Ultra-wideband (UWB) localization is one of the most promising indoor localization methods. Yet, non-line-of- sight (NLOS) positioning scenarios remain a challenge and can potentially cause significant localization errors. In this work, we leverage the collaborative paradigm of a multi-robot system by sharing relative positioning information, and thus alleviating error susceptibility in NLOS ranging scenarios. In particular, we detail a decentralized particle filter based localization algorithm which combines an UWB range model with a robot detection model. Finally, we test both collaborative and non-collaborative versions of our algorithm in simulation, in mixed LOS/NLOS scenarios. Results show superior performance for the collabora- tive system when compared to non-collaborative systems utilizing only UWB ranging