Jun Won Choi

Junho Koh, Junhyung Lee, Youngwoo Lee et al.

Most scanning LiDAR sensors generate a sequence of point clouds in real-time. While conventional 3D object detectors use a set of unordered LiDAR points acquired over a fixed time interval, recent studies have revealed that substantial performance improvement can be achieved by exploiting the spatio-temporal context present in a sequence of LiDAR point sets. In this paper, we propose a novel 3D object detection architecture, which can encode LiDAR point cloud sequences acquired by multiple successive scans. The encoding process of the point cloud sequence is performed on two different time scales. We first design a short-term motion-aware voxel encoding that captures the short-term temporal changes of point clouds driven by the motion of objects in each voxel. We also propose long-term motion-guided bird's eye view (BEV) feature enhancement that adaptively aligns and aggregates the BEV feature maps obtained by the short-term voxel encoding by utilizing the dynamic motion context inferred from the sequence of the feature maps. The experiments conducted on the public nuScenes benchmark demonstrate that the proposed 3D object detector offers significant improvements in performance compared to the baseline methods and that it sets a state-of-the-art performance for certain 3D object detection categories. Code is available at https://github.com/HYjhkoh/MGTANet.git

Youngseok Kim, Sanmin Kim, Jun Won Choi et al.

Camera and radar sensors have significant advantages in cost, reliability, and maintenance compared to LiDAR. Existing fusion methods often fuse the outputs of single modalities at the result-level, called the late fusion strategy. This can benefit from using off-the-shelf single sensor detection algorithms, but late fusion cannot fully exploit the complementary properties of sensors, thus having limited performance despite the huge potential of camera-radar fusion. Here we propose a novel proposal-level early fusion approach that effectively exploits both spatial and contextual properties of camera and radar for 3D object detection. Our fusion framework first associates image proposal with radar points in the polar coordinate system to efficiently handle the discrepancy between the coordinate system and spatial properties. Using this as a first stage, following consecutive cross-attention based feature fusion layers adaptively exchange spatio-contextual information between camera and radar, leading to a robust and attentive fusion. Our camera-radar fusion approach achieves the state-of-the-art 41.1% mAP and 52.3% NDS on the nuScenes test set, which is 8.7 and 10.8 points higher than the camera-only baseline, as well as yielding competitive performance on the LiDAR method.

Youngseok Kim, Juyeb Shin, Sanmin Kim et al.

Autonomous driving requires an accurate and fast 3D perception system that includes 3D object detection, tracking, and segmentation. Although recent low-cost camera-based approaches have shown promising results, they are susceptible to poor illumination or bad weather conditions and have a large localization error. Hence, fusing camera with low-cost radar, which provides precise long-range measurement and operates reliably in all environments, is promising but has not yet been thoroughly investigated. In this paper, we propose Camera Radar Net (CRN), a novel camera-radar fusion framework that generates a semantically rich and spatially accurate bird's-eye-view (BEV) feature map for various tasks. To overcome the lack of spatial information in an image, we transform perspective view image features to BEV with the help of sparse but accurate radar points. We further aggregate image and radar feature maps in BEV using multi-modal deformable attention designed to tackle the spatial misalignment between inputs. CRN with real-time setting operates at 20 FPS while achieving comparable performance to LiDAR detectors on nuScenes, and even outperforms at a far distance on 100m setting. Moreover, CRN with offline setting yields 62.4% NDS, 57.5% mAP on nuScenes test set and ranks first among all camera and camera-radar 3D object detectors.

Jisong Kim, Minjae Seong, Geonho Bang et al.

While LiDAR sensors have been successfully applied to 3D object detection, the affordability of radar and camera sensors has led to a growing interest in fusing radars and cameras for 3D object detection. However, previous radar-camera fusion models were unable to fully utilize the potential of radar information. In this paper, we propose Radar-Camera Multi-level fusion (RCM-Fusion), which attempts to fuse both modalities at both feature and instance levels. For feature-level fusion, we propose a Radar Guided BEV Encoder which transforms camera features into precise BEV representations using the guidance of radar Bird's-Eye-View (BEV) features and combines the radar and camera BEV features. For instance-level fusion, we propose a Radar Grid Point Refinement module that reduces localization error by accounting for the characteristics of the radar point clouds. The experiments conducted on the public nuScenes dataset demonstrate that our proposed RCM-Fusion achieves state-of-the-art performances among single frame-based radar-camera fusion methods in the nuScenes 3D object detection benchmark. Code will be made publicly available.

