Continuous Maneuver Control and Data Capture Scheduling of Autonomous Drone in Wireless Sensor Networks
Abstract : Thanks to flexible deployment and excellent maneuverability, autonomous drones are regarded as an effective means to enable aerial data capture in large-scale wireless sensor networks with limited to no cellular infrastructure, e.g., smart farming in a remote area. A key challenge in drone-assisted sensor networks is that the autonomous drone's maneuvering can give rise to buffer overflows at the ground sensors and unsuccessful data collection due to lossy airborne channels. In this paper, we propose a new Deep Deterministic Policy Gradient based Maneuver Control (DDPG-MC) scheme which minimizes the overall data packet loss through online training instantaneous headings and patrol velocities of the drone, and the selection of the ground sensors for data collection in a continuous action space. Moreover, the maneuver control of the drone and communication schedule is formulated as an absorbing Markov chain, where network states consist of battery energy levels, data queue backlogs, timestamps of the data collection, and channel conditions between the ground sensors and the drone. An experience replay memory is utilized onboard at the drone to store the training experiences of the maneuver control and communication schedule at each time step.
? In mobile case, data collection algorithms are based on sinks that are usually moving on the ground with lower speeds and static sensors.
? UAVs differ with the traditional mobile sinks as the UAVs fly at a given height with a higher speed, thus, there have been some limitations if existing data collection schemes are fully applied in UAV-based scenario.
? Moreover, most of existing data gathering algorithms aim to improve various performance metrics of static networks.
? The contention based protocols may accommodate a large number of terminals with sporadic traffic but they are not scalable makes a comparison between the existing schemes according to the main MAC performance criteria.
? The trajectory planning is decoupled between multiple subproblems which separately schedule the ground sensors’ transmissions, trajectories, and altitudes of the drone.
? Q-learning typically supports discrete state and action spaces, and therefore is not suitable for the continuous state and action spaces in the drone maneuver problem considered here.
? As part of the network state, the TTA can have a strong impact on the actions of the drone.
? In , the trajectory planning is formulated as a mixed integer non-linear programming to reduce the average path loss between the drone and the ground sensor.
• The proposed solutions have proved superior to conventional scheduling and routing solutions.
• The proposed solutions can be implemented and applied in different application contexts for which the ground nodes are mobile or easily adapted to the case where the nodes are static.
• UAVs have also been proposed for delivering broadband data rates in emergency situations through lowaltitude platforms.
• In the proposed adaptive hybrid MAC protocols, we fully take into account the real-time dynamic topology of the network.
• UAVs have also been proposed as an effective solution for delivering broadband data rates in emergency situations through low-altitude platforms.
? The performance gains keep growing with N. The reason is that DDPG-MC learns the ground sensors’ buffer lengths, battery levels, and channel states, so that the maneuver control and the node selection can minimize the data packet loss of the entire network.
? The reason is that DDPG-MC optimizes the future maneuver control and communication schedules at every location of the drone by taking advantage of the learning experience in the replay memory, which controls ?(a) and v(a) adapting to the data traffic.
? It can also be observed from that the performance gap decreases with ?.
? The observation and evaluation are also used to update the experience replay memory of the drone.
? A desktop with 4-core Intel i7-6700K 4GHz CPUs and 16G memory based on 64-bit Ubuntu 16.04 is used for the TensorFlow setup.
|