Presentation Information
[BCS-1-08]Mathematical Modeling of Sensing Quality and Resource Allocation Problem in Multi-Modal ISAC
〇Masanori Nakajima1, Takayuki Nishio1, Mehdi Bennis2 (1. Institute of Science Tokyo, 2. University of Oulu)
Keywords:
Multi-Modal ISAC,Resource Allocation,Sensing Quality,CSI Sensing
In sensing with wireless sensor terminals such as wireless cameras, collecting observations consumes wireless communication resources, causing resource competition with data communication terminals. This study focuses on multimodal ISAC integrated with sensing using communication signals, where channel state information (CSI) obtained from radio resources allocated to non-sensor communication terminals can also contribute to sensing. Based on this observation, we formulate a resource allocation problem that maximizes communication capacity while satisfying a sensing-quality constraint. We then model resource-dependent sensing quality from both compressed sensor features and CSI observations using conditional mutual information under a Gaussian approximation. Furthermore, we validate the proposed mathematical model through a pose-estimation task. The results show that both RGB resolution and CSI subcarriers reduce the estimation error. In addition, the proposed sensing-quality prediction model shows a strong correlation with the estimation error. These results demonstrate the usefulness of the proposed model for evaluating allocation-dependent task performance in multi-modal ISAC.