A Distributed Sensor Network Architecture for UAVs with Heterogeneous Processing Hardware

Many industrial applications require mobile sensors that can provide high coverage capacity and operational flexibility. Unmanned aerial vehicles (UAVs) represent a promising solution due to their precise maneuverability and ability to operate in targeted environments. The effectiveness of these mobile sensors is further enhanced when multiple UAVs collaborate to form a networked mobile sensor platform (NMSP), in which each UAV shares its local measurements with others. However, the heterogeneous nature of NMSPs in terms of hardware and software makes it difficult to communicate individual measurement data among UAVs in a unified and structured framework. This work provides a framework for sensor data exchange among UAVs. The design is organized into three layers: Low-level control, process management, and communication. The low-level control layer is centered on a flight control unit (FCU), while for the process management is implemented through customized modules under the robotic operating system 2 (ROS 2) middleware running on the OBCs. The communication layer focuses on the design of the sensors' networking and is developed within the ROS 2 framework. We present experimental results using three UAVs sharing measurements from their GPS sensors as well as an onboard light detection and ranging (LIDAR) sensor.