Sensing context information and making it available to the people, involved in coordinating a collaborative task, is a preliminary phase in making a system adaptable to the prevailing situation in pervasive environments. However, the diversity of the sources of context information, the characteristics of pervasive environments, and the nature of collaborative tasks pose a stern challenge to the efficient management of context information by sensing a lot of redundant and conflicting information. Quality of Context parameters can be used to resolve the conflicts in context information. In this paper, we present a context aggregation system that detects and removes the duplicates and conflicts from context information by using the policies based on Quality of Context parameters. This system effectively aggregates the continuously evolving context information and efficiently uses the scare resources in pervasive environments.