Componentizing Autonomous Underwater Vehicles by Physical-runnning Algorithms

The research experiment outlined on this page presents an architecture designed for componentizing autonomous underwater vehicles, utilizing direct physical-runnning algorithms. Specifically, a genetic algorithm (GA) operates directly on a physical component to identify its notable operational points. The associated diagram delineates the interrelation of seven key elements in accordance with this architectural framework, facilitating evaluation in a cyber-physical loop. These elements include a computer, a controller, two sensors (for power consumption and thrust acquisition), a motor driver, a power supply, and a DC motor. For this experiment, the chosen controller was the Atmel SAMD21 board (Arduino MKR1000). This controller is adept at transmitting and receiving multiple data streams over serial communication interfaced with either Matlab or R software. This software is hosted on the computer where the genetic algorithm operates. The employed sensors comprised a strain gauge, utilized for assessing the thrust generated by the DC motor, and a current sensor, used for power consumption measurement. The combined function of the power supply and motor driver is to energize the motor. Using a PWM signal, the GA has the capability to adjust the motor's rotational intensity.