Why need it?
Internet of Things networks can be controlled in a variety of ways, most commonly through a mobile or web-based application. The goal of this project is to instead use gesture recognition to control IoT networks using MQTT for connectivity and AWS SageMaker for model creation and inference from sensor data.
What is it?
The system uses a wearable instead of the current trend of intense image processing to allow for greater freedom of movement and complexity in gesture motions. As with most wearables, hardware design is a main concern as moving parts need to be secured correctly for reliability. Gestures must also be determined as a gesture within a certain time frame in order to properly utilize machine learning models for prediction. Deterministic design of these cases were developed to mitigate these challenges. The system is ultimately able to control multiple IoT devices through MQTT based off the predictions from AWS SageMaker models invoked by the main wearable.
Internet of Things networks can be controlled in a variety of ways, most commonly through a mobile or web-based application. The goal of this project is to instead use gesture recognition to control IoT networks using MQTT for connectivity and AWS SageMaker for model creation and inference from sensor data.
What is it?
The system uses a wearable instead of the current trend of intense image processing to allow for greater freedom of movement and complexity in gesture motions. As with most wearables, hardware design is a main concern as moving parts need to be secured correctly for reliability. Gestures must also be determined as a gesture within a certain time frame in order to properly utilize machine learning models for prediction. Deterministic design of these cases were developed to mitigate these challenges. The system is ultimately able to control multiple IoT devices through MQTT based off the predictions from AWS SageMaker models invoked by the main wearable.