Path Planning Algorithms for Automated Agricultural Machines Performing Sequentially Dependent Operations in Arable Farming

• Applicant: Professor Dr.-Ing. Timo Oksanen
• Subject Area: Automation, Mechatronics, Control Systems, Intelligent Technical Systems, Robotics
• Term: since 2023
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 528103308

This project focuses on developing computational coverage path planning algorithms for automated agricultural machines, including autonomous tractors. The main objective of this project is to study novel algorithms for the emerging problem of sequentially dependent agricultural robotic operations. Sequentially dependent means that one operation must be completed before the other. This problem arises especially in arable farming where several coverage operations are performed consecutively on the field during the cultivation cycle. Some of the operations, such as harrowing and seeding, can be performed immediately one after the other. The conventional way in arable farming is to perform only one operation at a time on the entire field. Several studies consider the problem where a single coverage operation such as seeding is divided among multiple machines, and the machines are working simultaneously on the same operation on the same field. A typical solution is to assign each machine their own workspace, allowing them to work independently without risk of conflicts or collisions. However, enabling the machines to perform different operations in parallel can reduce the total completion time of multiple operations, and therefore, improve the work efficiency. The goal of this project is to enable multiple machines to work on different operations simultaneously on essentially the same area of land without waiting for a previous operation to be finished on the entire field. This means that the simple solution of assigning each machine their individual, static workspace is infeasible. A significant part of the challenge is that the machines are heterogeneous, and oftentimes have different working widths. This means that one machine cannot simply follow another. The work in this project consists of deriving a formal definition of the sequentially dependent robotic problem, deriving an algorithmic solution to thereof, and finally, verifying and demonstrating the feasibility of the algorithm in real life with real agricultural machinery. To demonstrate the practical applicability of the results, this project includes empirical tests with real-life agricultural equipment, including tractors and implements for arable farming operations.

• DFG Programme: Research Grants