TeLL-Drive.github.io/index.html at main · PerfectXu88/TeLL-Drive.github.io · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
<!DOCTYPE html>
<html lang="en">
<head>
    <meta charset="UTF-8">
    <meta name="viewport" content="width=device-width, initial-scale=1.0">
    <title>TeLL-Drive: Enhancing Autonomous Driving</title>
    <link href="https://fonts.googleapis.com/css2?family=Roboto:wght@400;700&display=swap" rel="stylesheet">
    <style>
        body {
            font-family: 'Roboto', sans-serif;
            margin: 0;
            padding: 0;
            background-color: #f7f8fc;
            color: #333;
        }
        header {
            background: linear-gradient(135deg, #3a3f44, #009688);
            color: #fff;
            padding: 4rem 4rem;
            text-align: center;
            box-shadow: 0 2px 4px rgba(0, 0, 0, 0.2);
        }
        header h1 {
            margin: 1.5rem 0;
            font-weight: 700;
            font-size: 2.6rem;
            line-height: 1.2;
        }
        .github-button, .paper-button {
            display: inline-block;
            background-color: #333;
            color: #fff;
            padding: 0.8rem 1.5rem;
            border-radius: 5px;
            text-decoration: none;
            font-weight: 700;
            transition: background-color 0.3s ease;
            margin-top: 2rem;
            margin-right: 1rem;
        }
        .github-button:hover, .paper-button:hover {
            background-color: #555;
        }
        .container {
            max-width: 1000px;
            margin: 2rem auto;
            padding: 2rem;
            background-color: #fff;
            border-radius: 8px;
            box-shadow: 0 4px 8px rgba(0, 0, 0, 0.1);
        }
        .section {
            margin-bottom: 2rem;
        }
        .section h2 {
            color: #3a3f44;
            text-align: center;
            font-size: 2.4rem;
            margin-bottom: 1rem;
        }
        .section p {
            line-height: 1.8;
            font-size: 1.1rem;
            color: #555;
        }
        .authors {
            display: flex;
            justify-content: center;
            gap: 1.5rem;
            flex-wrap: wrap;
            margin-bottom: 1.5rem;
        }
        .authors a {
            color: #3a3f44;
            text-decoration: none;
            font-weight: 700;
            transition: color 0.3s ease;
        }
        .authors a:hover {
            color: #0056b3;
        }
        .images img {
            max-width: 100%;
            border-radius: 8px;
            margin-bottom: 1.5rem;
            box-shadow: 0 2px 4px rgba(0, 0, 0, 0.1);
        }
        .images h3 {
            font-size: 1.1rem;
            color: #555;
            text-align: left;
            margin-top: 0.5rem;
        }
        video {
            width: 100%;
            height: auto;
            border-radius: 8px;
            margin-bottom: 1.5rem;
            box-shadow: 0 2px 4px rgba(0, 0, 0, 0.1);
        }
        .video h3, .video h4 {
            color: #555;
            font-size: 1.2rem;
            text-align: center;
        }
    </style>
</head>
<body>
    <header>
        <h1>TeLL-Drive: Enhancing Autonomous Driving with</h1>
        <h1>Teacher LLM-Guided Deep Reinforcement Learning</h1>
        <div>
            <a href="https://github.com/PerfectXu88/TeLL-Drive.github.io" class="github-button" target="_blank">View on GitHub</a>
            <a href="https://arxiv.org/pdf/2502.01387" class="paper-button" target="_blank">Read the Paper on Arxiv</a>
        </div>
    </header>
    <div class="container">
        <div class="section info">
            <div class="authors">
                <a href="https://perfectxu88.github.io/" target="_blank">Chengkai Xu</a>
                <a href="https://jiaqiliu-aca.netlify.app/" target="_blank">Jiaqi Liu</a>
                <a href="https://fangshiyuu.github.io/" target="_blank">Shiyu Fang</a>
                <a href="https://tops.tongji.edu.cn/info/1131/1818.htm" target="_blank">Yiming Cui</a>
                <a href="https://tops.tongji.edu.cn/info/1031/1383.htm" target="_blank">Peng Hang</a>
                <a href="https://tops.tongji.edu.cn/info/1031/1187.htm" target="_blank">Jian Sun</a>
            </div>
        </div>
        <div class="section abstract">
            <h2>Abstract</h2>
            <p>
                Although Deep Reinforcement Learning (DRL) and Large Language Models (LLMs) each show promise in addressing decision-making challenges in autonomous driving, DRL often suffers from high sample complexity, while LLMs have difficulty ensuring real-time decision making. To address these limitations, we propose TeLL-Drive, a hybrid framework that integrates an Teacher LLM to guide an attention-based Student DRL policy. By incorporating risk metrics, historical scenario retrieval, and domain heuristics into context-rich prompts, the LLM produces high-level driving strategies through chain-of-thought reasoning. A self-attention mechanism then fuses these strategies with the DRL agent’s exploration, accelerating policy convergence and boosting robustness across diverse driving conditions. Our experimental results, evaluated across multiple traffic scenarios, show that TeLL-Drive outperforms existing baseline methods, including other LLM-based approaches, in terms of success rates, average returns, and real-time feasibility. Ablation studies underscore the importance of each model component, especially the synergy between the attention mechanism and LLM-driven guidance. Finally, we build a virtual-real fusion experimental platform to verify the real-time, robustness, and reliability of the model when deployed in real vehicles through vehicle-in-loop experiments. These findings suggest that TeLL-Drive significantly enhances both the adaptability and safety of autonomous driving systems, while offering a more efficient and scalable approach for policy learning.
            </p>
        </div>
        <div class="section images">
            <h2>Images</h2>
            <img src="https://github.com/PerfectXu88/TeLL-Drive.github.io/blob/main/framework_v3.png?raw=true" alt="Conceptual Framework">
            <h3>The overall conceptual framework of TeLL-Drive, where a DRL student agent is guided by the LLM teacher for better decision making in autonomous driving.</h3>

