The fusion of human-centric design and artificial intelligence (AI) capabilities has opened up new possibilities for next-generation autonomous vehicles that go beyond transportation. These vehicles can dynamically interact with passengers and adapt to their preferences. This paper proposes a novel framework that leverages Large Language Models (LLMs) to enhance the decision-making process in autonomous vehicles. By utilizing LLMs' linguistic and contextual understanding abilities with specialized tools, we aim to integrate the language and reasoning capabilities of LLMs into autonomous vehicles. Our research includes experiments in HighwayEnv, a collection of environments for autonomous driving and tactical decisionmaking tasks, to explore LLMs' interpretation, interaction, and reasoning in various scenarios. We also examine realtime personalization, demonstrating how LLMs can influence driving behaviors based on verbal commands. Our empirical results highlight the substantial advantages of utilizing chainof-thought prompting, leading to improved driving decisions, and showing the potential for LLMs to enhance personalized driving experiences through ongoing verbal feedback. The proposed framework aims to transform autonomous vehicle operations, offering personalized support, transparent decisionmaking, and continuous learning to enhance safety and effectiveness. We achieve user-centric, transparent, and adaptive autonomous driving ecosystems supported by the integration of LLMs into autonomous vehicles. The experiment videos are available on https://youtube.com/playlist?list=PLgcRcf9w8BmLJi fqTGq-7KCZsbpEIE4a&si=dhH9lgaeSmB5K94t.

Please send your proposal on profiling research activities of your or other intelligent transportation systems research groups and labs for the “ITS Research Labs” column to Yisheng Lv at yisheng.lv@ ia.ac.cn.

Navigation or route guidance systems are designed to provide drivers with real-time travel information and the associated recommended routes for their trips. Classical route choice models typically rely on utility theory to represent drivers' route choice behavior. Such choices, however, may not be optimal from both the individual and the system perspectives. This is simply due to the fact that drivers usually have imperfect knowledge about the time-varying traffic conditions. In this article, we explore and propose a new model-free deep reinforcement learning (DRL) approach to solving the adaptive route guidance problem based on microsimulation. The proposed approach consists of three interconnected algorithms, including a network edge labeling algorithm, a routing plan identification algorithm, and an adaptive route guidance algorithm. Simulation experiments on both a toy network and a real-world network of Suzhou, China, are performed to demonstrate the effectiveness of the proposed approach in terms of guiding a single vehicle as well as multiple vehicles through complex traffic environments. Comparative results confirm that the DRL approach outperforms the traditional shortest path method by further reducing the average travel time in the network.

Autonomous vehicles (AVs) are expected to reshape future transportation systems, and decision making is one of the critical modules toward high-level automated driving. To overcome those complicated scenarios that rule-based methods could not cope with well, data-driven decision-making approaches have aroused more focus. The datasets to be used in developing data-driven methods dramatically influence the performance of decision making; hence, it is necessary to have a comprehensive insight into the existing datasets. From the aspects of collection sources, driving data can be divided into vehicle-, environment-, and driver-related data. This study compares the state-of-the-art datasets of these three categories and summarizes their features, including sensors used, annotation, and driving scenarios. Based on the characteristics of the datasets, this survey also discusses potential applications of datasets on various aspects of AV decision making, assisting researchers in finding appropriate ones to support their own research. The future trends of AV dataset development are summarized.

Please send your proposal on profiling research activities of your or other intelligent transportation systems research groups and labs for the “ITS Research Labs” column to Yisheng Lv at yisheng.lv@ ia.ac.cn.

Traffic congestion has become a major concern in most cities all over the world. The proper guidance of cars with an effective route planning method has become a fundamental and smart way to alleviate congestion under existing urban road facilities. Current route planning methods mainly focus on a single car, but ignoring the dynamic effect between cars may lead to severe congestion during the actual driving guidance. In this article, we extend the study of route planning to the case of multiple cars and present a novel multicar shortest travel-time routing problem. The objective is to minimize the average travel time by considering the dynamic effect of the induced traffic congestion on travel speed, while ensuring that each car’s travel distance is within an acceptable range. We construct a time-hierarchical graph model for structuring the spatiotemporal dynamic properties of the urban road network and then develop a two-level multicar route planning optimization method for complex problem solving. The experimental results show that our path recommendations reduce the average travel time by 51.74% and 38.87% on average compared to two representative methods. Our research will become more important in the years ahead as self-driving cars become more commonplace.