# Talk2Sensors: 3D Visual Grounding in Autonomous Driving via Sensor-Adaptive Physical Cue Matching

arXiv: https://arxiv.org/abs/2608.04568 · HTML: https://mlsanity.com/paper/2608.04568/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 25 of its 41 references that are in the map.
- First posted on arXiv 5 Aug 2026.

## Builds on (most cited first, 25 of 25)

- [End-to-End Object Detection with Transformers](https://mlsanity.com/paper/2005.12872/) (2020)
- [Squeeze-and-Excitation Networks](https://mlsanity.com/paper/1709.01507/) (2017)
- [PointPillars: Fast Encoders for Object Detection from Point Clouds](https://mlsanity.com/paper/1812.05784/) (2018)
- [BEVFusion: Multi-Task Multi-Sensor Fusion with Unified Bird's-Eye View Representation](https://mlsanity.com/paper/2205.13542/) (2022)
- [DETR3D: 3D Object Detection from Multi-view Images via 3D-to-2D Queries](https://mlsanity.com/paper/2110.06922/) (2021)
- [ScanRefer: 3D Object Localization in RGB-D Scans using Natural Language](https://mlsanity.com/paper/1912.08830/) (2019)
- [Gated Multimodal Units for Information Fusion](https://mlsanity.com/paper/1702.01992/) (2017)
- [FUTR3D: A Unified Sensor Fusion Framework for 3D Detection](https://mlsanity.com/paper/2203.10642/) (2022)
- [EDA: Explicit Text-Decoupling and Dense Alignment for 3D Visual Grounding](https://mlsanity.com/paper/2209.14941/) (2022)
- [SeeGround: See and Ground for Zero-Shot Open-Vocabulary 3D Visual Grounding](https://mlsanity.com/paper/2412.04383/) (2024)
- [VLM-Grounder: A VLM Agent for Zero-Shot 3D Visual Grounding](https://mlsanity.com/paper/2410.13860/) (2024)
- [Refer-it-in-RGBD: A Bottom-up Approach for 3D Visual Grounding in RGBD Images](https://mlsanity.com/paper/2103.07894/) (2021)
- [OmniHD-Scenes: A Next-Generation Multimodal Dataset for Autonomous Driving](https://mlsanity.com/paper/2412.10734/) (2024)
- [A Survey on Text-guided 3D Visual Grounding: Elements, Recent Advances, and Future Directions](https://mlsanity.com/paper/2406.05785/) (2024)
- [Mono3DVG: 3D Visual Grounding in Monocular Images](https://mlsanity.com/paper/2312.08022/) (2023)
- [Language-Guided 3D Object Detection in Point Cloud for Autonomous Driving](https://mlsanity.com/paper/2305.15765/) (2023)
- [Talk2Radar: Bridging Natural Language with 4D mmWave Radar for 3D Referring Expression Comprehension](https://mlsanity.com/paper/2405.12821/) (2024)
- [TSP3D: Text-guided Sparse Voxel Pruning for Efficient 3D Visual Grounding](https://mlsanity.com/paper/2502.10392/) (2025)
- [NuGrounding: A Multi-View 3D Visual Grounding Framework in Autonomous Driving](https://mlsanity.com/paper/2503.22436/) (2025)
- [Ges3ViG: Incorporating Pointing Gestures into Language-Based 3D Visual Grounding for Embodied Reference Understanding](https://mlsanity.com/paper/2504.09623/) (2025)
- [Talk to Parallel LiDARs: A Human-LiDAR Interaction Method Based on 3D Visual Grounding](https://mlsanity.com/paper/2405.15274/) (2024)
- [Availability-aware Sensor Fusion via Unified Canonical Space](https://mlsanity.com/paper/2503.07029/) (2025)
- [Enhanced Vision-Language Models for Diverse Sensor Understanding: Cost-Efficient Optimization and Benchmarking](https://mlsanity.com/paper/2412.20750/) (2024)
- [MMDrive: Interactive Scene Understanding Beyond Vision with Multi-representational Fusion](https://mlsanity.com/paper/2512.13177/) (2025)
- [LidaRefer: Context-aware Outdoor 3D Visual Grounding for Autonomous Driving](https://mlsanity.com/paper/2411.04351/) (2024)

## Abstract (arXiv)

As a key capability for embodied intelligence, 3D visual grounding (3DVG) has been predominantly studied in indoor scenes with RGB-D or point-cloud inputs, while existing outdoor extensions largely rely on monocular images alone. Both settings fall short of real-world outdoor perception, where heterogeneous sensors capture complementary yet distinct physical properties, such as visual texture, 3D geometry, and object kinematics, that are indispensable for flexible and robust query-adaptive grounding but remain under-exploited. To bridge this gap, we introduce Talk2Sensors, the first multi-sensor 3D visual grounding dataset built upon camera, LiDAR, and 4D radar. It contains 8,682 language instructions and 20,558 referred objects, with diverse prompts explicitly aligned with sensor-specific physical cues. Furthermore, we propose TSFormer, a unified Transformer-based framework for language-guided 3D visual grounding in autonomous driving. TSFormer adopts a coarse-to-fine property-aware fusion strategy: the Language-Routed Property Sampler first performs coarse text-conditioned feature retrieval by modulating sensor sampling weights with query-level linguistic cues, while the subsequent Sparse-Preserving Modality Arbiter module conducts fine-grained modality arbitration and text-guided refinement to determine the precise referred spatial location. This design enables dynamic routing of appearance, geometry, and motion cues according to the semantic requirements of each prompt, preventing dense modalities from overwhelming sparse but critical sensor signals. Extensive experiments demonstrate that TSFormer achieves state-of-the-art performance across multiple benchmarks: it improves over the strongest baseline by 8.05 mAP on Talk2Sensors, and transfers to the monocular Mono3DRefer benchmark with 53.05% Acc@0.5.