Economic, biophysical, and behavioral drivers for robotic weed control

The heavy reliance on herbicides for weed control has led to an increase in resistant weeds in the US, which results in over $1 billion annual yield loss. Robotic weed control is emerging as an alternative technology for removing weeds mechanically using artificial intelligence. This dissertation consists of three studies, in which we develop an Integrated Weed Ecological and Economic Dynamic (I-WEED) model using optimal control theory to examine the economic, biophysical, and behavioral drivers for robotic weed control. In the first study of this dissertation, we examine the impacts of initial resistance level, farmer behavior, fitness cost, immigrant seed resistance level, technology attributes, and market conditions on robotic adoption by farmers that differ in the extent to which they are forward looking or myopic about the consequences of current decisions. We find that compared to a myopic farmer, a forward-looking farmer adopts robots earlier, adopts fewer robots, and deploys robots on a smaller portion of the land, and achieves higher profitability, greater resistance control, and lower yield loss in the long run. A forward-looking farmer adopts robotic mechanical weeding as a complement to chemical weeding whereas a myopic farmer adopts it as a substitute. Our sensitivity analyses indicate that a lower fitness cost or a higher immigrant seed resistance level will increase farmer’s incentives for robotic weed control adoption, while the effect of discount rate is ambiguous. In the second study of this dissertation, we extend the I-WEED model to allow farmers to hire robotic weeding services in lieu of owning robots, where the services are more flexible but less certain, and analyze the impact of having more additional adoption methods on farmer’s adoption choice and resulting outcomes. We find that the uncertainty of robotic services timeliness and the robotic service fee rates are the main factors that affect farmer’s decision of whether to own robots or to hire services. Compared to a myopic farmer, a forward-looking farmer adopts robots earlier, but is less likely to hire services. By having the option to choose between owning robots or hiring services, both types of farmers have a better outcome in profitability, resistance management, and yield loss control than if they are restricted to only owning or only hiring. Though farmers have the flexibility to choose, only a forward-looking farmer might combine owning and hiring in the long run and obtain the highest total profit. In the third study of this dissertation, we develop a two-farmer I-WEED model to examine the mutual impact of a farmer’s weeding behavior on their neighboring farmer’s adoption choice under different levels of distance between the two farms and different biophysical conditions in each farm. We simulate the weed management strategies to reveal under what conditions the cooperation can occur and quantify the transfer payment associated with the cooperation. We compare individually optimal weeding strategies with socially optimal weeding strategies and their corresponding economic outcomes. We find that the distance between two farms increases, the effects of the neighbor’s weeding behavior decline, and a farmer’s weeding decision is more independent. By internalizing the seed dispersion over space, a social planner would have an earlier robotic weeding adoption than a private farmer, which suggests the importance of farmer education and training programs that can help farmers internalize the negative impact of chemical weeding over time and space.

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Agricultural & Applied Econ

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2023

Author dc:creator

• Yu, Chengzheng

Contributors dc:contributor

• Khanna, Madhu
• Atallah, Shady S
• Hutchins, Jared
• Bagavathiannan, Muthukumar

• Robotic
• Weed Management
• Technology Adoption
• Bioeconomic

Statement dc:rights

• Copyright 2023 Chengzheng Yu

Language dc:language

en, eng

Handle dc:identifier

https://hdl.handle.net/2142/121223

OAI record oai:identifier

https://www.ideals.illinois.edu/oai-pmh?verb=GetRecord&metadataPrefix=etdms&identifier=oai%3Awww.ideals.illinois.edu%3A2142%2F121223