M.Sc Thesis

M.Sc StudentBarenboim Moran
SubjectSteerable Burrowing Robot: Design, Modeling and Experiments
DepartmentDepartment of Autonomous Systems and Robotics
Supervisor ASSOCIATE PROF. Amir Degani


Locomotion of robots on land, water, and air has been possible for years. However, locomotion underground is a challenging task for robots and has yet to be fully addressed. Underground robots can be useful in several applications, such as laying underground cables, detonating mines, sensing for precision agriculture, or even gas exploration. Several studies were conducted in the field, yet little was reported on the ability of robots to steer underground. 

Most of the recent studies that focused on designing a mechanism that can locomote underground can be categorized into three types: drilling robots, biomimetic robots and hammering robots. From the three types of mechanisms, the hammering robots were shown to be able to penetrate substantially deeper than other mechanisms.

The current study investigates the design and modeling of a simple, self-propelled, steerable robot in granular medium. In order to be able to steer inside the granular medium we use an asymmetric bevel-tip that, by interacting with the granular medium, causes the robot to move along a curved trajectory. To thrust within the granular medium we use a hammering mechanism, which comprises a linear thrusting actuator connected through a spring to the robot's frame. Supplying a square-wave signal to the thrusting actuator causes the robot to move in an oscillatory motion that result in a net displacement over time.

To describe the robot's motion, we developed a simplified model for the steering and the thrusting mechanisms. We model the steering mechanism as a simple kinematic car, also known as the bicycle model, which uses nonholonomic constraints to describe the robot's trajectory. We modeled the thrusting mechanism assuming two rigid bodies connected by a spring. To complete the thrusting mechanism calculation, we showed an approximation of the normal forces acting on it in the granular medium and assumed linear relation to the friction forces.

Using the models for the steering and thrusting mechanisms we performed numerical simulations demonstrating a typical motion of the robot. Using the simulation, we performed parameter sweep on the input signal parameters of the thrusting mechanism. By doing so, we discovered that certain values of the input signal parameters may affect not only on the robot's velocity, but also on the robot's direction of motion and cause it to move in reversal direction.

We conducted experiments to validate the design and model of the robot. In the experiments, we verified that different parameters of the input signal may indeed cause the robot to move in reversal direction as was suggested in the simulation. We have concluded with a proof-of-concept experiment, showing the robot's ability to locomote underground and perform a change in the robot's trajectory by changing its bevel-tip orientation. These experiments demonstrate the feasibility of the robot's design and fidelity of the model.