Presentation Title

Can​ ​Robots​ ​Learn​ ​What​ ​the​ ​Shortest​ ​Path​ ​Is?

Faculty Mentor

Nolan Tsuchiya

Start Date

18-11-2017 1:30 PM

End Date

18-11-2017 1:45 PM

Location

9-245

Session

Engineering/CS 2

Type of Presentation

Oral Talk

Subject Area

engineering_computer_science

Abstract

The application of robotics in businesses has been popular as a way of increasing productivity. With this in mind, the need for a system of efficient navigation is ever more necessary. This research intends to contribute in the area of robotics to improve its intelligence. Specifically, this project intends to answer the research questions: "whether or not a robot can learn the shortest path between two points on a grid of diverging and converging lines" Here, the grid would resemble a maze in appearance with its lines forming right angles. In order to build the robot, the researcher utilized the Arduino platform and built the robot from scratch. The research utilizes the algorithm that either the left or right hand rule will ultimately lead to the chosen destination. Preliminary results indicate that using a set of algorithms for encountered scenarios, the robot can learn to shorten redundant actions. Results of this research can be applied to many industries or businesses including manufacturing and retail wherein an item needs to be delivered from one location to another separated by some obstacles.

Summary of research results to be presented

To address the research question of “whether or not a robot can learn the shortest path between two points on a grid of diverging and converging lines, first, I built a robot (see the attached or hyperlinked two photos of the top and side view of the robot, respectively) that is powered by the Arduino platform, an open-source electronic prototyping platform. Four light sensors were placed collinearly parallel to the axis of the front wheels to give the robot an expanded view of the environment. The fact that this robot has multiple sensors should shorten the travel time. With the robot building process completed, the next step was to implement the algorithm that either the left or right hand rule will ultimately lead to the chosen destination. This algorithm should enhance the robot’s navigation intelligence. For instance, the robot was able to replace three right turns with a left turn until it developed shortened directions. This preliminary result suggests that the algorithm appear to perform as expected and that robots can learn the shortest path between two points on a grid of diverging and converging lines.

This document is currently not available here.

Share

COinS
 
Nov 18th, 1:30 PM Nov 18th, 1:45 PM

Can​ ​Robots​ ​Learn​ ​What​ ​the​ ​Shortest​ ​Path​ ​Is?

9-245

The application of robotics in businesses has been popular as a way of increasing productivity. With this in mind, the need for a system of efficient navigation is ever more necessary. This research intends to contribute in the area of robotics to improve its intelligence. Specifically, this project intends to answer the research questions: "whether or not a robot can learn the shortest path between two points on a grid of diverging and converging lines" Here, the grid would resemble a maze in appearance with its lines forming right angles. In order to build the robot, the researcher utilized the Arduino platform and built the robot from scratch. The research utilizes the algorithm that either the left or right hand rule will ultimately lead to the chosen destination. Preliminary results indicate that using a set of algorithms for encountered scenarios, the robot can learn to shorten redundant actions. Results of this research can be applied to many industries or businesses including manufacturing and retail wherein an item needs to be delivered from one location to another separated by some obstacles.