Autonomous Hanging Tether Management and Experimentation for a UAV-USV Team: Sea Trials

This paper considers the mission of a UAV-USV team where a small unmanned surface vehicle (USV) is tethered to an unmanned air vehicle (UAV). In contrast to a majority of existing tethered UAV work, which assumes a taut tether, this paper addresses the challenge of managing a slack, hanging tether in a dynamic ocean environment up to sea state 4 on the Douglas scale. Our previous work developed and experimentally validated a prototype smart-reel system for a UAV-USV team in a lab-based environment. That system used a reference model for tether control based on static catenary cable theory that equally weighted an increase in tether tension with the risk of fouling the tether with the USV or water. In conjunction, a Kalman filter was developed – fusing slow (4 Hertz) differential GPS (dGPS) relative position measurements with fast (100 Hertz) inertial measurements. The output of the Kalman filter, a 100 Hertz estimate of relative position, was fed into the tether length reference model and a gain-scheduled proportional-derivative (PD) controller drove the smart-reel system to the desired tether length. This paper presents an extension of that work to controlled on-water dynamic wave experimentation and randomized excitation through sea trials.A surrogate UAV was used in wave-pool experimentation to tune the Kalman filter for on-water use. Using this tuned filter, nine separate UAV flights successfully validated the slack tether management approach in varying head-on wave heights up to 1.0 m as well as two trials of random period irregular wave heights of 0.6 m. The trials were performed using alternating feedback methods for relative position: first using ground-truth motion capture feedback and then using Kalman filter feedback with de-rated ground-truth to replicate the dGPS sensor at 4 Hertz. The experiments showed robustness to the heave and pitch motions of the USV, effectively decoupling its motion from that of the UAV. The resulting effect is an increased payload capacity and altitude limits for the UAV. In addition to successfully demonstrating the ability to manage a slack tether, reflective markers were added to the tether to generate a unique UAV-tether-USV dynamic motion dataset. This data will be used to validate tether dynamic simulation models in ongoing work.Following indoor wave-pool testing, the dGPS system was upgraded to a newer model which outputs an 8 Hertz measurement. Significant effort was put into the setup of the radios on each end of the tether (TCP/IP with 921600 Baud) to limit latency to ensure that no dGPS packets were dropped. Our initial sea-based experimental testing revealed that the catenary reference model for tether control needed to be modified to be gain-scheduled for relative position ratios (∆r/∆z) up to 10. Lastly, significant effort was put into the software system to ensure that the timing of the digital control system was as deterministic as possible. By ensuring that the Kalman filter, controller, and output command to the motor all occur at their target frequencies, any buildup of errors and delays was minimized – resulting in improved performance. For sea trials, a new waterproof electronics enclosure was developed using industrial electronic components (as we learned during wave-pool testing that Murphy’s Law tends to hold true: anything that can break will break), and the smart-reel system was installed on the bimini of a small 6.5 m Boston Whaler. Three separate 30 minute flights were performed at anchor just outside San Diego Bay. The tether management system operated effectively and the experiment was a success.The development and validation of an autonomous hanging tether management system for a UAV-USV team operating in a dynamic ocean environment up to sea state 4 is a significant technical achievement with far-reaching implications. The system’s ability to manage a slack, hanging tether has been validated through both wave-pool testing and on-water sea trials, demonstrating robustness to heave and pitch motions of the USV, increasing the UAV’s payload capacity and altitude limits. Additionally, the generation of a unique UAV-tether-USV dynamic motion dataset provides a valuable resource for validating tether dynamic simulation models. Beyond military applications, this system has potential benefits for various industries including fishing, environmental conservation efforts, and other oceanographic research. This work represents a significant step forward in the development of autonomous, long-endurance, tethered UAVs.