My idea for the Intel Edison is to use it as a Cognitive Server for IoT Sensor Networks.

Data will remain just data unless it is used to come to a certain conclusion. Nowadays, with low-cost IoT sensor nodes generating huge amounts of data, we rely

on off-site servers and corresponding cloud platforms for analytics. What if you could have a personal, dedicated computation device to make sense out of the data  being generated by the nodes which are of relevance only to you, say, your house?

Enter the Intel Edison. I plan to leverage the Intel Edison's wealth of communication avenues and sheer computing horsepower to introduce the element of artificial intelligence to the conventional architecture of wireless sensor networks.

With data generated from household devices, such as when a person switches on the lights or at what temperature does he switch off the fan, the Edison will learn the preferences of an individual and optimize the power usage of such resources by only switching them on when they are needed. It will also show the usage of said resources on a dedicated web server accessible via a browser like Chrome on devices on the same network.

The procedure is fairly simple.

1. Data is generated by household sensors, including when the resources are in use and when they aren't.
2. This data is relayed to the Edison via BLE or WiFi (TCP stack)
3. The received data is stored in an SQL table locally on the Edison
4. A separate machine learning script, running continually on the Edison regularly queries the database and maps the temporal and sensor value dependency of these values (eg : At what time does the user switch on the lights? What is the ambient light value when he does?)
5. Using this trend, the Edison trains itself to use the resources only when the values of environmental parameters are within the constraints.
6. Using relays, it will control the resources as needed.
7. The values of stored sensor data from all data sources as well as the state of the resources will be displayed on a web page hosted on the Edison.

My idea for the Intel Edison is to use it as a Cognitive Server for IoT Sensor Networks.

Data will remain just data unless it is used to come to a certain conclusion. Nowadays, with low-cost IoT sensor nodes generating huge amounts of data, we rely

on off-site servers and corresponding cloud platforms for analytics. What if you could have a personal, dedicated computation device to make sense out of the data  being generated by the nodes which are of relevance only to you, say, your house?

Enter the Intel Edison. I plan to leverage the Intel Edison's wealth of communication avenues and sheer computing horsepower to introduce the element of artificial intelligence to the conventional architecture of wireless sensor networks.

With data generated from household devices, such as when a person switches on the lights or at what temperature does he switch off the fan, the Edison will learn the preferences of an individual and optimize the power usage of such resources by only switching them on when they are needed. It will also show the usage of said resources on a dedicated web server accessible via a browser like Chrome on devices on the same network.

The procedure is fairly simple.

1. Data is generated by household sensors, including when the resources are in use and when they aren't.
2. This data is relayed to the Edison via BLE or WiFi (TCP stack)
3. The received data is stored in an SQL table locally on the Edison
4. A separate machine learning script, running continually on the Edison regularly queries the database and maps the temporal and sensor value dependency of these values (eg : At what time does the user switch on the lights? What is the ambient light value when he does?)
5. Using this trend, the Edison trains itself to use the resources only when the values of environmental parameters are within the constraints.
6. Using relays, it will control the resources as needed.
7. The values of stored sensor data from all data sources as well as the state of the resources will be displayed on a web page hosted on the Edison.

Would love to build an Intel Edison garage monitoring and alert system as showcased on Instructables!