Abstract:Large Language Models (LLMs), capable of handling multi-modal input and outputs such as text, voice, images, and video, are transforming the way we process information. Beyond just generating textual responses to prompts, they can integrate with different software platforms to offer comprehensive solutions across diverse applications. In this paper, we present ChatSUMO, a LLM-based agent that integrates language processing skills to generate abstract and real-world simulation scenarios in the widely-used traffic simulator - Simulation of Urban MObility (SUMO). Our methodology begins by leveraging the LLM for user input which converts to relevant keywords needed to run python scripts. These scripts are designed to convert specified regions into coordinates, fetch data from OpenStreetMap, transform it into a road network, and subsequently run SUMO simulations with the designated traffic conditions. The outputs of the simulations are then interpreted by the LLM resulting in informative comparisons and summaries. Users can continue the interaction and generate a variety of customized scenarios without prior traffic simulation expertise. For simulation generation, we created a real-world simulation for the city of Albany with an accuracy of 96\%. ChatSUMO also realizes the customizing of edge edit, traffic light optimization, and vehicle edit by users effectively.
What problem does this paper attempt to address?
The paper attempts to address the issues of complexity, time consumption, and the need for specialized traffic knowledge in the process of generating and modifying traffic simulations in modern transportation systems. Specifically:
1. **Complexity and Time Consumption of Traffic Simulation Generation**: Traditional traffic simulation software like SUMO requires users to manually define parameters such as networks, vehicles, and routes. This not only increases the difficulty of use, especially for beginners, but also consumes a significant amount of time.
2. **Lack of Intuitive Human-Computer Interaction**: Existing traffic simulation tools typically require users to have specialized traffic knowledge, which limits the effective use by non-professionals. The emergence of large language models (LLMs) offers the possibility to simplify this process through natural language processing technology, enabling more intuitive human-computer interaction.
3. **Improving the Efficiency of Traffic Planning and Management**: By automating and intelligentizing the generation and modification of traffic simulations, the efficiency of traffic planning and management can be significantly improved, reducing the cost of actual traffic planning implementation and helping to alleviate traffic congestion and environmental pollution.
To this end, the paper proposes ChatSUMO, an intelligent assistant based on large language models (LLM), aimed at simplifying the process of generating and modifying SUMO traffic simulations. The main contributions of ChatSUMO include:
- **Converting Text Descriptions into SUMO Simulation Scenarios**: Users can generate complex traffic simulation scenarios through simple text input without needing specialized traffic simulation knowledge.
- **Supporting Various Custom Modifications**: Users can modify simulation scenarios through text commands, including road editing, traffic signal optimization, vehicle adjustments, and more.
- **Providing Real-Time Analysis and Dynamic Adjustments**: ChatSUMO can analyze simulation results in real-time and make dynamic adjustments based on user needs, providing detailed reports and recommendations.
Through these features, ChatSUMO significantly lowers the barrier to traffic simulation, allowing users of different levels to easily create and modify traffic simulation scenarios, thereby improving the efficiency of traffic planning and management.