ProgPrompt: Generating Situated Robot Task Plans using Large Language Models

Task planning can require defining myriad domain knowledge about the world in which a robot needs to act. To ameliorate that effort, large language models (LLMs) can be used to score potential next actions during task planning, and even generate action sequences directly, given an instruction in natural language with no additional domain information. However, such methods either require enumerating all possible next steps for scoring, or generate free-form text that may contain actions not possible on a given robot in its current context. We present a programmatic LLM prompt structure that enables plan generation functional across situated environments, robot capabilities, and tasks. Our key insight is to prompt the LLM with program-like specifications of the available actions and objects in an environment, as well as with example programs that can be executed. We make concrete recommendations about prompt structure and generation constraints through ablation experiments, demonstrate state of the art success rates in VirtualHome household tasks, and deploy our method on a physical robot arm for tabletop tasks. Website at progprompt.github.io

Authors

• Ishika Singh
• Valts Blukis
• Arsalan Mousavian
• Ankit Goyal 0001
• Danfei Xu
• Jonathan Tremblay
• Dieter Fox
• Jesse Thomason
• Animesh Garg

Keywords

• Automation
• Natural languages
• Manipulators
• Planning
• Task analysis
• Language Model
• Task Planning
• Large Language Models
• Natural Language
• Sequence Of Actions
• Ablation Experiments
• Household Tasks
• Physical Robot
• Robot Capabilities
• Plan Generation
• Environmental Conditions
• Chicken
• Programming Language
• Point Cloud
• Task Execution
• Active Recovery
• Goal State
• Action Execution
• State Feedback
• Relevant Tasks
• Virtual Agent
• Static Graph
• List Of Objects
• Object Affordances
• Planning Of Robots
• Execution Plan
• Current State Of The Environment
• API Calls