arXiv:2502.07631v3 Announce Type: replace 
Abstract: Perceiving the environment and its changes over time corresponds to two fundamental yet heterogeneous types of information: semantics and motion. Previous end-to-end autonomous driving works represent both types of information in a single feature vector. However, including motion related tasks, such as prediction and planning, impairs detection and tracking performance, a phenomenon known as negative transfer in multi-task learning. To address this issue, we propose Neural-Bayes motion decoding, a novel parallel detection, tracking, and prediction method that separates semantic and motion learning. Specifically, we employ a set of learned motion queries that operate in parallel with detection and tracking queries, sharing a unified set of recursively updated reference points. Moreover, we employ interactive semantic decoding to enhance information exchange in semantic tasks, promoting positive transfer. Experiments on the nuScenes dataset with UniAD and SparseDrive confirm the effectiveness of our divide and merge approach, resulting in performance improvements across perception, prediction, and planning. Our code is available at https://github.com/shenyinzhe/DMAD.

يتناول المقال نهجًا مبتكرًا للقيادة الذاتية من النهاية إلى النهاية يفصل بين التعلم الدلالي والتعلم الحركي لتحسين أداء الكشف والتتبع. الطريقة المقترحة، وهي فك تشفير الحركة Neural-Bayes، تستخدم استعلامات حركة متعلمة بالتوازي مع استعلامات الكشف والتتبع، مما يعزز تبادل المعلومات من خلال فك التشفير الدلالي التفاعلي. هذا يعالج مشكلة النقل السلبي التي لوحظت في التعلم متعدد المهام، والتي يمكن أن تعيق الأداء في مهام القيادة الذاتية.

El artículo aborda un enfoque novedoso para la conducción autónoma de extremo a extremo que separa el aprendizaje semántico y el aprendizaje del movimiento para mejorar el rendimiento en la detección y el seguimiento. El método propuesto, decodificación de movimiento Neural-Bayes, utiliza consultas de movimiento aprendidas en paralelo con consultas de detección y seguimiento, mejorando el intercambio de información a través de la decodificación semántica interactiva. Esto aborda el problema de transferencia negativa observado en el aprendizaje multitarea, que puede obstaculizar el rendimiento …

L'article traite d'une approche novatrice pour la conduite autonome de bout en bout qui sépare l'apprentissage sémantique et l'apprentissage du mouvement afin d'améliorer les performances de détection et de suivi. La méthode proposée, le décodage de mouvement Neural-Bayes, utilise des requêtes de mouvement apprises en parallèle avec des requêtes de détection et de suivi, améliorant l'échange d'informations grâce à un décodage sémantique interactif. Cela aborde le problème de transfert négatif observé dans l'apprentissage multitâche, qui peut nuire aux performances des tâches de conduite autono…

The article discusses a novel approach to end-to-end autonomous driving that separates semantic and motion learning to improve detection and tracking performance. The proposed method, Neural-Bayes motion decoding, utilizes learned motion queries in parallel with detection and tracking queries, enhancing information exchange through interactive semantic decoding. This addresses the negative transfer issue seen in multi-task learning, which can hinder performance in autonomous driving tasks.

Divide and Merge: Motion and Semantic Learning in End-to-End Autonomous Driving

arXiv:2503.07168v3 Announce Type: replace 
Abstract: As an essential component of autonomous driving systems, high-definition (HD) maps provide rich and precise environmental information for auto-driving scenarios; however, existing methods, which primarily rely on query-based detection frameworks to directly model map elements or implicitly propagate queries over time, often struggle to maintain consistent temporal perception outcomes. These inconsistencies pose significant challenges to the stability and reliability of real-world autonomous driving and map data collection systems. To address this limitation, we propose a novel end-to-end tracking framework for global map construction by temporally tracking map elements' historical trajectories. Firstly, instance-level historical rasterization map representation is designed to explicitly store previous perception results, which can control and maintain different global instances' history information in a fine-grained way. Secondly, we introduce a Map-Trajectory Prior Fusion module within this tracking framework, leveraging historical priors for tracked instances to improve temporal smoothness and continuity. Thirdly, we propose a global perspective metric to evaluate the quality of temporal geometry construction in HD maps, filling the gap in current metrics for assessing global geometric perception results. Substantial experiments on the nuScenes and Argoverse2 datasets demonstrate that the proposed method outperforms state-of-the-art (SOTA) methods in both single-frame and temporal metrics. The project page is available at: https://yj772881654.github.io/HisTrackMap.

