[Paper Review] Visual Servoing of Unmanned Surface Vehicle from Small Tethered Unmanned Aerial Vehicle

Haresh Karnan, Aritra Biswas+3 more|arXiv (Cornell University)|Oct 9, 2017

Robotics and Sensor-Based Localization8 references3 citations

TL;DR

This paper presents a motor schema for autonomous visual servoing of an Unmanned Surface Vehicle (USV) using a small tethered Unmanned Aerial Vehicle (UAV), enabling the UAV's camera to track the USV with minimal pan/tilt movements. The system uses relative position error from visual localization to generate yaw and pitch control signals via a serial interface, achieving 100% tracking success in lab trials with normalized sensitivity of 0.034.

ABSTRACT

This paper presents an algorithm and the implementation of a motor schema to aid the visual localization subsystem of the ongoing EMILY project at Texas A and M University. The EMILY project aims to team an Unmanned Surface Vehicle (USV) with an Unmanned Aerial Vehicle (UAV) to augment the search and rescue of marine casualties during an emergency response phase. The USV is designed to serve as a flotation device once it reaches the victims. A live video feed from the UAV is provided to the casuality responders giving them a visual estimate of the USVs orientation and position to help with its navigation. One of the challenges involved with casualty response using a USV UAV team is to simultaneously control the USV and track it. In this paper, we present an implemented solution to automate the UAV camera movements to keep the USV in view at all times. The motor schema proposed, uses the USVs coordinates from the visual localization subsystem to control the UAVs camera movements and track the USV with minimal camera movements such that the USV is always in the cameras field of view.

Motivation & Objective

• To automate UAV camera control for continuous visual tracking of a USV during search and rescue operations.
• To reduce operator workload by eliminating manual camera steering during USV navigation.
• To develop a low-power, computationally efficient control system for real-time pan/tilt adjustments based on visual feedback.
• To validate the system in controlled laboratory environments with measurable tracking performance.
• To enable autonomous USV-UAV teaming for emergency response in marine casualty scenarios.

Proposed method

• A motor schema computes relative position error $P$ between the USV and the center of the camera's region of interest (ROI), using visual localization data.
• Control is activated only when $P > 1$, minimizing unnecessary camera movements.
• Yaw and pitch commands are generated based on the sign and magnitude of $P$ to steer the camera toward the USV.
• Commands are transmitted via USART serial interface at 9600 bps to the Fotokite UAV ground station.
• The system uses a fixed-gain control law with pan/tilt velocities of 0.3 rad/s, implemented in real time during trials.
• The algorithm is tested in two laboratory arenas with a red object simulating the USV, using a tethered UAV with a PTZ camera.

Experimental results

Research questions

• RQ1 Can a low-complexity motor schema autonomously track a moving USV in real time using visual feedback from a tethered UAV?
• RQ2 How effective is the system in minimizing camera movements while maintaining continuous USV visibility?
• RQ3 What is the sensitivity and success rate of the tracking system under varying motion profiles in controlled environments?
• RQ4 Can the system maintain tracking across different distances and trajectories without manual intervention?
• RQ5 How does the system perform in terms of response time and control effort in real-world test conditions?

Key findings

• The system achieved 100% tracking success in 13 trials in Arena 1 and 17 trials in Arena 2, with no loss of target.
• The mean sensitivity across all trials was 0.6045 in Arena 1 and 0.4536 in Arena 2, with a normalized sensitivity of 0.034.
• Pitch and yaw movements were largely mutually exclusive, reducing redundant camera motion and improving energy efficiency.
• The control system responded effectively within a few seconds after the USV exited the ROI, demonstrating real-time responsiveness.
• The algorithm successfully maintained the USV within the camera's field of view using minimal control signals, even during rapid movements.
• The results confirm the feasibility of automating UAV camera tracking for USV support in marine search and rescue operations.