Temporal data alignment and association for event-streaming in ubiquitous environments based on fuzzy-causal dependencies

Advances in sensors as well as wireless networks and ad-hoc networks have allowed new applications based on ubiquitous environments (UE) to be conceived. The deployment of a UE implies the exchange of numerous data streams asynchronously generated by multiple distributed sources. Such data must be fusioned at a certain stage. To perform the data fusion, the involved devices need to agree on some common temporal references of the whole system. To attain such references, the current solutions propose the use of centralized schemes and/or global references. Unfortunately, in a UE it is difficult to get global references mainly due to the asynchronous execution nature of the distributed systems. In this paper, we propose a distributed data alignment and association approach that establishes temporal references among the exchanged data streams, without requiring the use of synchronized clocks or centralized schemes. This is achieved by translating temporal/spatial references based on a time-line and physical locations to fuzzy-causal dependencies among streams. Through the establishment of the fuzzy causal dependencies, we infer a degree of temporal closeness among the data streams, which can be useful to correlate, filter or combine such data at a later processing.