Yecheol Kim, Konyul Park, Minwook Kim et al.

Fusing data from cameras and LiDAR sensors is an essential technique to achieve robust 3D object detection. One key challenge in camera-LiDAR fusion involves mitigating the large domain gap between the two sensors in terms of coordinates and data distribution when fusing their features. In this paper, we propose a novel camera-LiDAR fusion architecture called, 3D Dual-Fusion, which is designed to mitigate the gap between the feature representations of camera and LiDAR data. The proposed method fuses the features of the camera-view and 3D voxel-view domain and models their interactions through deformable attention. We redesign the transformer fusion encoder to aggregate the information from the two domains. Two major changes include 1) dual query-based deformable attention to fuse the dual-domain features interactively and 2) 3D local self-attention to encode the voxel-domain queries prior to dual-query decoding. The results of an experimental evaluation show that the proposed camera-LiDAR fusion architecture achieved competitive performance on the KITTI and nuScenes datasets, with state-of-the-art performances in some 3D object detection benchmarks categories.

Youngseok Kim, Sanmin Kim, Sangmin Sim et al.

Recent advances in monocular 3D detection leverage a depth estimation network explicitly as an intermediate stage of the 3D detection network. Depth map approaches yield more accurate depth to objects than other methods thanks to the depth estimation network trained on a large-scale dataset. However, depth map approaches can be limited by the accuracy of the depth map, and sequentially using two separated networks for depth estimation and 3D detection significantly increases computation cost and inference time. In this work, we propose a method to boost the RGB image-based 3D detector by jointly training the detection network with a depth prediction loss analogous to the depth estimation task. In this way, our 3D detection network can be supervised by more depth supervision from raw LiDAR points, which does not require any human annotation cost, to estimate accurate depth without explicitly predicting the depth map. Our novel object-centric depth prediction loss focuses on depth around foreground objects, which is important for 3D object detection, to leverage pixel-wise depth supervision in an object-centric manner. Our depth regression model is further trained to predict the uncertainty of depth to represent the 3D confidence of objects. To effectively train the 3D detector with raw LiDAR points and to enable end-to-end training, we revisit the regression target of 3D objects and design a network architecture. Extensive experiments on KITTI and nuScenes benchmarks show that our method can significantly boost the monocular image-based 3D detector to outperform depth map approaches while maintaining the real-time inference speed.

Sehwan Choi, Jungho Kim, Hongjae Shin et al.

In this paper, we introduce Mask2Map, a novel end-to-end online HD map construction method designed for autonomous driving applications. Our approach focuses on predicting the class and ordered point set of map instances within a scene, represented in the bird's eye view (BEV). Mask2Map consists of two primary components: the Instance-Level Mask Prediction Network (IMPNet) and the Mask-Driven Map Prediction Network (MMPNet). IMPNet generates Mask-Aware Queries and BEV Segmentation Masks to capture comprehensive semantic information globally. Subsequently, MMPNet enhances these query features using local contextual information through two submodules: the Positional Query Generator (PQG) and the Geometric Feature Extractor (GFE). PQG extracts instance-level positional queries by embedding BEV positional information into Mask-Aware Queries, while GFE utilizes BEV Segmentation Masks to generate point-level geometric features. However, we observed limited performance in Mask2Map due to inter-network inconsistency stemming from different predictions to Ground Truth (GT) matching between IMPNet and MMPNet. To tackle this challenge, we propose the Inter-network Denoising Training method, which guides the model to denoise the output affected by both noisy GT queries and perturbed GT Segmentation Masks. Our evaluation conducted on nuScenes and Argoverse2 benchmarks demonstrates that Mask2Map achieves remarkable performance improvements over previous state-of-the-art methods, with gains of 10.1% mAP and 4.1 mAP, respectively. Our code can be found at https://github.com/SehwanChoi0307/Mask2Map.

