In recent years, machine learning is a broad field that has seen tremendous progress. It is based on the principle that a computer can autonomously improve its performance on a given task by learning from data – sometimes even beyond the capabilities of humans. Machine learning’s applications include but not limited to image classification, object detection, audio scene, speech recognition, weather or stock market forecasting, and anomaly detection.
Edge AI refers to the use of artificial intelligence in the form of machine learning algorithms running directly on edge devices instead of the cloud. Bringing AI from the cloud to the edge, can not only reduce the requirement for latency, bandwidth, and power consumption (which means cost down) but also increase data security and reliability of the whole IoT system.
As machine learning develops, many exciting possibilities will now extend to edge devices as well. Thanks to the TinyML, a subset of machine learning that employs optimization techniques to reduce the computational space and power required by machine learning models, Edge AI is fitting artificial intelligence applications into smaller and less powerful computers, even ultra-low-power microcontrollers, such as our Wio Terminal.
Wio Terminal is an ATSAMD51-based microcontroller with wireless connectivity supported by Realtek RTL8720DN. It is equipped with a 2.4” LCD Screen, onboard IMU(LIS3DHTR), Microphone, Buzzer, microSD card slot, Light sensor, and Infrared Emitter(IR 940nm).
Now, Wio Terminal is officially supported by Edge Impulse, a leading development platform that enables developers to create the next generation of intelligent edge device solutions with embedded Machine Learning.
If you’re new or you would like to learn more about the Wio Terminal, check out “Resource Roundup for Wio Terminal: Tutorials, Reviews, and Projects from Community” where we cover different documentations, reviews, community projects and also courses for the Wio Terminal!
Earlier this year, Seeed organised a contest for attendees to get a free trial of using the Wio Terminal by submitting their own creations. The contest aims to build project tutorials for Wio Terminal and other embedded hardware using TinyML/Edge Impulse/Azure/IoT Central/IoT Plug-and-Play. We also included projects from the contest winners. A huge thank you to all the attendees who have participated in the competition and submitted their projects. To the winners, congratulations on winning a coupon refund for your project! We will contact you soon with details on your coupon.
Seeed will organise more events in the future for everyone. Do stay tuned!
Without further ado, let’s explore how the community solves real problems using Machine learning with Wio Terminal at the very edge.
Applications of Wio Terminal x Edge Impulse (Updated on August 26th, 2021)
Francesco Azzola builds this KNN classifier with Wio Terminal. This tutorial describes how to implement a Wio Terminal KNN classifier. In more detail, it covers how to use a KNN classifier to classify objects using colors. To implement this Wio Terminal Machine Learning example, we will use a color sensor (TCS3200). Moreover, we will use the built-in LCD display to show the result. Check out the detailed tutorial here.
Salman Faris applied Edge Impulse and a Wio Terminal to build up this immersive game project. When you jump, the T-Rex jumps! Having gaming fun and burning calories at the same time, what we could ask more!
?Project tutorial: https://lnkd.in/gTKxDDm
Fun audio classification project from Francesco Azzola! This tutorial describes how to build a machine learning model to detect chirping birds using Wio Terminal and Edge Impulse. To simplify the steps to build the machine learning, Francesco uses Edge Impulse with a Colab script. And he uses Environmental Sound Classification 50. This is a classification problem where we have to classify a sound and determine if it belongs to the chirping birds class or to the noise class.
This is a great project. Learn more about it here.
With the Wio Terminal equipped MIC Sensor, Selvakumar created a shower timer with machine learning. He documented the steps he took to collecting the data for the machine learning model, and also the challenges he faced while doing so.
Learn more about this ingenious Wio Terminal Shower Timer here.
This project aims to provide better rehabilitation for people differently-abled. A Neurosensory Buzz will be used as the main interface for patients, which will be receiving BLE signals from several ESP32 microcontrollers. The Wio Terminal is used here for its BLE capabilities to send the sensor data.
This is a great project backed up by a great cause. Learn more about it on hackster.io.
Air Purifier Control Poc by Mithun
He incorporates #TinyML in a TensorFlow Lite Model to detect intruders in the dark with a low-resolution thermal camera! For Data collection for training, Naveen has used Wio Terminal to collect data. The 3 buttons on the Wio Terminal were used to label the 3 classes (Person, Object, and Background). The captured data is saved to the files on an inbuilt micro SD card on the Wio Terminal.
Benjamin’s AI Nose on Make: Magazine
Benjamin Cabé applied Wio Terminal with artificial intelligence and tinyML into an AI Nose that detects surrounding smells. It can be used for a wide variety of applications, from helping folks suffering from anosmia to spot the smell of burning food or spoiled milk, to monitor the cleanliness of office buildings, etc.
While Wio Terminal makes for a fine IoT development kit, Shawn tested it to run embedded machine learning applications (TinyML). Specifically, he tested its ability to handle speech recognition (or more specifically, keyword spotting).
He trained a convolutional neural network to identify 4 keywords: forward, left, right, stop. These words were chosen to demonstrate what you might find on something like a robot. However, any words could have been chosen (within some limits). It seems that the neural network gets a little too big or inaccurate after about 5 different keywords.
A digital stethoscope that auscultates and detects abnormalities in the respiratory system using tinyML at the edge.
This project tutorial shows how you can build a machine learning-based handwriting recognition device with the Wio Terminal, Edge Impulse, and Arduino. Follow this detailed guide to learn how a single time of flight sensor can allow you to recognize handwriting gestures and translate them to text!
Tiny Machine Learning Capabilities to Wearable IoT Devices for Boxing Technique Management Anthony Joseph
Anthony Joseph, CTO of My House Geek, has recently presented a micro-sized, wearable device designed to improve your boxing skills. With the goal of reinforcing good technique in mind, Joseph designed a device that is capable of detecting when one hand is in a blocking position, while the other hand is throwing a punch.
Typically, a weather station project would involve connecting the sensor to our microcontroller platform and reading out this data in real-time, either on an LCD display or over an internet connection to an online database.
Today, I want to share the smart version of this weather station project that I created by adding the capability for our Wio Terminal to predict the current weather conditions. With an onboard TensorFlow Lite model, the Wio Terminal becomes able to use real-time temperature and relative humidity data for making half-hourly weather predictions!
“In this wiki, we will introduce how to use Wio Terminal with Edge Impulse to simply deploy a machine learning project. Wio Terminal with the Grove systems can be very powerful, which brings hundreds of sensor data in for analysis and to possible evaluate different scenarios!”
“Let’s quickly train and deploy a simple neural network for classifying rock-paper-scissors gestures with just a single light sensor.”
This project aims to show that TinyML can perform gesture recognition functions with only a single pixel from a single light sensor. The TinyML Course is a video series by Dmitry Maslov teaching about TinyML on the Wio Terminal. The various episodes teach several projects, with a detailed explanation about how everything works. You can watch the videos here.
“In this tutorial, I’ll teach you how to train and deploy an audio scene classifier with Wio Terminal and Edge Impulse. For more details and video tutorial, watch the corresponding video!”
“In this project, we will create a people counting system by using Wio Terminal, an ordinary Ultrasonic ranger and special Deep Learning sauce to top it off and make it work. We will also utilize Microsoft Azure IoT Central service to store the room occupancy data in the cloud and visualize it on PC.”