Modeling, Control and Energy Efficiency of Underwater Snake Robots

摘要

This thesis is mainly motivated by the attribute of the snake robots that they
\nare able to move over land as well as underwater while the physiology of the robot
\nremains the same. This adaptability to different motion demands depending on the
\nenvironment is one of the main characteristics of the snake robots. In particular,
\nthis thesis targets several interesting aspects regarding the modeling, control and
\nenergy efficiency of the underwater snake robots.
\nThis thesis addresses the problem of modeling the hydrodynamic effects with
\nan analytical perspective and a primary objective to conclude in a closed-form
\nsolution for the dynamic model of an underwater snake robot. Two mathematical
\nmodels of the kinematics and dynamics of underwater snake robots swimming in
\nvirtual horizontal and vertical planes aimed at control design are presented. The
\npresented models are derived in a closed-form and can be utilized in modern modelbased
\ncontrol schemes. In addition, these proposed models comprise snake robots
\nmoving both on land and in water which makes the model applicable for unified
\ncontrol methods for amphibious snake robots moving both on land and in water.
\nThe third model presented in this thesis is based on simplifying assumptions in
\norder to derive a control-oriented model of an underwater snake robot moving in a
\nvirtual horizontal plane that is well-suited for control design and stability analysis.
\nThe models are analysed using several techniques. An extensive analysis of the
\nmodel of a fully immersed underwater snake robot moving in a virtual horizontal
\nplane is conducted. Based on this analysis, a set of essential properties that characterize
\nthe overall motion of underwater snake robots is derived. An averaging
\nanalysis reveals new fundamental properties of underwater snake robot locomotion
\nthat are useful from a motion planning perspective.
\nIn this thesis, both the motion analysis and control strategies are conducted
\nbased on a general sinusoidal motion pattern which can be used for a broad class
\nof motion patterns including lateral undulation and eel-like motion. This thesis
\nproposes and experimentally validates solutions to the path following control problem
\nfor biologically inspired swimming snake robots. In particular, line-of-sight
\n(LOS) and integral line-of-sight (I-LOS) guidance laws, which are combined with
\na sinusoidal gait pattern and a directional controller that steers the robot towards
\nand along the desired path are proposed. An I-LOS path following controller for
\nsteering an underwater snake robot along a straight line path in the presence of
\nocean currents of unknown direction and magnitude is presented and by using a
\nPoincaré map, it is shown that all state variables of an underwater snake robot,
\nexcept for the position along the desired path, trace out an exponentially stable periodic orbit. Moreover, this thesis presents the combined use of an artificial potential
\nfields-based path planner with a new waypoint guidance strategy for steering
\nan underwater snake robot along a path defined by waypoints interconnected by
\nstraight lines. The waypoints are derived by using a path planner based on the
\nartificial potential field method in order to also address the obstacle avoidance
\nproblem.
\nFurthermore, this thesis considers the energy efficiency of underwater snake
\nrobots. In particular, the relationship between the parameters of the gait patterns,
\nthe forward velocity and the energy consumption for the different motion patterns
\nfor underwater snake robots is investigated. Based on simulation results, this thesis
\npresents empirical rules to choose the values for the parameters of the motion
\ngait pattern of underwater snake robots. The experimen