يقدم HisTrackMap إطار عمل جديدًا للتتبع من البداية إلى النهاية لبناء خرائط عالية الدقة العالمية، حيث يتناول التحديات المتعلقة بالحفاظ على نتائج الإدراك الزمني المتسقة في أنظمة القيادة الذاتية. يستخدم هذا الإطار مسارات تاريخية لعناصر الخريطة لتحسين استقرار وموثوقية جمع بيانات الخرائط.

HisTrackMap presenta un nuevo marco de seguimiento de extremo a extremo para la construcción de mapas de alta definición globales, abordando los desafíos de mantener resultados de percepción temporal coherentes en los sistemas de conducción autónoma. Este marco utiliza las trayectorias históricas de los elementos del mapa para mejorar la estabilidad y la fiabilidad de la recopilación de datos cartográficos.

HisTrackMap présente un nouveau cadre de suivi de bout en bout pour la construction de cartes haute définition mondiales, répondant aux défis de maintien de résultats de perception temporelle cohérents dans les systèmes de conduite autonome. Ce cadre utilise les trajectoires historiques des éléments de la carte pour améliorer la stabilité et la fiabilité de la collecte de données cartographiques.

HisTrackMap introduces a novel end-to-end tracking framework for global high-definition map construction, addressing the challenges of maintaining consistent temporal perception outcomes in autonomous driving systems. This framework utilizes historical trajectories of map elements to enhance the stability and reliability of map data collection.

HisTrackMap: Global Vectorized High-Definition Map Construction via History Map Tracking

arXiv:2506.18124v3 Announce Type: replace-cross 
Abstract: Multiobject tracking (MOT) is an important task in applications including autonomous driving, ocean sciences, and aerospace surveillance. Traditional MOT methods are model-based and combine sequential Bayesian estimation with data association and an object birth model. More recent methods are fully data-driven and rely on the training of neural networks. Both approaches offer distinct advantages in specific settings. In particular, model-based methods are generally applicable across a wide range of scenarios, whereas data-driven MOT achieves superior performance in scenarios where abundant labeled data for training is available. A natural thought is whether a general framework can integrate the two approaches. This paper introduces a hybrid method that utilizes neural networks to enhance specific aspects of the statistical model in Bayesian MOT that have been identified as overly simplistic. By doing so, the performance of the prediction and update steps of Bayesian MOT is improved. To ensure tractable computation, our framework uses belief propagation to avoid high-dimensional operations combined with sequential Monte Carlo methods to perform low-dimensional operations efficiently. The resulting method combines the flexibility and robustness of model-based approaches with the capability to learn complex information from data of neural networks. We evaluate the performance of the proposed method based on the nuScenes autonomous driving dataset and demonstrate that it has state-of-the-art performance.

يقدم مقال جديد طريقة هجينة لتتبع الكائنات المتعددة باستخدام نماذج بايزي، حيث يتم دمج الشبكات العصبية لتعزيز النماذج الإحصائية التقليدية، مما يعالج القيود الموجودة في الأساليب الحالية. يهدف هذا التطور إلى تحسين الأداء في تطبيقات متنوعة، بما في ذلك القيادة الذاتية والمراقبة الجوية.

Un nuevo artículo presenta un método híbrido para el seguimiento multiobjetos bayesiano (MOT) que integra redes neuronales para mejorar los modelos estadísticos tradicionales, abordando las limitaciones de los enfoques existentes. Este desarrollo busca mejorar el rendimiento en diversas aplicaciones, incluidas la conducción autónoma y la vigilancia aeroespacial.

Un nouvel article présente une méthode hybride pour le suivi multi-objets bayésien (MOT) qui intègre des réseaux neuronaux pour améliorer les modèles statistiques traditionnels, en répondant aux limitations des approches existantes. Ce développement vise à améliorer les performances dans diverses applications, y compris la conduite autonome et la surveillance aérospatiale.

A new paper introduces a hybrid method for Bayesian Multiobject Tracking (MOT) that integrates neural networks to enhance traditional statistical models, addressing limitations in existing approaches. This development aims to improve performance in various applications, including autonomous driving and aerospace surveillance.

Bayesian Multiobject Tracking With Neural-Enhanced Motion and Measurement Models

One More Thing in AI – Your Shortcut to AI Mastery

Divide and Merge: Motion and Semantic Learning in End-to-End Autonomous Driving

Was this article worth reading? Share it

One More Thing in AI

LucidQuery AI

Attentive AI

Https

The Visualizer

Zemith-3bda3b

Ready to build your own newsroom?