As You Wish: Mission Planning with Formal Verification using LLMs in Precision Agriculture
Marcos Abel Zuzu\'arregui, Stefano Carpin
Why It Matters
What makes this one worth your time
Improving the reliability of mission planning in precision agriculture can lead to more effective automation and better resource management in farming practices.
This work enhances LLM-based mission planning in agriculture by integrating formal verification techniques.
Summary
The paper extends a mission planning system that uses large language models (LLMs) for precision agriculture by incorporating feedback loops and linear temporal logic (LTL) to improve the clarity and reliability of mission specifications derived from natural language.
Key contributions
- Introduction of feedback loops in mission planning architecture.
- Application of linear temporal logic (LTL) for formal verification of mission plans.
- Evaluation of the integration of LLMs in autonomous mission planning systems.
Notable insights
- The use of multiple LLMs for specification and verification tasks may help mitigate biases inherent in single models.
- Incorporating LTL into the planning process could provide a structured way to handle ambiguities in natural language.
Possible limitations
- The abstract does not specify the experimental setup or the metrics used for evaluation.
- Potential challenges in generating valuable LTL formulas from LLM outputs are mentioned but not deeply explored.
Abstract
arXiv:2606.18519v1 Announce Type: cross Abstract: Though robotic systems are now being commercialized and deployed in various industries, many of these systems are highly specialized and often require an advanced skill set to operate and ensure they perform as instructed. To mitigate this problem, we recently introduced a mission planner leveraging LLMs to synthesize mission plans in precision agriculture based on mission descriptions provided in natural language. While the system demonstrates impressive performance, it also suffers from the inherent ambiguities of natural language. In this paper, we extend our system to address this issue by introducing multiple feedback loops in the planning architecture that leverage linear temporal logic (LTL) to ensure the mission planning system meets the specifications formulated by the user while still using natural language. To mitigate potential bias, this is achieved by using two different commercial LLMs in charge of the specification and verification subtasks. Through extensive experiments, we highlight the strengths and limitations of integrating mission verification into a fully autonomous pipeline, particularly regarding an LLM's ability to generate valuable LTL formulas, and show how our proposed implementation addresses and solves these challenges.