Section: Research Program
Organization of context
Building a context results in the creation of contextual information. The TACOMA research team is focused on placing this information in the pervasive systems, and in mechanisms to allow applications to access and develop it.
Placement. Our approach seeks to promote local placement while allowing the recovery of data in the cloud whenever relevant. We explore two possibilities. The first is to store information (or semantics) directly on objects (a physical object, embedded computer, sensor, etc.). This involves addressing complex issues for the objects and storing some information with very limited computing power. The other approach is to rely on one or more local coordinators (a node placed in the vicinity of objects). These coordinators can be organized into a network or hierarchical system, in order to provide sufficient processing capacity to build and store a context.
Access control. The system should provide access control mechanisms for information collected in the environment and for the various actuators available. This is an essential aspect for context-aware applications, such as medical applications that can use or build patient information. We study mechanisms to restrict access only to authorized applications. For example, we address the problem of access priority for the actuators during the interaction with the environment. Critical applications such as fire detection applications must be able to preempt the actuators, even when these are normally operated by other applications.
Context enrichment. The construction of the context may be carried out permanently for all applications. In fact, each application has specific needs and a vision of the particular context. It is impossible to predict all the data that will be necessary for the proper operation of applications a priori. Each application must therefore build a set of relevant contextual information independently. We seek to provide software mechanisms that enable those applications to enrich the context and share it with other applications if necessary. For example, a set of services created by a public building operator can calculate contextual information related to the use of energy in the building and share it. The storing and organizing of this information (on objects, coordinators, or in the cloud) is supported by the pervasive system, thus alleviating the burden of building applications for developers.