            <img src="https://github.com/PerfectXu88/TeLL-Drive.github.io/blob/main/baselines.png?raw=true" alt="Baselines Comparison">
            <h3>Comparison of the performance of this model with traditional DRL training results.</h3>

            <img src="https://github.com/PerfectXu88/TeLL-Drive.github.io/blob/main/Ablation.png?raw=true" alt="Ablation Study">
            <h3>Comparison of performance results during the ablation experiment training process.</h3>

            <img src="https://github.com/PerfectXu88/TeLL-Drive.github.io/blob/main/success_rate_comparison.png?raw=true" alt="Success Rate Comparison">
            <h3>Comparison of testing success rate results between the teacher agent and the student agent.</h3>
        </div>
        <div class="section video">
            <h2>Videos</h2>
            <h3>Simulation Experiment</h3>
            <video controls>
                <source src="https://github.com/PerfectXu88/TeLL-Drive.github.io/blob/main/Unsignalized%20intersection.mp4?raw=true" type="video/mp4">
            </video>
            <h4>Unsignalized intersection: The agent must execute an unprotected left turn at an unsignalized intersection, requiring conflict resolution and time-slot preemption to navigate crossing traffic safely.</h4>

            <video controls>
                <source src="https://github.com/PerfectXu88/TeLL-Drive.github.io/blob/main/High-Speed%20Ramp%20Merging.mp4?raw=true" type="video/mp4">
            </video>
            <h4>High-Speed Ramp Merging: The agent operates on an acceleration lane, performing speed matching and gap selection to merge seamlessly into highway traffic at elevated velocities.</h4>

            <video controls>
                <source src="https://github.com/PerfectXu88/TeLL-Drive.github.io/blob/main/Four-Lane%20Adaptive%20Cruise.mp4?raw=true" type="video/mp4">
            </video>
            <h4>Four-Lane Adaptive Cruise: The agent focuses on fine-grained control of inter-vehicle distances and speeds across four lanes, highlighting precision in longitudinal control and continuous lane tracking.</h4>

            <h3>Vehicle-in-loop Experiment</h3>
            <video controls>
                <source src="https://github.com/PerfectXu88/TeLL-Drive.github.io/blob/main/Vehicle-in-loop%20experiment%20case%201.mp4?raw=true" type="video/mp4">
            </video>
            <h4>In Case 1, the autonomous vehicle equipped with TeLL-Drive begins from a standstill and accelerates toward the intersection. As it approaches the stop line, the vehicle slows down to create sufficient observation and decision space, enhancing its ability to assess the surrounding traffic. By the 7th second, the vehicle encounters an oncoming vehicle. Upon assessing the situation, the vehicle decides to slow further at the 12th second to yield and avoid a collision. After the oncoming vehicle passes, the autonomous vehicle resumes acceleration and approaches the exit road of the intersection by the 15th second. To maintain a safe distance from the vehicle in front, the system performs adaptive acceleration and deceleration, ensuring both safety and traffic efficiency. In this case, the autonomous vehicle is the last to leave the intersection, but the maneuver was executed safely and efficiently.</h4>

            <video controls>
                <source src="https://github.com/PerfectXu88/TeLL-Drive.github.io/blob/main/Vehicle-in-loop%20experiment%20case%202.mp4?raw=true" type="video/mp4">
            </video>
            <h4>In Case 2, the autonomous vehicle follows similar actions up to the point before entering the intersection. However, at the 6th second, the vehicle observes fewer vehicles in the intersection and determines it can pass first, so it accelerates. By the 8th second, a vehicle on the left side approaches, prompting an interaction. After a brief period of strong interaction between the two vehicles, the simulation vehicle (SV) decides to slow down and stop, while our autonomous vehicle continues to pass first, successfully navigating the intersection.</h4>
        </div>
    </div>
</body>
</html>