English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Recent advancements in large language models (LLMs) have spurred interest inusing them for generating robot programs from natural language, with promisinginitial results. We investigate the use of LLMs to generate programs forservice mobile robots leveraging mobility, perception, and human interactionskills, and where accurate sequencing and ordering of actions is crucial forsuccess. We contribute CodeBotler, an open-source robot-agnostic tool toprogram service mobile robots from natural language, and RoboEval, a benchmarkfor evaluating LLMs' capabilities of generating programs to complete servicerobot tasks. CodeBotler performs program generation via few-shot prompting ofLLMs with an embedded domain-specific language (eDSL) in Python, and leveragesskill abstractions to deploy generated programs on any general-purpose mobilerobot. RoboEval evaluates the correctness of generated programs by checkingexecution traces starting with multiple initial states, and checking whetherthe traces satisfy temporal logic properties that encode correctness for eachtask. RoboEval also includes multiple prompts per task to test for therobustness of program generation. We evaluate several popular state-of-the-artLLMs with the RoboEval benchmark, and perform a thorough analysis of the modesof failures, resulting in a taxonomy that highlights common pitfalls of LLMs atgenerating robot programs. We release our code and benchmark athttps://amrl.cs.utexas.edu/codebotler/.

Deploying and Evaluating LLMs to Program Service Mobile Robots