A Predictive Control Strategy to Offset-Point Tracking for Agricultural Mobile Robots
Stephane Ngnepiepaye Wembe 1,2 {}^{\textbf{1,2}}, Vincent Rousseau 1 {}^{\textbf{1}}, Johann Laconte 1 {}^{\textbf{1}} and Roland Lenain 1 {}^{\textbf{1}} · 2026
Read A Predictive Control Strategy to Offset-Point Tracking for Agricultural Mobile Robots by Stephane Ngnepiepaye Wembe 1,2 {}^{\textbf{1,2}}, Vincent Rousseau 1 {}^{\textbf{1}}, Johann Laconte 1 {}^{\textbf{1}} and Roland Lenain 1 {}^{\textbf{1}} (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-28439.txt.
Introduction
Wembe and three co-authors design a controller that steers farm robots by their tool, not just their center. A weeder or cultivator rides at an offset, and ignoring that offset can crush crops. Their closed-form predictive method models the tool as a rigid offset point, with sideways slip and lever-arm effects included. In field tests with two tools, median tracking error fell by 24 to 56 percent. Peak errors at curve changes dropped by up to 70 percent.
Catalog record
| CREATOR | Stephane Ngnepiepaye Wembe 1,2 {}^{\textbf{1,2}}, Vincent Rousseau 1 {}^{\textbf{1}}, Johann Laconte 1 {}^{\textbf{1}} and Roland Lenain 1 {}^{\textbf{1}} |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 93483 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.28439
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
A Predictive Control Strategy to Offset-Point Tracking for Agricultural Mobile Robots Stephane Ngnepiepaye Wembe 1,2 {}^{\textbf{1,2}}, Vincent Rousseau 1 {}^{\textbf{1}}, Johann Laconte 1 {}^{\textbf{1}} and Roland Lenain 1 {}^{\textbf{1}} Abstract Robots are increasingly being deployed in agriculture to support sustainable practices and improve productivity. They offer strong potential to enable precise, efficient, and environmentally friendly operations. However, most existing path-following controllers focus solely on the robot’s center of motion and neglect the spatial footprint and dynamics of attached implements. In practice, implements such as mechanical weeders or spring-tine cultivators are often large, rigidly mounted, and directly interacting with crops and soil; ignoring their position can degrade tracking performance and increase the risk of crop damage. To address this limitation, we propose a closed-form predictive control strategy extending the approach introduced in [ 25 ] . The method is developed specifically for Ackermann-type agricultural vehicles and explicitly models the implement as a rigid offset point, while accounting for lateral slip and lever-arm effects. The approach is benchmarked against state-of-the-art baseline controllers, including a reactive geometric method, a reactive backstepping method, and a model-based predictive scheme. Real-world agricultural experiments with two different implements show that the proposed method reduces the median tracking error by 24 % 24\text{\,}\% to 56 % 56\text{\,}\% , and decreases peak errors during curvature transitions by up to 70 % 70\text{\,}\% . These improvements translate into enhanced operational safety, particularly in scenarios where the implement operates in close proximity to crop rows. {IEEEkeywords} Field Robotics, Agricultural Robotics, Autonomous Navigation, Nonholonomic Robots, Trajectory Tracking, Predictive Control, Agricultural Implement Control † † corresponding: Corresponding author: Stephane Ngnepiepaye Wembe (email: [email protected]) † † note: This work has been funded by the french National Research Agency (ANR), under the grant ANR-19-LCV2-0011, attributed to the joint laboratory Tiara ( www6.inrae.fr/tiara ). It has also received the support of the French government research program ”Investissements d’Avenir” through the IDEX-ISITE initiative 16-IDEX-0001 (CAP 20-25), the IMobS3 Laboratory of Excellence (ANR-10-LABX-16-01). This work has been partially supported by ROBOTEX 2.0 (Grants ROBOTEX ANR-10-EQPX-44-01 and TIRREX ANR-21-ESRE-0015) 1 INTRODUCTION \IEEEPARstart Agriculture as a whole faces a multitude of challenges that arise, directly or indirectly, from the realities of the modern world, such as global population growth or pollution. Among the most pressing issues are environmental constraints, such as soil preservation, and the imperative to ensure sufficient production to meet the ever-increasing global food demand. This last challenge is underscored by recent data from the FAO ( FAO ), which report that approximately 733 million people suffered from hunger in 2023, representing one in eleven individuals globally [ 8 ] . In response to this challenge, two main approaches have emerged in agriculture. One approach prioritizes environmentally friendly farming practices, such as conservation agriculture. By promoting sustainable soil management and reducing chemical inputs, it helps limit ecological degradation. However, this method often results in relatively modest yields that fall short of meeting the growing global food