Objectives

This project pushes the borders of accessibility in the harsh and scientifically as well as economically fruitful underwater world. Autonomous underwater robots are a central puzzle piece on the way to learn more about our (by far) largest ecological system on earth. Scientifically, there is still a vast amount to explore and to learn here, while many practical applications are close at hand as well.

The technical objective of this project is to enable underwater vehicles to localize, and move without external control. This is especially important (and hard) over long distances. While the vehicle is self-sufficient in terms of energy, sensors, actuators and computational power, an understanding of the spatial environment and of any potential interaction with it, needs to be build up. This understanding (the model) needs to be suitable for ...

Benchmarks for these systems are:

1. Demonstrate diving and self-sufficiency (passed)
2. Stabilize itself in six degrees of freedom (passed)
3. Collect data via all on board individual sensor systems (passed)
4. Model sensor data abstractions for specific tasks (passed, continuing)
5. Demonstrate visual servoing (passed, continuing)
6. Model multiple sensor-actuator control loops for specific tasks (on going)
7. Employ all sensor systems in closed control loops for specific tasks (in preparation)
8. Demonstrate short term operations in sea environments
9. Demonstrate long term operations in sea environments

See also the publications and theses for current state of the project.

Since real world underwater projects require many different competencies and technical systems we are specifically open for collaborations. If you like to discuss with us about the project or its applications or if you are interested in being a part of the team on this challenging journey please contact us.

Assumptions

• Open systems – autonomous robust systems considered here are operating in open environments, i.e. environments which are not under full control of the system designer. Technically speaking, the interpretation of sensing and senor-actuator coupling can not be pre-designed completely, and needs to be adjusted or reorganized during the systems lifetime.

• Adaptability – as a direct result of the open environments assumption, a robot needs to come up with generic parts and/or adaptive behaviors.

• Complete systems – most components and designs defining a robot operating in open environments matter; including morphology, sensors, actuators, internal structures, operation environment, initial states, kinematics and dynamics – all components are strictly interdependent and fulfil their tasks in a highly specific setup and environment only. Although sub-systems might be experimentally evaluated individually, as long as they describe a closed control loop, including vehicle and environment.

• Specific systems – Current useful autonomous systems are specifically designed according to their actual tasks. Nevertheless, theoretically sound approaches and generic as well as general methods are requested in the frame of the specific 'ecological niche'
  – philosophical remark: Although the idea of (or fear from) 'universal', or even humanoid robots is basically as old as civilized mankind, it is still (almost) as far from reality as it was 3000 years ago. Nevertheless it seemed to be common sense for hundreds of years that such a robot will be created in the near future (see historic sketch for ideas and arts from many centuries).

• Dynamics – interactions in open environments can neither be formulated, nor designed, or analyzed fully using stationary or static models. This touches especially criticisms about the finite automata theory, but also models of asynchronous, parallel, or even analogous processing. The theoretical basis for dynamical systems beyond the terms of attractors of different complexity is very weak still. Nevertheless dynamics are identified as a central aspect and will be attacked and employed wherever possible.

• Time – as motion and dynamics cannot even be defined without a concept of time closely related to the current spatial environment, real time systems and closed control loops are the underlying architectures, which are intrinsic parts of the design.

• Simplicity – robust solutions are simple. This pragmatic and in itself simple sounding insight needs to be taken seriously in a research field, where the whole systems are usually out of the range of analytical methods and engineered designs.

Applications

Out of the rich field of applications, some major future impacts are discussed briefly:

Observations

Marine ecosystems
While most marine ecosystems are not accessible for humans without a significant amount of technical equipment and even then it is only an a short term basis, autonomous underwater vehicles are the alternative at hand. Due to their ability to stay in a certain environment over a long time and to collect data with minimal effects to the biological sub-system itself, these systems might open a new area of marine understanding.

Environmental effects (pollutions)
The influence of humankind on the global environment is only partly understood (although it is obviously - generally speaking - 'negative'). In order to identify the critical substances and areas, a more specific sampling is essential. Autonomous underwater vehicles will enable wide range deep sea sampling strategies as well as dedicated measurements for instance close to polluted sea ports, cities or chemical industries.

Inspections

If the object of investigation is already well known, but needs to be inspected regularly like sea cables, ship hulls, harbour constructions, oil riggs, etc. autonomous vehicles could support in the standard inspection process or enable inspections at places or times, where it is too dangerous or ineffective to employ ships, and divers.