Scholarlink

<p>This work presents a novel algorithm for local path planning for autonomous vehicles (AVs), prioritizing safety and adherence to traffic regulations. The algorithm integrates the Twin Delayed Deep Deterministic Policy Gradient (TD3) with sensor fusion based on Nvidia Convolutional Neural Network (NCNN). Using the CARLA simulator and real-world datasets like KITTI and WAYMO, the proposed algorithm leverages Imitation Learning (IL) and Deep Reinforcement Learning (DRL). IL uses human driving data for rule-abiding behavior, while DRL refines decisions through real-time interaction. This integrated approach overcomes the limitations of individual techniques, such as the lack of generalizability in Supervised Learning (SL) methods. CARLA simulations show 100% route completion, zero collisions, and perfect traffic rules adherence. Real-world tests confirmed its ability to generalize from simulations to real driving conditions. Sensor fusion mitigates individual sensor weaknesses, enhancing safety and efficiency.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Navigating the Rules: Integrating TD3 and Sensor Fusion for Traffic-Aware Autonomous Vehicle Path Planning

  • 期刊论文

  • 发表日期：

    2025年10月6日

查看全文
  • Mahmoud Elsayed,
  • Amr El-Mougy

摘要

This work presents a novel algorithm for local path planning for autonomous vehicles (AVs), prioritizing safety and adherence to traffic regulations. The algorithm integrates the Twin Delayed Deep Deterministic Policy Gradient (TD3) with sensor fusion based on Nvidia Convolutional Neural Network (NCNN). Using the CARLA simulator and real-world datasets like KITTI and WAYMO, the proposed algorithm leverages Imitation Learning (IL) and Deep Reinforcement Learning (DRL). IL uses human driving data for rule-abiding behavior, while DRL refines decisions through real-time interaction. This integrated approach overcomes the limitations of individual techniques, such as the lack of generalizability in Supervised Learning (SL) methods. CARLA simulations show 100% route completion, zero collisions, and perfect traffic rules adherence. Real-world tests confirmed its ability to generalize from simulations to real driving conditions. Sensor fusion mitigates individual sensor weaknesses, enhancing safety and efficiency.