AoI-Aware Trajectory Planning for Smart Agriculture Using Proximal Policy Optimization

In the realm of smart agriculture, the integration of Internet of Things Devices (IoTDs) for crop surveillance is pivotal for enhancing agricultural quality and output. However, the limited processing and transmission capabilities of these devices require alternative mechanisms for efficient data collection and transfer. Unmanned Aerial Vehicles (UAVs) emerge as a proficient solution, bridging the gap between IoTDs and more sophisticated systems, such as Multi-access Edge Computing (MEC) servers. Within this framework, the meticulous design of UAV flight paths is imperative to circumvent the processing of obsolete data. Such precision ensures timely access to vital data, for instance, in pest detection, and prevents erroneous actions such as unsuitable pesticide dispensation. This study delves into a UAVMEC-integrated system tailored for data acquisition, transfer, and processing from a multitude of IoTDs dispersed throughout a smart farm. The flight path of the UAV is traced by employing the Proximal Policy Optimization (PPO) approach, based on the data’s timeliness in the IoTDs, quantified by the average Age of Information (AoI). Furthermore, we analyze the potential advantages of this optimized UAV trajectory concerning its battery longevity. Simulations reveal that our model adeptly formulates an effective UAV trajectory, leading to a reduction in the average AoI and simultaneously conserving the UAV’s propulsion energy in designated scenarios.