Privacy-Preserving Multi-Agent Marine Data Collection via Differential Privacy

Data collection is a resource-intensive process that significantly impacts the detection and mitigation of natural disasters. To overcome limitations in available datasets, it becomes necessary to involve mutually distrusting parties in collaborative data collection missions. This paper presents two contributions. Firstly, we propose a pipeline based on differential privacy to enable privacy-preserving multi-agent data collection in marine environments. This pipeline ensures that data collected by external participants, acting as semi-honest third parties, is securely pooled and anonymized by a trusted data steward. Secondly, we demonstrate the implementation of two cost-effective unmanned surface vessel designs using readily available components for conducting hardware-based experiments. The resulting anonymized data can be utilized in various marine applications, thereby facilitating improved decision-making processes. Through this research, we aim to address the challenges associated with data collection in marine environments and foster collaboration among stakeholders for effective disaster management.