Institute for

The automated shell: design of a burrowing robot

A successful cooperation in the framework of a research project at the University of Zurich.


While bachelor students of the HSR gain insight into an interdisciplinary field of research, the Artificial Intelligence Lab at the University of Zurich has benefited from their excellent practical and technical skills.

Clams are commonly known as rather exotic dish or as producers of shimmering pearls. But the muscular body not only delivers edible meat, and the deposition of lime not only pearls. The muscles and the different-shaped calcareous shells allow digging and anchoring in the sediment. The digging into the sediment protects the clam from its natural predators. Some clams bore into rocks, anchor themselves to other objects, lie loosely on the seabed or even swim through the water by flapping their shell halves.


There is an affinity between the locomotion of clams and worms: while a part of the body expands and serves as an anchor, another part is pushed or pulled forward. Clams use two anchors: shell and foot. The latter is a tongue-like extension of the soft body that drills into the sediment. It corresponds to the elongated foot used by snails to crawl, which also belong to the group of shell molluscs. Shell and foot alternately anchor themselves in the sediment (by opening or swelling), while the other part of the body is dragged along in the course of movement.

Interdisciplinary Project

At the Artificial Intelligence Laboratory of the University of Zurich the influence of the body shape on the behavior of an organism is generally examined. For this project, the relation between the shape of the shell and the digging behavior of clams was investigated. In contrast to biologists, who would do this by observing real clams, we tried to build the clams ourselves. With this "synthetic approach" in contrary to the "analytical approach", the natural system is not only analyzed, but also synthesized, thus modeled artificially. Although this approach is more complicated, it has the advantage that it faster reveals if the developed theory is correct. In case the replica does not function, one has obviously not yet understood the phenomenon properly.


The Institute for Laboratory Technology has designed and built an experimental system to simulate the digging behavior of clams. This system consists of an aquarium filled with water and sand, a clam-model and an external drive, which transmits the digging motions onto the model. Two linear motors, with attached ropes pull alternately the model into the sand. The model illustrates the shell behavior. The main advantage of a self-built clam is that the used shape is completely controllable. A real clam-shell can serve as a model, or artificial clam forms can be generated using mathematical models. The shapes are then realized by a 3D printer. Experiments, in which only one parameter is changed at a time, allow a systematic study of the influence of the shape on the efficiency of digging. A mechanism for opening and closing of the shell as well as an artificial foot are planned.


Another involved discipline is paleontology. For some extinct clam species it is not clear how they lived and on what surface. With our experimental environment, it would be possible to scan and replicate the relevant fossils, in order to subsequently find the optimal movement patterns and sediment for a specific shape. It is also planned to imitate evolution by applying mutation and selection on some clam shells. It will be interesting to see to what optimized shell shapes this experiment leads.

Successful Collaboration

This dissertation at the University of Zurich is funded by the Swiss National Science Foundation (SNSF). The project was supported by the ILT under supervision of Prof. Dr. Agathe Koller, in the course of three previous bachelor theses. The students applied a systematic development methodology and were very motivated to participate in this research project combining artificial intelligence, robotics, biology and paleontology.


The collaboration allowed optimally use of the complementary skills of the University of Zurich and the University of Applied Sciences Rapperswil. The cooperation also led to a publication at ICRA 2010, Conference on Robotics and Automation, one of the largest robotics conferences worldwide.

Autor: Daniel Germann; Doktorand, Labor für Künstliche Intelligenz, Universität Zürich

Please contact us for further information.

Download Project Outline