Abstract
Unmanned ocean vehicles can be guided and controlled autonomously or remotely, and even remote operation can be automated significantly. Classical methods use trajectory tracking errors in negative feedback. Recently published methods are proposed instead. Deterministic (non-stochastic) artificial intelligence (DAI) combines optimal learning with an asserted self awareness statement in the form of the governing mathematical model (based on physics in this instantiation) to allow control that can be alternatively adaptive (i.e., capable of reacting to changing system dynamics) or learning (i.e., able to provide information about what aspects of the system dynamics have changed). In this manuscript, deterministic artificial intelligence is applied to the heading control of a simulated remotely operated underwater vehicle (ROV). Research is presented illustrating autonomous control of a Seabotix vLBV 300 remotely operated vehicle within milli-degrees on the very first step of a shaped square wave command, and error decreased an additional sixty-two percent by the third step of the square wave command.
Highlights
Unmanned vehicles are increasingly popular in society and have proven especially useful in dangerous environments
Materials and Methods This section builds the proposed methodology from first principles cited in the Introduction, naturally starting with the system dynamics which lead to the governing differential equations of motion, which are quickly re-parameterized to permit two-norm optimal feedback learning
The inputs are applied to the system, and measurements of the system state are fed into a state estimator, and the estimated system state is used to optimally update the estimates for the parameters of the system dynamics model
Summary
Unmanned vehicles are increasingly popular in society and have proven especially useful in dangerous environments. Hamilton introduced alternative formulation methods in 1834 using energy [7], while Lagrange introduced a presentation using energy that more closely resembled Newton’s original formulation [8,9], and in the last century Kane [10–12] provided the latest parameterization of the same natural relationships very often referred to as a Lagrangian formulation [8,9] of D’Alembert’s principle, which was introduced in 1743 and published in 1747 [13,14] and 1750 [15] These fundamental and well-understood principles of mechanics embodied the approaches to control themselves in an open-loop, feedforward sense, as expressed in the 1950’s and 1960’s, respectively, by Bellman in the West [16] and Pontryagin in the East [17]. Pontryagin’s work applied to highly flexible robotic vehicles [18] sprung from their foundation of the core principles of adaptive techniques late in the last century [19–21]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
