Qualcomm Gimbal: Creating Context-Aware Mobile Applications
Qualcomm Gimbal is a context awareness platform for Android and iOS enabling mobile developers to add context sensitive functionality to their applications. Basically, with Gimbal, an application will be able to push rich media notifications to a device based on user’s preferences, habits, location and time.
Gimbal operates with several concepts:
- Places are locations on the world map defined by latitude and longitude. There are two types of places: Public and Private. Public places are created and managed by the application developer and are associated with all application users or a group of them. Private places are created by the application on the user’s device and are not shared with anybody else.
- Geofences are digital boundaries created around Places, either as a circle with a certain radius or a polygon. The application is informed when the user crosses a geofence, and can use this event to push a notification to the user. Geofence events are generated even when the application is not running, since they are triggered by a background service.
- Interest Sensing. Gimbal determines what the user is interested in by collecting and analyzing data from all hardware sensors, browsing history, applications installed, and other sources. Data is analyzed at night when the device is plugged-in based on a set of rules.
- Image Recognition (IR). Gimbal is capable of recognizing some objects when the device’s camera is pointed at it.
- Communication. The application developer may decide what rich media notification to send to specific user groups when a certain geofence event is triggered.
Gimbal forces applications to asks users if they want and which features they want to be enabled. Also, certain data, such as the browsing history, is not made available to applications, only the analysis result.
Gimbal has an online Manager used to generate API keys for applications, to create geofences, to setup notifications, and to analyze user activity.
Gimbal comes with an associated SDK and API, including a sample application and API documentation and usage examples. The SDK is currently free, but it will cost $0.08/user/month for more than 5,000 users for using each of this services: geofence, interest sensing or IR.
The SDK allows app developers to make sense of vast amount of data that may be available to them leading to a large number of possible uses, as Robert Scoble noted on a post entitled Mobile 3.0 arrives: How Qualcomm just showed us the future of the cell phone:
In the future our cell phones will know us at a very deep level. Already I’ve told Facebook more than 5,000 things I like. Check out my list. It’s public. On it you’ll see which startups I like. But also that I like Round Table Pizza. Think about that one for a moment.
In the future my cell phone will know I ordered a pizza. Will know when I get in my car. Will know who is in the car with me. And will give me contextual data that will make my life better. For instance, on my todo list I might have put “pick up a hammer at the hardware store.” It will know that Round Table Pizza is near the hardware store. It will know I have an extra 15 minutes. It can use Waze to route me to the hardware store first, tell me to pick up my hammer, and then head to Round Table to pick up that pizza. All while measuring how many steps I took (Nike Fuel points!) and telling me who has crossed my path.
One of the current problems with such solutions is the energy needed to collect and analyze lots of data. Scoble mentioned discussing with Roland Ligtenberg, Product Developer at Qualcomm Labs, who suggested that data collection and analysis might consume less battery power if it was implemented in hardware rather than software, a hint to possible future chips that Qualcomm might produce to enhance Gimbal.
Brandon Holt, Preston Briggs, Luis Ceze, Mark Oskin May 21, 2015