Editorial: Agricultural robotics

Editorial: Agricultural Robotics • • • • • • • • • • • • • • • • •

The mechanization of agriculture over the past century has resulted in significant increases in productivity and convenience that allow a small percentage of farmers to produce the majority of the food required to sustain mankind. Mobile machinery, introduced in the early 1900s, was a breakthrough innovation that afforded much of that evolution. This has been one of the main factors that allowed food production to keep in pace with the fast-growing population of the world. In December 2008, the U.S. Census Bureau projected that the world’s population would grow from its current number of 6 billion to 9 billion people in 2040, despite the talk of falling birth rates in developed nations. Extrapolating the figures linearly, it is roughly estimated that a 50% increase in food production will be required to provide for the world’s population. Recent advances in automation have provided increases in productivity over the past 10 years through precision guidance and management of machine functions. These advances provide part of the driving force leading to field robotics as an enabling technology for food production. While demand increases, food production faces the challenge of reduced resources due to the flow of manpower to urban areas and the expansion of the boundaries of urban areas at the expense of agricultural land. The trend is for agricultural industries to move farther away from urban centers, to areas with less manpower to attend to the increasingly demanding tasks of feeding the growing population. With the intention of achieving integrated autonomous agricultural systems, agricultural robotics provides not only the means to carry out the required agricultural tasks with less manpower, but also the

means to improve food quality and to exert more control over crops by reducing dependence on the elements. The agricultural application domain provides a fascinating challenge to robotic capabilities. Agricultural applications often involve a semistructured environment that can be partially tailored to simplify the robotics systems requirements. This is in contrast to the fully structured factory automation robotics at one end of the spectrum and the completely unstructured human environment at the other end. As such, agricultural tasks may create unique applicationspecific challenges. Ultimately, it is crucial to understand the appropriate role of field robotics as part of a worksite solution. The focus of this enabling technology must always be to drive increases in productivity or convenience. Current trends display incremental increases in productivity and/or convenience through the use of robotics technologies (e.g., navigation control, sensing, and perception) to augment human-operated machines. However, future systems will also need to consider opportunities for new machine forms and worksite configurations specifically developed for autonomous food production systems. The technological challenges include the following: •

•

Precision control of machinery over terrain that has significant changes in surface characteristics over the annual production cycle. Determining cost-effective navigation solutions for diverse environmental conditions, which vary from open fields to tree canopies.

Journal of Field Robotics 26(6–7), 501–503 (2009) C 2009 Wiley Periodicals, Inc. Published online in Wiley InterScience (www.interscience.wiley.com). • DOI: 10.1002/rob.20302

•

502 • • •

Journal of Field Robotics—2009

Robust and cost-effective machine perception technologies to provide situation awareness that enables safe operation. Autonomous decision-support systems that implement in-field decision management that the farmer-in-the-loop provides. Precision control of tasks, especially in the replacement of human dexterity in selective management of crops.

In this special issue of the Journal of Field Robotics, six papers have been selected to represent the current trends and advances in agricultural robotics and automation. These papers address several of the challenges mentioned above. Two papers are focused on the modeling and control of autonomous farm vehicles, which impacts precision control across the terrain. Three papers provide concepts of robotic worksite solutions on unique agricultural applications. One paper addresses the sensing and perception (vision) challenges in an agricultural task. The editors note that this still leaves significant issues and challenges that should inspire the field robotics community to explore opportunities that will enable new breakthrough innovations that impact the well-being of mankind. The first paper, by Cariou et al., utilizes the fourwheel structure of a tractor to estimate and compensate for slippage between the tractor and the ground. With a kinematic model derived for the tractor, an observer was designed to estimate the amount of sliding, and a backstepping controller was utilized. A predictive algorithm is developed to address the delays induced by the steering actuators, compensating for transient overshoots on curves. The proposed algorithm was implemented on two-wheeled and four-wheeled mobile robots, and experimental results were collected in the field, compared, and discussed. In the second paper, by Derrick and Bevly, MRAC adaptive control was utilized to control the yaw rate of farming tractors. The proposed algorithm aims to compensate for the variations in farm vehicle dynamics through adaptive control strategies. Critical analysis of simulation results is presented along with the results of field implementation on a tractor. The first of the application papers is presented by Nagasaka et al. The paper explores the use of an autonomous rice transplanting tractor in a rice paddy. In this application, the tractor is required to run multiple

paths along the rice paddy and maintain its heading in such a way that each pass along the rice paddy is as much as possible parallel to the other passes. This will ensure that the rice seedlings are planted in rows that do not cross each other and that the seedlings are spread evenly across the field. The paper reports on the hardware and software implementation of an industrial rice transplanter, and the experimental results in a rice paddy are presented. The paper by Johnson et al. demonstrates how robotics becomes an enabler for increasing worksite system productivity. The paper reports on a system of multiple tractors developed to automate a peat moss harvesting operation, with a human operator placed in charge of the tractor team. Each tractor is capable of lower level intelligence, while the human operator monitors and decides on the operational procedures. This again reduces the manpower requirement of having an experienced driver for each tractor in an application limited by labor availability. Shapiro et al. report the outcomes of an experimental project to automate the task of spraying palm trees. Termed “elevated agrobotics” by the authors, the task addresses not only the need for skilled manpower in this labor-intensive and expensive process, but also the safety of human labor operating at such heights. A scaled-down prototype of a visually guided robotic mechanism equipped with a long reaching arm to perform the spraying process was fabricated and tested for its effectiveness. Simulated and field experimental results and possible implementation issues are reported. The last paper, by Jin and Tang, reports on a stereo vision algorithm used in corn planting. The main problem is in the identification process to separate individual corn plants within overlapped plant canopies. Stereo vision was chosen to provide the additional information of depth, allowing the algorithm to construct the plant skeleton structures and hence identifying individual plants. The six papers in this special issue were reviewed not only for their quality, but also for their suitability in presenting the state of the art of today’s agricultural robotics capabilities, as well as in providing field-tested results with insight into practical problems of agricultural robotics. The editors of this special issue would like to note the excellent response to the call for papers, indicating a growing research interest in this important area of field robotics. We would like to thank all the reviewers who contributed their time and effort to the special issue Journal of Field Robotics DOI 10.1002/rob

Editorial

for their evaluations and constructive suggestions, which are instrumental in setting the high standard of our research work in this field, as well as the quality of this technical journal. We also would like to express our appreciation to the authors of the papers,

Journal of Field Robotics DOI 10.1002/rob

•

503

who worked so hard to present us with their very best results in this exciting field of study. Denny Oetomo, John Billingsley, and John F. Reid Guest Editors