Program Results

Introduction to the event

The goal of this project is to enable decentralized intelligence across a wide variety of devices, mobile and computing platforms, and systems, ultimately creating a mor e adaptive and resilient decentralized system. To achieve this ambitious vision, the project focused on two fundamental research questions during the first two years. The first research question investigates how different sensing data modalities and data domains can interact with each other to get richer user behavior understanding. This involves integrating diverse data sources such as wireless, visual, and audio signals. For example, combining wireless sensing data with visual trajectories can reveal patterns of user mobility, while adding audio cues can further understand subtle aspects of social interaction among users. To support this, the project designed a generic framework for synergistic perception through cross domain data interaction. This framework is highly modular, allowing new data types or sensing modalities to be seamlessly incorporated. By employing the developed data compensatory techniques, the framework creates an effect of “1 + 1 > 2,” where weaknesses in one modality are compensated by strengths in another. Additionally , data augmentation techniques were explored to improve data quality and robustness. The proposed generic framework has great potential to flexibly combine multiple data modalities, improve data quality via augmentation, and dynamically manage sensing resources. The second research question addresses how sensing, computing, and communication resources can collabo rate to enhance the performance and resiliency of services. For example, computing the paths of mobile relays (e.g., UAVs or robots) according to the sensed mobility patterns and communication demands can provide just in time connectivity in dynamic environments. By analyzing trajectory patterns of mobile platforms, the framework can proactively compute optimal relay paths, ensuring seamless communication between devices, mobile platforms, and access points. The proposed framework has great potential to be applied to flexibly manage sensing, computing, and communication resources between connected user devices, mobile platforms, edge platforms, and systems.