demand. In contrast, a second approach, widely adopted worldwide, relies on intensive agriculture. It focuses on maximizing production through the extensive use of pesticides, chemical fertilizers, and other potentially harmful substances. While this strategy enables large-scale output, it also causes significant environmental harm, threatening the long-term sustainability of natural resources. Reconciling these two fundamental objectives, i.e., high productivity and environmental preservation, has long posed a challenge. This involves strategic trade-offs, forcing stakeholders to choose between prioritizing agricultural output and safeguarding ecosystems. To address this dual challenge of ensuring consistent agricultural production while respecting the environment, a new approach has emerged: agroecology. Today, this vision has become the guiding principle of sustainable agriculture. Agroecology seeks to achieve sufficient production levels while incorporating practices that respect ecological constraints, such as soil preservation and minimizing environmental impact [ 40 ] . In this context, Agriculture 4.0 plays a pivotal role, integrating advanced technologies such as robotics and automated equipment to support sustainable and environmentally friendly farming [ 34 ] . Agricultural robotics, in particular, has emerged as a promising technological solution [ 19 ] . It aims to transform the agricultural sector by automating and optimizing key tasks, thereby addressing pressing challenges such as labor shortages and the demand for more sustainable agricultural practices [ 21 ] . Robotic systems are increasingly becoming capable of performing repetitive operations, such as planting, weeding, spraying, and harvesting, with high precision. These innovations contribute to the reduction of pesticide use, thereby mitigating environmental impact. Moreover, they alleviate the physical burden on workers, supporting the transition toward more agroecological and sustainable farming systems. The growing interest in agricultural robotics is reflected in a multitude of research projects aimed at developing high-performance robotic systems for field operations. These projects aim to improve both productivity and sustainability in the agricultural sector [ 4 ] . To perform an agricultural task, such as weeding, a robot must be equipped with several essential functions [ 24 ] . In agricultural applications, the tasks assigned to robots, whether sowing, weeding, or harvesting, share a fundamental objective: precise path-following. This capability is critical to ensure effective interventions. During an agricultural operation, the implement attached to the robot must often interact directly with crops. In weeding operations, the end effector, such as a blade or plowshare, must follow a precise path that allows it to remove weeds while preserving crops. Figure 1 depicts a robot performing an inter-row weeding operation. In this scenario, the effector elements are the plowshares, which carry out the weeding task. These plowshares follow a designated path, allowing them to operate near the crop rows. However, if the path-following algorithm is ineffective, high magnitude oscillations may occur, potentially causing the effector elements of the agricultural implement to damage the crops. An optimized path-following controller ensures not only the efficiency of the operation but also minimizes the risk of damage to crops. Figure 1 : An autonomous agricultural system, consisting of a robot linked to a mounted implement. In this configuration, the control accuracy is required at the implement-soil interaction point rather than at the center of the robot, which renders center-based control laws inadequate. Traditional control strategies typically focus on the rear axle center of the robot. This choice has the advantage of allowing linearizable modeling using exact linearization techniques [ 33 ] . However, as illustrated in Figure 1 , agricultural applications introduce a specific constraint: it is the implement, not the robot’s center, that directly interacts with the soil and crops. As a result, the control algorithm must be specifically designed to account for the position and actions of the agricultural implement. This necessitates shifting the focus of control from the rear axle center to the agricultural implement itself to ensure maximum precision in performing agricultural tasks, such as weeding or sowing, where direct contact with crops is critical. Controlling the agricultural implement directly, rather than the rear axle center, involves managing an offset point of the system while maintaining the vehicle’s overall kinematics. As such, this requires the development of specific control laws that enable the offset point to follow a predefined path accurately. In the literature, the issue of controlling an offset point relative to the robot is related to the control of trailer systems. However, trailer systems