A new arXiv paper introduces Robusto-2, a benchmark evaluating both humans and vision-language models on autonomous driving tasks in Lima, Peru and New York City. The work targets how well multi-modal systems generalize when deployed in unfamiliar, out-of-distribution urban environments. It is a cross-listed replacement submission on arXiv cs.AI.
A new arXiv paper proposes combining Interacting Multiple Model (IMM) filtering with a state prediction neural network for multiple object tracking in autonomous driving. The approach targets radar sensing, which keeps working in poor weather and provides relative velocity via the Doppler effect. The work aims to improve tracking robustness for obstacle avoidance and route planning.
A new arXiv paper proposes using convolutional spiking neural networks with temporal data augmentation to predict whether a pedestrian intends to cross the road, based on event-based camera input. The authors frame the task as safety-critical for autonomous driving, where inference must hold up under motion blur, high dynamic range scenes, and imbalanced classes. The work is positioned as an alternative to conventional frame-based deep learning pipelines for this prediction problem.
A new arXiv paper claims that a common mistake in assessing deployed AI systems is treating data gathered for operational monitoring or regulatory compliance as though it were collected for comparative evaluation. Using automated driving as its main example, the work calls for clearer measurement validity standards so that compliance-oriented datasets are not used to make comparative performance claims. The authors frame this as a recurring evaluation failure rather than an isolated incident.
A new arXiv paper addresses computational cost and multi-scale feature extraction limits in bird's-eye-view (BEV) 3D object detection for autonomous driving. The authors combine bi-level routing with sparse spatial attention to improve the efficiency of dense 2D-to-BEV view transformation. The work targets multi-view camera setups used in self-driving perception.
A new arXiv preprint proposes using predictive audio representations to spot and track hidden dynamic objects before they come into view. The authors argue that occluded traffic agents can appear too late for detection systems, and that audio cues offer an earlier warning signal. The work targets safety-critical settings such as autonomous driving.
A new arXiv paper introduces DiffAdapterVLA, a method that folds continuous trajectory planning directly into the backbone of a pretrained driving vision-language model. The authors argue that while driving VLMs absorb rich visual, route, language and driving context, their training objectives stay disconnected from continuous planning, so prior work tends to bolt planning on separately. The approach targets native continuous trajectory generation rather than a detached planning head.
A new arXiv paper introduces GT-Space, a technique for multi-agent collaborative perception in autonomous driving where vehicles with different sensor modalities share perceptual data. The method uses ground truth feature space to address the challenge of aligning heterogeneous features across agents. It appears across arXiv cs.AI and cs.LG as a cross-listing revision.
Tesla's Cybercab robotaxis have reportedly been operating in New York since August 2026 without any formal announcement from the company. Unlike the previously shown design, the vehicles are equipped with a steering wheel and carry a person sitting behind it, which the report attributes to a specific reason. No official confirmation of the service has been issued.
Tesla's robotaxis and Cybercabs are not permitted to operate on highways, forcing longer surface-street routes. One passenger's trip that would normally take about 10 minutes stretched to roughly 70 minutes because of the detour. The restriction highlights the limited operational area of Tesla's autonomous ride service.
A newly updated arXiv paper introduces MAVEN-T, a method combining reinforcement learning with heterogeneous knowledge distillation to forecast the future paths of multiple agents simultaneously. The work targets real-time deployment in autonomous vehicles, where anticipating surrounding traffic informs collision checking, planning, and control. The approach is designed to remain dependable in dense scenarios with diverse and multimodal agent behaviors.
A new arXiv paper proposes a perception pipeline for Formula Student Driverless vehicles that relies only on LiDAR and a random forest classifier, avoiding deep learning and multi-sensor fusion. The authors aim to cut latency and hardware demands so the system can run without GPU acceleration.
A revised arXiv paper presents a multi-modal perception pipeline aimed at object detection and tracking for autonomous racing vehicles. The work addresses robustness challenges such as degraded visibility, sensor noise, and sensor failures that remain difficult for driving perception systems. It falls under the cs.AI category and appears as a replacement submission.