The general dynamic model of an underwater vehicle follows the 6 degrees of freedom template presented in (Fossen, 1994). It is a complete model that takes into account all the forces and moments acting on a submerged body, but for which it is not easy to design adequate controllers. Typically, AUVs evolve according to two dimensional motions, either in the vertical or in the horizontal plane. Therefore, the traditional approach for the control of theses vehicles is based on mode decoupling (Healey & Lienard, 1993). Following such approach, the MARES control system is organized into four basic controllers: surge, heading, pitch and depth. The first two determine the horizontal motion of the vehicle and their outputs are combined to obtain the actuation for the horizontal thrusters. The outputs of the pitch and depth controllers are combined to provide the actuation for the vertical thrusters. Each one of the four basic controllers can operate in either closed or open loop mode.
At the lowest level, all four controllers are assumed independent. Surge is controlled by the common mode of horizontal thrusters, heading is controlled by the differential mode of these thrusters, while depth is controlled by the common mode of vertical thrusters, and pitch by their differential mode. In this approach, the couplings between modes are treated as external disturbances, which must be taken into account when designing decoupled feedback controllers. This typically results in a small reduction of performance, which is largely balanced by the simplicity of the design and by the modularity of the approach. Each one of these controllers can operate in different modes, ranging from a pure open loop operation to more complex structures with more than one feedback loop, as follows: Surge:
Open loop — input is a direct common mode command for the horizontal thrusters
Velocity loop — input is a surge velocity reference
Open loop — input is a direct differential mode command for the horizontal thrusters
Velocity loop — input is a heading rate reference Position loop — input is a heading reference Line tracking loop — input is a directed horizontal line reference
Open loop — input is a direct common mode command for the vertical thrusters
Velocity loop — input is a heave velocity reference Position loop — input is a depth reference
Open loop — input is a direct differential mode command for the vertical thrusters
Velocity loop — input is a pitch rate reference Position loop — input is a pitch reference
The autonomous operation of MARES is defined by a mission plan. Besides configuring a large set of variables that affect the vehicle behaviour (such as controller gains, maximum operating depth, operating frequencies of the acoustic systems, etc.), the mission plan also includes a set of elemental maneuvers that the vehicle should execute in sequence. Each maneuver prescribes the behaviour all the four basic controllers (therefore defining the vehicle motion). It also defines its end condition, and a timeout for safety reasons. Besides some maneuvers that are mainly used for debugging purposes, the basic MARES maneuvers are:
• dive - a downwards maneuver, typically executed at the start of a mission or in depth transitions;
• surface - an upwards maneuver, typically executed at the end of a mission or in depth transitions;
• hovering - a maneuver that stops the vehicle at the current position;
• gotoxy - a horizontal plane maneuver that drives the vehicle along a straight line.
The possibility of independently defining each basic controller allows for very different vehicle behaviours, making the operation of MARES very flexible. For example, a pure vertical motion can be easily obtained by a dive maneuver with a closed loop pitch with zero reference, and a zero surge command; a combined vertical and horizontal motion can be achieved with a dive maneuver with a closed loop pitch with a negative (downward looking) reference and an appropriate surge command.
Furthermore, each basic controller is already prepared to accept inputs defined by external processes. This allows for the implementation of unconventional guidance strategies which can be based on payload data collected in real time.
Was this article helpful?