Integrating Foundation Models with Open RAN for Robots-Based Mobile Scenarios

The advancement of Artificial Intelligence (AI) has rendered robots an appealing paradigm for various applications, such as parcel sorting and handling in warehouse management. In this context, rapidly evolving mobile communication technology facilitates robot cooperation for efficient parcel sorting and handling. However, traditional AI solutions often struggle with unreasonable resource utilization, which makes it challenging to achieve both accurate parcel sorting and low-latency parcel handling. To address this issue, we propose a Foundation Model (FM)-empowered Open Radio Access Network (O-RAN) framework to achieve highly accurate and real-time robot cooperation in a pipeline manner. We first propose an adaptive FM splitting algorithm that decouples the sorting mission into multiple tasks to enable the robots to perform sequential task implementation for high-accuracy parcel sorting. Next, we present a cooperative path planning algorithm that allows a feasible number of robots to implement cooperative parcel handling with low latency and reduced energy consumption. Finally, we demonstrate the effectiveness of our proposed framework. The simulation results indicate that our solution achieves parcel sorting accuracy of up to 90% while reducing parcel handling latency by an average of 13.9% compared to traditional FM solutions.