Duration: Jun 2013 - Sep 2017
Awards: 2
Publications: 8
Funding: MYR 137,900

The Robot Operating System (ROS) for open-field Agriculture

Major agricultural operations are often performed by tractors, harvesters, ploughers, and other vehicular machinery. Automating these tasks using the Robot Operating System (ROS) requires the successful implementation of ROS for outdoor mobile robots. This is the focus of this research.??



The Main Challenge?

ROS is an open-source, community-driven Platform for Robot Software development, its community is made of developers from all around the world, which include hobbyists, students, academics, as well as professional practitioners. ROS-industrial is a sub-community of ROS, as the name implies, focusing entirely on Industrial Robots and their R&D. It is also supported by a conglomerate of companies that specialize in the field of robot design and development.

In most current ROS applications, robots implementing ROS are mostly indoor robots, such as Social robots, Industrial robots, humanoids, and others. Therefore, their system architecture and implementation are designed for a local network and close proximity between robots and human users or operators.

Implementing ROS for outdoor applications, such as agriculture, requires the robots to cover vast distances in open fields, working independently but also in collaboration with human workers or operators.?


These requirements present a number of system challenges that were not addressed by the ROS community because of the focus on indoor applications.?In this work, my PhD research, I tackled those challenges and proposed a number of procedures and protocols to overcome them. Thus, allowing for the successful implementation of ROS in outdoor applications (agriculture and others), and so expanding ROS's benefits.

The Proposed Solution

The first step of the implemented solution involves hardware abstraction of agriculture machinery into the ROS framework; that means implementing a number of protocols to enable allow these machines to become ROS-ready, and so benefit from its features.?

The Second step involves developing the remote ROS network, which is an expansion of the indoor ROS network. Thus, allowing for remote robot operations and remote interactions between robots operating in vast open fields and human operators. This was achieved through PortForwarding.?

The third and final step is to develop an effective Human-Robot-Interaction Framework to facilitate the interaction between human workers and these ROS-ready robots while on the field. This was achieved by combining ROS with the android framework.?

Research Output

This work took 6 years to complete; I completed my PhD on a part-time basis while I worked as a full-time lecturer. Nevertheless, Outcome of this work includes 5 papers, 1 patent (filed), and innovation awards. Furthermore, part of this work (remote ROS networking) has been incorporated into ROS as part of its learning and training material.?


Comments

Leave a comment