Disturbance Rejection

Dagon-AUV in the test basin (Source: DFKI GmbH)
Dagon-AUV in the test basin (Source: DFKI GmbH)

State of the Art:

An Autonomous Underwater Vehicle (AUV) expected to perform operations such as cleaning, inspection & (minor) repair has to compensate for the forces & torques applied on it by its natural environment or due to utilization of specific tools.

ROBOCADEMY aims to detect, interpret and exploit local flow to improve control stability and energy efficiency, contributing to this to date very poorly researched topic.

Regarding the control of under‐actuated AUVs in a steering control scheme for tracking 2D smooth trajectories using back stepping techniques is designed. In a feedback control law was developed, which asymptotically stabilizes longitudinal‐axis translation of a streamlined underwater vehicle in any desired, non‐vertical direction. In both approaches, the vehicle follows a predefined path, without taking into consideration crucial limitations such as field of view requirements.

A guidance and control architecture integrated with a single camera sensor is introduced in order to handle the high degree of uncertainty which characterizes underwater operations.

Progress beyond the State of the Art:

In order to make the low‐level UUV platform as “transparent” and stiff as possible to higher levels of decision making, we will address:

Disturbance compensation: Regarding stabilization we plan to introduce the viable solutions. Viability will require less control effort, leading to higher energy efficiency. The Navigation Functions framework used and further developed by NTUA for non‐holonomic systems along with tools from viability theory will be combined and extended.

Interaction Control: The Compliance / Interaction Control principles will incorporate the non‐holonomic actuation specific to the under-actuated setup, composed of the combined UUV‐end effector system, that may not be able to readily generate the required forces & torques.

Researchers:

© DFKI GmbH
last updated 26.08.2015
to top