Optimization Shortest One-Way Path for Energy Saving Auto Robot Collecting Floating Garbage using Fast Approximate Nearest Neighbor Search

Abstract

Floating garbage is a global problem that needs to be cleared and disposed of from the sea and rivers or canals. Mostly the floating plastic garbage directly affects the ecology and the aquatic animals. This paper proposes to develop a low-cost auto robot from waste materials in the household, school, or agriculture, then combines with the Internet of Things technology and focusing on the power energy saving with the highest accuracy in operation of floating garbage collection. ESP32 development kit board is the primary controller with a camera module and ultrasonic sensor for detecting the floating garbage while moving on the water’s surface in the boundary of the specified area, including the small canal, rectangle area, and no rectangle area. The mixed mode is a new method to define the shortest one-way path using a fast approximate nearest neighbor search for the auto robot to detect and collect the floating garbage on the water surface. It was combined between the stationary and the cover model. The mixed mode can eliminate the blind spot area that occurs when using the stationary mode and cover mode. As a result, the auto robot system’s accuracy in mixed mode is 98.3%. It was found that the development of the shortest one-way path for the auto robot collecting floating garbage by mixed mode provides the highest efficiency in detecting floating garbage and also helps to save overall costs, and the resource consumption is nearly half of cover mode consumption.