Exploring GPT-4 for Robotic Agent Strategy with Real-Time State Feedback and a Reactive Behaviour Framework
This work addresses practical deployment issues for LLM-driven robotic agents, but it is incremental as it builds on existing methods for executability and correctness.
The paper tackled the problem of using GPT-4 to control a humanoid robot by addressing practical concerns like safety and task transitions, resulting in a method that produces executable outputs for feasible requests with smooth transitions and achieves user goals most of the time across various time horizons.
We explore the use of GPT-4 on a humanoid robot in simulation and the real world as proof of concept of a novel large language model (LLM) driven behaviour method. LLMs have shown the ability to perform various tasks, including robotic agent behaviour. The problem involves prompting the LLM with a goal, and the LLM outputs the sub-tasks to complete to achieve that goal. Previous works focus on the executability and correctness of the LLM's generated tasks. We propose a method that successfully addresses practical concerns around safety, transitions between tasks, time horizons of tasks and state feedback. In our experiments we have found that our approach produces output for feasible requests that can be executed every time, with smooth transitions. User requests are achieved most of the time across a range of goal time horizons.