Digital traces of processes executed in the physical world without Information Systems (IS) support are composed of high-level events derived from low-level physical world events (e.g., from IoT). Due to this, frequently these high-level events are ambiguous, i.e., they yield multiple interpretations. As current IS lack ambiguity-awareness, ambiguity in digital traces can compromise process analytics. Motivated by this and current trends toward multi-dimensional analytics, we introduce an ambiguity-aware object-centric representation of digital traces by extending Event Knowledge Graphs with ambiguity. We integrate its construction into a framework to enable analytics for ambiguous IoT-based digital traces. A prototype implementation shows the applicability of the construction from ambiguous process event streams in online settings.