Optimal Proactive Caching for Multi-View Streaming Mobile Augmented Reality

Mobile Augmented Reality (MAR) applications demand significant communication, computing and caching resources to support an efficient amalgamation of augmented reality objects (AROs) with the physical world in multiple video view streams. In this paper, MAR service is decomposed anchored at different edge cloud locations optimally explore scarce resources, especially during congestion episodes. that way, proposed scheme enables processing popular streams embedded AROs. More specifically, we explicitly utilize notion content popularity not only synthetic but also case, are cached a proactive manner, together preferred/popular AROs, selected improve overall user experience mobility events. To achieve that, joint optimization problem considering mobility, decomposition, balance between delay preference AROs proposed. tackle curse dimensionality problem, nominal long short-term memory (LSTM) neural network proposed, which trained offline optimal solutions provides high-quality real-time decision making within gap 5.6% 9.8% inference. Evidence from wide set numerical investigations shows schemes owns around 15% 38% gains hence substantially outperforms schemes, oblivious inherent multi-modality potential decomposition services.