<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/"><channel><title>Trajectory on Xu'Blog</title><link>https://xuquant.com/tags/trajectory/</link><description>Recent content in Trajectory on Xu'Blog</description><image><title>Xu'Blog</title><url>https://xuquant.com/og-default.png</url><link>https://xuquant.com/og-default.png</link></image><generator>Hugo -- 0.152.2</generator><language>zh</language><lastBuildDate>Sat, 21 Feb 2026 10:00:00 +0800</lastBuildDate><atom:link href="https://xuquant.com/tags/trajectory/index.xml" rel="self" type="application/rss+xml"/><item><title>Driving JEPA 综述：V-JEPA 系列方法在自动驾驶场景的应用</title><link>https://xuquant.com/posts/world-models/driving-jepa/</link><pubDate>Sat, 21 Feb 2026 10:00:00 +0800</pubDate><guid>https://xuquant.com/posts/world-models/driving-jepa/</guid><description>V-JEPA 系列在自动驾驶 benchmark 上的迁移综述：因果未来掩码、motion-aware mask、temporal-coherent mask 等 driving-specific 变体的 fine-tune 结果对比，以及 driving 与通用视频自监督在 mask 假设上的根本 mismatch。</description></item><item><title>Trajectory Tokenization for Autoregressive Planning: Clustering, Matching, and the AR+Diffusion Paradigm</title><link>https://xuquant.com/posts/autonomous-driving/ar-trajectory-tokenization/</link><pubDate>Sat, 28 Jun 2025 10:00:00 +0800</pubDate><guid>https://xuquant.com/posts/autonomous-driving/ar-trajectory-tokenization/</guid><description>深入探讨自回归驾驶规划器的轨迹分词方法：从基于 k-means 聚类的状态离散化，到 token 匹配与重建，再到 AR+Diffusion 范式与基于 GRPO 的强化学习后训练。</description></item></channel></rss>