A simulation platform interaction method and system based on a large language model

By constructing a high-confidence index structure through a simulation platform interaction method based on a large language model, and combining retrieval and generative robot technologies, user intent is parsed and mapped to simulation platform interface parameters. This solves the problem that traditional robots cannot understand user needs, and enables zero-based use and efficient simulation operation.

CN122287342APending Publication Date: 2026-06-26ZHUOYI ZHINENG
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
ZHUOYI ZHINENG
Filing Date
2026-03-31
Publication Date
2026-06-26

AI Technical Summary

Technical Problem

Traditional question-and-answer robots on existing drone simulation platforms cannot deeply understand users' specific flight control parameter configurations, complex environmental variable settings, and personalized mission scenario requirements, resulting in difficulties for users to operate, high learning costs, and an inability to provide precise technical guidance.

Method used

We adopt a simulation platform interaction method based on a large language model. By constructing a high-confidence index structure and combining retrieval-type and generative robot technologies, we optimize the indexing process, suppress the risk of illusion, parse the user's semantic intent, and map it to the standardized interface parameters of the simulation platform to realize interface calls and provide zero-based use.

Benefits of technology

The simulation platform has achieved lightweight operation and control, which lowers the threshold for use, improves application efficiency and R&D output quality, and allows users to quickly start simulation tasks without having to learn a complicated operation interface.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122287342A_ABST
    Figure CN122287342A_ABST
Patent Text Reader

Abstract

This invention belongs to the field of intelligent speech and discloses a simulation platform interaction method and system based on a large language model. The method includes: enhancing retrieval by integrating a large language model; combining the characteristics of retrieval-oriented and generative robot technologies; optimizing a single index into a multi-level retrieval optimization process; generating inference strategies using a high-confidence index structure and embedding them into the large language model; performing generation processing on the retrieval results of the recall and rearrangement processes; using a retrieval enhancement mechanism to suppress the risk of hallucinations; parsing the results of suppressing hallucinations to obtain the user's precise semantic intent; extracting standardized sequences from the simulation platform; a smart agent interface adaptation layer mapping the semantic parsing results and simulation operation sequence instructions to standardized interface parameters for simulation recognition; determining the interface call logic and timing rules; and calling the platform interface based on the standardized interface call scheme to obtain the simulation platform interaction results. This invention lowers the barrier to entry for using simulation platforms through natural language understanding generation.
Need to check novelty before this filing date? Find Prior Art