Publication: Autonomous surface vessel for floating rubbish collecting
Loading...
Date
2020-08-01
Authors
Cheong, Chun Keat
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Substantial amount of waste generated by human annually leads to water pollution. This scenario attracts a lot of companies’ attention to develop such cleaning system in massive scale. However, this is not feasible for some consumers as their cleaning workspaces are small. Hence, this project presented autonomous surface vessel that could cruise through the user’s defined waypoints to collect the floating rubbish on the waterways. This autonomous system comprised three subsystems which were sensor’s value filtering to reduce the noise problem in sensor’s measurements, motion controllers to track the yaw angle and surge speed and lastly a guidance system with path interpolation to solve the under actuation presented in the ASV. The sensor’s value filtering system used Madgwick’s filter and loosely coupled GPS/INS integration to obtain reliable yaw angles and the position as well as velocity of the ASV respectively during cruising process. To avoid distortion of magnetometer’s reading, a magnetometer calibration program that can find the non-orthogonal matrix, offset biases, and misalignment matrix between triads of IMU was included. Besides that, PID controllers with back–calculation were designed for surging and yawing motion. Then, the LOS guidance system followed the path generated from straight –line and natural cubic-spline interpolation. Simulation on the ASV had been done
using ROS Gazebo environment. Based on the results, the sensors’ filtering part reported estimated states information with high reliability and the mean errors recorded were
5.480 × 10−6°N, −1.383 × 10−5°E and -0.733 m for the position estimates of north coordinates, east coordinates and altitude, respectively. Meanwhile small lags of velocity estimates were observed with the maximum recorded error of 0.020 𝑚𝑠−1. For the yaw angle tracking, no oscillation was observed, and very fast response time were archived while surging speed tracking showed small oscillation but it stabilized with a mean error of 0.01 𝑚𝑠−1 with a reference velocity of 1.50 𝑚𝑠−1 after travelling for 20 𝑠. Moreover, a very good response of tracking straight – line path was obtained but high oscillations were observed for some portions of the curved path, although it still could track back the path after the oscillation. Lastly, the yaw values determined from magnetometer’s data after calibrating the magnetometer were quite accurate with the maximum absolute mean error of 1.5°.
This project aims to provide guidelines for developing autonomous system of the rubbish cleaning ASV to promote the use of this unmanned system that allows high efficiency of cleaning process.