SmartLight

Enshuo ZOU et Thai Ha NGUYEN

• When the lamp SmartLight ​detects the change of light or the presence of the user, it sends raw data to the remote PC services.
• When the Remote ​PC Service ​received ​data, it turns into states: bright ​or not bright, present or not, and transform these states as Json forma to send to the second remote ​PC services.
• The 2nd Remote PC Service ​receive ​the Json. It stores ​the Json into a local file, and user can check it. In addition ​the service uploads this file on internet, into the doctor'​s ​dropbox. Then the doctor can treating a patient remotely.

arnaud BILLAUDEL

Youtube SmartLight Video

There are two services on the object:

• One service is to capture the raw datas from the distance sensor and tranform the datas to real distance whose unit is cm.
• The other service is the value from light sensor, which presents the level of brightness. These two services are sent with a specified frequency via upnp (to avoid the problem of overrun on the embedded card phidget).

There are two high added value services for our project:

• Save the treated data in a local file to persist data. All these data are in form of json. So that the user can follow his behaviors.
• Upload the data file to a dropbox. This is useful when a doctor need to monitor remotely his patient and intervene when there is a need.

To realize there services, we created a intermediate layer to receive the data sent from the object, and treat these data to get the different states. With the distance data, we can get the states like: the user is present or not. For the brightness data, the states are: bright and not bright. Then we mix the two types of states to get a mixed state, and this is the data that we need to send to the layer of high added value service. Note : (including a figure on the orchestration between services on objects and information systems)

• Embedded

• Remote PC

• 2nd Container on Remote PC ( to simulate the data base and distance service - upload2dropbox)

• NoiseFilter.dll
• ConverToDistance.dll
• NoiseFilter Source code
• ConverToDistance Source code

• DLLs
• source code

• Copy SmartLight.wcc, ConvertToDistance.dll and NoiseFilter.dll on the phidget
• When the SmartLight device is turned on, the embedded service will be started automatically, by modify the script /etc/rc.local (to add mono Container.exe –r ../Beans –l smartLight2.wcc –n SmartLight)

(N.B. As there is a bug from Wcomp for our embedded service, this step is not necessary for us.)

• Copy the CheckBrillantState.dll, CheckPresentState.dll and StateController.dll into the directory Beans of SharpDevlop
• Open Device Spy and copy the functional URL of the object
• Open SharpDevelop and create a new container, then import service.wcc
• Modify the component smartLight_Functional_01's uri with the copied with one (should become green now)
• Bind this container to UPnP
• create another container and import BD.wcc
• Modify the component sl_services_functional_11's uri with the functional uri of the last container.

The collaboration with Reims learned us:

• How to make the nearly commercial product
• How to deal the prototype (we had written a specification file to describe our project)
• To communicate (we discuss on the web site of adhoc frequently to make the decisions and to talk about the changes for the project)

In result, the prototype was made ​​as commended, except that there are two small remarks:

• The high part of the lamp to put in the light sensor is not well done to capture light.
• The lamp was too fragile (lamp was already broken when we received it)

• add the actuators into the object.
• add the motion sensor.
• RFID to distinguish users.

• Service to alert the user when he stays too long in front of his desk.
• Service to alert when the user is in front of the lamp and doctor wants to alert the user.

• Mixer weather information with states to provide data more relevant.
• Combine Calendar service with actuators on objects to advertise user events in the Calendar.
• Combine user's personal calendar with the treated data to get a mixed calendar.

• Create a similar, kind GUI Calendar by adding statements of the weather information and events from the user.
• Transform the JSON data into a excel file and create the table to show the different data.