An investigation of the critical timescales needed for digging in wet and dry soil using a biomimetic burrowing robot

Abstract

The Atlantic razor clam, Ensis directus, burrows underwater by expanding and contracting its valves to fluidize the surrounding soil. Its digging method uses an order of magnitude less energy than would be needed to push the clam directly into soil, which could be useful in engineering applications such as anchoring and sensor placement. The first chapter of this thesis presents the theoretical basis for the timescales necessary to achieve such efficient digging and gives design parameters for a device to validate the timescales. It then uses RoboClam, a robot designed to imitate the razor clam's movements, to test the design rules. It was found that the minimum contraction time is the most critical timescale for efficient digging, and that efficient expansion times vary more widely. The results of this chapter can be used as design rules for other robot architectures for efficient digging, optimized for the size scale and soil type of the specific application. The second chapter of this thesis examines whether it would be theoretically possible to use the same E. directus-inspired method to dig into dry soil, for applications such as sensor placement. The stress state of the soil around the robot was analyzed, and a target stress state for dry soil digging was found. Then, the two possible modes of soil collapse were investigated and used to determine how quickly the robot would have to contract to achieve the target stress state. It was found that for most dry soils, a RoboClam-like device would have to contract in 0.02 seconds, a speed slightly faster than the current robot is capable of, but still within the realm of possibility for a similar machine. These results suggest that the biomimetic approach successfully used by RoboClam to dig into submerged soil could feasibly be used to dig into dry soil as well.