Doctoral Defense - Eric Beskell

About the Event

The Department of Electrical and Computer Engineering  

Michigan State University  

Ph.D. Dissertation Defense 

Wednesday, April 30, 2025, at 10:30 am 

Mechanical Engineering Conference Room 2555D and Zoom

Contact Department or Advisor for Zoom Information

ABSTRACT

MODELING AND CONTROL OF AQUATIC SURFACE ROBOTS WITH APPLICATION TO ACOUSTIC TELEMETRY

BY: ERIC BESKELL

ADVISOR: DR. XIAOBO TAN

Aquatic robots have recently generated significant interest as sampling platforms for environmental studies since their mobility allows for a greater breadth of experiments than more established stationary sensors. Two types of surface vehicles are considered in this work: steerable drifters and uncrewed surface vehicles (USVs). Comparing to underwater vehicles, both drifters and USVs have the advantages of being able to communicate with RF signals as well as localize themselves readily with GPS. 

Steerable and active drifters have emerged as promising new platforms that can modify their trajectories using rudders and through sparing application of thrusters. Such drifters retain the energy-efficiency advantage of passive drifters with added maneuverability. This work focuses on the modeling, analysis, and control of drifters with the motivation to use them as mobile sampling platforms in environments with pronounced ambient flows such as rivers and lakes with circulation structures. A dynamic model is developed for steerable and active drifters, which is then used for adaptive parameter estimation and reachability estimation. Optimal control of drifters to maintain them as energy-efficient platforms is also investigated in a one-dimensional setting. Sets of state-space locations where the optimal control changes can be thought of as curves parameterized by one of the costate variables. Such curves enclose regions in the state-space for which the optimal control is the same, allowing the state-space to be mapped to the optimal control. The optimal control may then be determined by state feedback, which is shown to be more robust than open-loop optimal control in simulation. The proposed optimal control is further validated by data collected in experimental trials. 

We further investigate the use of USVs for sensing applications such as acoustic telemetry. Aquatic acoustic telemetry is a widely-used method for tracking movements of animals and studying their behaviors. In acoustic telemetry animals are implanted with an acoustic tag broadcasting a unique identification code, which is decoded by an acoustic receiver. USVs and and Autonomous Underwater Vehicles (AUVs) are being increasingly used as mobile receiver platforms, but their performance as acoustic telemetry platforms is not well studied.  

This work characterizes the performance of a USV as an acoustic telemetry platform, which is then used to improve estimates of tag locations and to optimally select waypoints. The USV considered is intended to be a low-cost, easy-to-operate platform that can be constructed by non-experts. Detection efficiency, or the probability of successfully detecting an acoustic tag, is examined as a function of distance between the tag and the receiver. The effect of thruster use is also examined. This characterization is then used with a genetic algorithm to determine the optimal set of waypoints for the USV to localize acoustic tags, considering the trade-off between distance traveled and estimated localization error. Data consisting of successful and failed detections are used to determine the maximum a posteriori estimates of tag locations. Tag locations estimated from experimental data using the proposed method are shown to be more accurate than the typical presence-absence estimates of tag locations.  

The control of USVs has been widely studied; however, prior work has been primarily focused on continuous-time controllers. Limitations on computational power and actuator response can make such controllers difficult to implement. Event-triggered control, a control paradigm in which the applied control is updated when a condition on the state (called an event-trigger) is met, addresses these limitations. Prior work on event-triggered control has required that the continuous-time closed-loop system be input-to-state stable with respect to sampling error in the control. Due to the nonlinear, non-holonomic and underactuated nature of existing USV models, however, it may be difficult to prove input-to-state stability. This work proposes a novel means of designing event-triggers that relaxes this stability requirement. The proposed event-triggers are designed directly from a Lyapunov function for the system rather than comparison functions used in previous work on event-triggered control. For a continuous-time controller that is globally asymptotically stable, the control sampled using the proposed event-trigger is also globally asymptotically stable. This approach is validated with a simulation example. To mitigate Zeno behavior, a minimum inter-sampling interval is included in the design of the event-trigger. Although stability of the closed-loop system cannot be guaranteed with a minimum inter-sampling interval, convergence of system states to desired values is observed in simulation. 

Persons with disabilities have the right to request and receive reasonable accommodation. Please call the Department of Electrical and Computer Engineering at 355-5066 at least one day prior to the seminar; requests received after this date will be met when possible.