EduSim-LLM: An Educational Platform Integrating Large Language Models and Robotic Simulation for Beginners

TL;DR

EduSim-LLM tackles the challenge of teaching beginners to control robots through natural language by integrating large language models with the CoppeliaSim simulator. The authors design a four-module pipeline—natural language interface, LLM-based instruction planner, simulation control backend, and user frontend—enabling zero-coding, language-driven control across single and multi-robot tasks. Through 108 instruction cases spanning simple to complex tasks, the system achieves 100% success for simple, 94.4% for composite, and 88.9% for complex instructions, and demonstrates significant reductions in human operation time compared to manual control. This work evidence the feasibility and educational value of NL-based robotics control in simulations, potentially lowering the barrier to entry for learners and instructors in robotics.

Abstract

In recent years, the rapid development of Large Language Models (LLMs) has significantly enhanced natural language understanding and human-computer interaction, creating new opportunities in the field of robotics. However, the integration of natural language understanding into robotic control is an important challenge in the rapid development of human-robot interaction and intelligent automation industries. This challenge hinders intuitive human control over complex robotic systems, limiting their educational and practical accessibility. To address this, we present the EduSim-LLM, an educational platform that integrates LLMs with robot simulation and constructs a language-drive control model that translates natural language instructions into executable robot behavior sequences in CoppeliaSim. We design two human-robot interaction models: direct control and autonomous control, conduct systematic simulations based on multiple language models, and evaluate multi-robot collaboration, motion planning, and manipulation capabilities. Experiential results show that LLMs can reliably convert natural language into structured robot actions; after applying prompt-engineering templates instruction-parsing accuracy improves significantly; as task complexity increases, overall accuracy rate exceeds 88.9% in the highest complexity tests.

Figures (10)

• Figure 1: System architecture of EduSim-LLM integrating LLMs and CoppeliaSim
• Figure 2: User interface of EduSim-LLM
• Figure 3: The Youbot1 robot first takes a picture at its initial position, then moves towards the negative y-axis, gets close to the target object, takes another picture, continues moving towards the target object, then moves laterally towards the negative x-axis, takes a picture before reaching the object to be grasped, then moves laterally towards the positive x-axis, takes a picture before reaching the obstacle to be removed, and finally moves back to the starting point using a combination of x and y axis movements.
• Figure 4: The Youbot2 robot starts from its initial position and moves in the negative y-axis direction to the obstacle to be removed. It then performs a combined x and y-axis movement to collect the obstacle into the fixed frame at the front of the robot and moves it to the designated parking area. Finally, it moves laterally in the positive x-axis direction into the parking area to complete the removal task.
• Figure 5: The Youbot3 robot starts from its initial position and moves in the negative y-axis direction to the object to be gripped. It then uses its robotic arm to grip and retrieve the object. After retrieving the object, it moves in the positive y-axis direction to the designated parking area, and finally moves in the negative x-axis direction into the parking area to complete the gripping task.