on the head. But to provide this context reliably usually requires multiple sensor inputs.
The SenslinQ platform eases development by centralizing the workloads that require an intimate understanding of the physical behaviors and anomalies of sensors. It collects data from multiple sensors within a device, including microphones, radars, inertial measurement units (IMUs), environmental sensors, and time-of-flight (ToF) depth sensors, and conducts front-end signal processing such as noise suppression and filtering on this data. SenslinQ then creates “context enablers” such as activity classification, voice and sound detection, and presence and proximity detection. These context enablers can then be fused on-device or otherwise sent wirelessly (Bluetooth, Wi-Fi, NB-IoT) to a local edge computer or the cloud for determining and adapting the device to the environment in which it operates.
The SenslinQ customizable hardware reference design is composed of three pillars connected using standard system interfaces: ARM of RISC-V microcontroller; Ceva DSPs; a wireless connectivity platform Bluetooth, narrow-band IoT or Wi-Fi.
SenslinQ includes a portfolio of ready-to-use software libraries from Ceva and partners, which includes:
Hillcrest Labs’ MotionEngine software packages for sensor fusion and activity classification; ClearVox front-end voice processing, WhisPro speech recognition, and other DSP and AI libraries. The libraries also include active noise cancellation (ANC) software, sound sensing, 3D audio and more.
The SenslinQ platform is accompanied by the SenslinQ framework, a Linux-based Hardware Abstraction Layer (HAL) reference code and APIs for data and control exchange between the multiple processors and the various sensors.
“Contextual awareness, which fuses motion, sound, environmental and other sensor-created data, provides the opportunity to capitalize on our technology excellence,” said Erez Bar-Niv, CTO of Ceva, in a statement. “SenslinQ frees up our customers from having to understand the physical behavior of sensors and the connectivity channels, allowing them to focus on the applications and advanced services possible with a context-aware device.”
SenslinQ will be available to license in 2Q20.
Related links and articles:
News articles:
Ceva goes non-DSP with neural processor
CEVA acquires Hillcrest Labs intelligent sensor technologies
Software provides motion-based gesture control
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
