Autonomous Motion Control of Bacteria Powered Microrobots Using Electric Fields

Abstract

Recently, small scale robots gain a lot of interests and attention due to their growing capability to perform microscale tasks. These micro/nanorobots could be useful for various specific purposes such as therapeutic targeting drug delivery, micro-manipulation, and micro-assembly. For example, magnetotactic bacterial robots had been used to transport therapeutic agents in the capillaries. However, most of research results have been implemented by manual control of users and there have not been lots of works that show that autonomous control algorithm for microswimming robots. In order to apply the cutting-edge technology of microrobotics to the practical fields, the autonomous motion control is essential. Comparing to the manual control, the autonomous motion control enables the developed microrobots to achieve the given mission with efficiency and optimal control inputs. In this dissertation, an obstacle-avoidance based approach for the control of bacteria powered microrobots (BPMs) using electric fields will be present. A BPM is an integrated cell-based robotic system, each of which consists of a SU-8 microstructure blotted with swarming bacteria. The concept of the BPM is to utilize inorganic structures as platforms to harness the collective propulsive power from the biomolecular motors of bacteria. The suggested obstacle-avoidance method enhances the controllability of the BPMs by allowing them to avoid collision with static obstacles and dynamic obstacles in real time. In case of static obstacle, artificial potential field was used in our approach to generate the objective function regarding the controllability of the BPMs under electric field. On the other hand, the artificial potential field is difficult to be applied to the dynamic obstacle avoidance algorithm because the potential field should be computed repeatedly using the sequence images and the sampling time is too short to finish the computation. Therefore, we used different method for dynamic obstacle avoidance. The C-space is utilized to calculate the probability of collision in static obstacle case. For dynamic obstacles, the redefined VFH function in robotics is added in the obstacle avoidance algorithm in terms of calculating the collision risk and restraint the control inputs which make BPMs head toward the obstacle in a certain range. There are a couple of factors to consider for developing obstacle avoidance algorithm. The first factor is a self-actuation of a BPM that is not controllable due to biomolecular actuator. The second factor is the physical effect of an electric field around the obstacle. The fist constraint can be resolved by the kinematic model of a BPM that helps to predict the position. For the second factor, I use COMSOL Multiphysics engineering simulation software to model an electric field applied across the testbed to characterize distortions of the field around the boundaries of static and dynamic obstacles. We demonstrate the feasibility of our obstacle avoidance algorithm through multiple experiments with different…