quicklogic-logo
Home > Products > QuickAI – Old

QuickAI HW/SW Platform

Artificial Intelligence (AI) and Cognitive Sensing
at the Endpoint

The QuickAI™ platform provides an all-inclusive
low power solution and development environment
to economically incorporate the benefits of Artificial
Intelligence (AI) in endpoint applications.


A Complete Artificial Intelligence (AI) System Platform with

Sensor processing

eFPGA for feature extraction

Neurons for AI computing

Data analytics SW for data training
and model/classifier building

Scaling across bigger AI
systems with thousands of
endpoint devices

No need for in-house
expertise of data analytics,
DSP processing, app coding

Delivers an AI module solution
that can be deployed at the
endpoint devices with
different connectivity, sensor

System Architecture


*/
?>

Addressing the Challenges of Endpoint Applications

  • Companies developing endpoint devices often lack the resources to work effectively with the complexities of cloud-based AI
  • Endpoint applications often benefit from local AI resources that can react quickly, lower endpoint power consumption and lower life-cycle costs
  • Endpoint design teams often lack the Data Science and Firmware Engineering resources needed to develop AI models
  • The diversity and uniqueness of endpoint use cases drive the need to develop application specific algorithms and models
  • Once deployed, endpoint companies must have a plan to manage the distributed endpoints, leverage the information they collect and often update AI models remotely

QuickAI Platform

QuickLogic– The EOS™ S3 voice and sensor processing platform make it the ideal host SoC for the NM500. It can implement the proprietary interface for the NM500, sample the IoT sensor data, and extract features using its embedded FPGA technology.

SensiML – The analytics toolkit quickly and easily trains the data, builds the model and classifier and programs the EOS S3 for endpoint AI.

Nepes – The NM500 implements the NeuroMem technology in an energy efficient, small form factor component. It can be trained in the field to recognize patterns in real time, and multiple devices can be chained to provide any number of neurons. The Knowledge Studio software tool can be used to configure and train the neurons in the NM500 device.

Merced Hardware Development Kit (HDK)

The Merced HDK is an evaluation platform for QuickLogic’s EOS S3AI SoC. The HDK comes with time-series (continuous) sensors, including:

  • Accelerometer/gyroscope/magnetometer motion sensors
  • Two digital microphones (PDMs)

The Merced HDK can store data locally using a μSD card or transmit sensor data via a built in BLE module.

Ready to Use – Out of the Box

The Merced HDK is an Arduino form-factor compatible module. This makes it an ideal hardware platform to expand features without developing new hardware. The Merced HDK includes a demo application that makes it ready to use out of the box. The device comes with a boot-from-flash feature to enable users to evaluate QuickAI™ applications. It also links directly with the SensiML Data Analytics Toolkit in order to collect data, develop AI models, algorithms and new classifiers. Updating the Merced HDK is possible via an on-board USB-2-UART connection. This enables the creation of demos and prototypes without any additional development time or resources.

Ordering Information

To order, refer to part number QAI-EVALKIT-AA-1.0 and contact sales.

Example Applications

Industrial Predictive Maintenance

  • Unique model doesn’t scale across similar motor differences in mounting or loading
  • Endpoint AI decreases system bandwidth, latency, power
Algorithm Development: SensiML Toolkit for Time Series
  • Data collection, segmenting, labeling
  • Sensor input: motion, audio, pressure, temp/humidity, other time series data
  • Feature extraction
  • Model building
FPGA Features
  • Sensor Data Creation → Feature Extraction → Feature Vector
  • Hardware accelerator (FFT & MFCC)
  • NM500 hard neuron interface
FFE Enabled Features
  • Event trigger (segmentation)
  • Feature extraction for simpler features
  • Ultra-low power AON function

Structural Health Monitoring

  • Damage detection
  • Structural Integrity reporting

Algorithm Development: SensiML Toolkit for Time Series

  • Data collection, segmenting, labeling
  • Sensor input: motion, audio
  • Feature extraction
  • Model building

FPGA Features

  • Sensor Data Creation → Feature Extraction → Feature Vector
  • Hardware accelerator (FFT)
  • NM500 hard neuron Interface

FFE Enabled Features

  • Event trigger (segmentation)
  • Feature extraction for simpler features
  • Vibration (high precision accel) analysis at 200Hz ODR

Features: Vision Inspection for Fruit/Vegetable Harvesting

* Available Q2 2019

Features:

  • Identify ripe fruit/ vegetable from unripe fruit/vegetable
  • QuickAI HDK Platform (with camera interface) can be used for evaluation and/or deployment
  • HDK using UART interface to drive servo motors

FPGA Features:

  • Sensor Data -> Feature Extraction -> Vector
  • Interface to neurons and camera