Smart Home Monitoring System (SHMS)
- 0 Collaborators
A home monitoring and notification system that monitors scheduled events at home, and updates users about activities status. Currently, our project scope is to detect and inform users about maintenance activities (Pest control, cleaning etc.) on user request. Our system also does connects and performs activities like switch on, switch off IoT connected devices, based on the user preferences. ...learn more
Project status: Under Development
Mobile, Internet of Things, Artificial Intelligence
Intel Technologies
AI DevCloud / Xeon,
Intel Opt ML/DL Framework,
Movidius NCS
Overview / Usage
In our project, a high definition web camera is set up with Raspberry Pi 3 Model B to monitor doorway. The system triggers and starts monitoring when a person opens door. The device takes several pictures of the entered persons and uses a pre-trained deep learning model deployed on MCS to recognize users and the objects carried by them. Based on the objects and tools carried by the person, the system recognizes the person’s task and checks in its database, if there is any scheduled maintenance task planned at the time. If any scheduled activity is found, the system alerts user about the activity status, such as start and end of the activity. This allows user to know that the activity was performed or being performed, without needing him to check frequently whether the task is formed or not. The Raspberry Pi 3 connects to an IoT bridge, which connects to various smart devices like lights, fans, door locks etc. Based on the maintenance task, the system provides access to certain resources and restricts access to other devices or places. If needed the person can also request access to additional resources to perform the maintenance activity. When a registered resident is detected by our system, it performs configured activities like, switch on and off devices like fans, lights, Microwave etc.
Methodology / Approach
Android Application:
- Users can sign up or sign in to their SHMS system using the mobile app.
- Choose preferred contact method.
- Schedule a monitoring activity.
- View or cancel scheduled activities.
Raspberry Pi 3:
- Process live video stream to recognize persons.
- Central command module for Android app, Camera, Intel AI DevCloud, and IoT devices.
Intel Movidius Compute Stick:
Recognize faces and objects, using a pre-trained model.
Intel AI DevCloud:
Used to train Neural Network model to recognize faces and objects.
Technologies Used
Intel Technologies used:
Intel Movidius Compute Stick (MCS), designed to run Machine Learning models on low powered embedded devices like Raspberry Pi, Beagle Board etc.
Intel Optimized Tensorflow for CPUs, a Tensorflow optimized version for Intel CPUs.
Intel Optimized Keras, a high-level library for Tensorflow.
Intel® AI DevCloud powered by Intel® Xeon® Scalable processors for your machine learning and deep learning training and inference compute needs.
.
Other Technologies used:
Python 3, OpenCV 3, Raspberry Pi 3 Model B, Camera, Android Studio, Android SDK