Kyeongyeon Kim, Jun Won Choi, Suleyman S. Kozat et al.

We introduce a low complexity approach to iterative equalization and decoding, or "turbo equalization", that uses clustered models to better match the nonlinear relationship that exists between likelihood information from a channel decoder and the symbol estimates that arise in soft-input channel equalization. The introduced clustered turbo equalizer uses piecewise linear models to capture the nonlinear dependency of the linear minimum mean square error (MMSE) symbol estimate on the symbol likelihoods produced by the channel decoder and maintains a computational complexity that is only linear in the channel memory. By partitioning the space of likelihood information from the decoder, based on either hard or soft clustering, and using locally-linear adaptive equalizers within each clustered region, the performance gap between the linear MMSE equalizer and low-complexity, LMS-based linear turbo equalizers can be dramatically narrowed.

Sehwan Choi, Jungho Kim, Junyong Yun et al.

Predicting the future motion of dynamic agents is of paramount importance to ensuring safety and assessing risks in motion planning for autonomous robots. In this study, we propose a two-stage motion prediction method, called R-Pred, designed to effectively utilize both scene and interaction context using a cascade of the initial trajectory proposal and trajectory refinement networks. The initial trajectory proposal network produces M trajectory proposals corresponding to the M modes of the future trajectory distribution. The trajectory refinement network enhances each of the M proposals using 1) tube-query scene attention (TQSA) and 2) proposal-level interaction attention (PIA) mechanisms. TQSA uses tube-queries to aggregate local scene context features pooled from proximity around trajectory proposals of interest. PIA further enhances the trajectory proposals by modeling inter-agent interactions using a group of trajectory proposals selected by their distances from neighboring agents. Our experiments conducted on Argoverse and nuScenes datasets demonstrate that the proposed refinement network provides significant performance improvements compared to the single-stage baseline and that R-Pred achieves state-of-the-art performance in some categories of the benchmarks.

Jaehyun Park, Konyul Park, Daehun Kim et al.

In autonomous driving, transparency in the decision-making of perception models is critical, as even a single misperception can be catastrophic. Yet with multi-sensor inputs, it is difficult to determine how each modality contributes to a prediction because sensor information becomes entangled within the fusion network. We introduce Layer-Wise Modality Decomposition (LMD), a post-hoc, model-agnostic interpretability method that disentangles modality-specific information across all layers of a pretrained fusion model. To our knowledge, LMD is the first approach to attribute the predictions of a perception model to individual input modalities in a sensor-fusion system for autonomous driving. We evaluate LMD on pretrained fusion models under camera-radar, camera-LiDAR, and camera-radar-LiDAR settings for autonomous driving. Its effectiveness is validated using structured perturbation-based metrics and modality-wise visual decompositions, demonstrating practical applicability to interpreting high-capacity multimodal architectures. Code is available at https://github.com/detxter-jvb/Layer-Wise-Modality-Decomposition.

Junhyung Lee, Junho Koh, Youngwoo Lee et al.

LiDAR sensors are widely used for 3D object detection in various mobile robotics applications. LiDAR sensors continuously generate point cloud data in real-time. Conventional 3D object detectors detect objects using a set of points acquired over a fixed duration. However, recent studies have shown that the performance of object detection can be further enhanced by utilizing spatio-temporal information obtained from point cloud sequences. In this paper, we propose a new 3D object detector, named D-Align, which can effectively produce strong bird's-eye-view (BEV) features by aligning and aggregating the features obtained from a sequence of point sets. The proposed method includes a novel dual-query co-attention network that uses two types of queries, including target query set (T-QS) and support query set (S-QS), to update the features of target and support frames, respectively. D-Align aligns S-QS to T-QS based on the temporal context features extracted from the adjacent feature maps and then aggregates S-QS with T-QS using a gated attention mechanism. The dual queries are updated through multiple attention layers to progressively enhance the target frame features used to produce the detection results. Our experiments on the nuScenes dataset show that the proposed D-Align method greatly improved the performance of a single frame-based baseline method and significantly outperformed the latest 3D object detectors.

Jae Soon Baik, In Young Yoon, Jun Won Choi

There is growing interest in the challenging visual perception task of learning from long-tailed class distributions. The extreme class imbalance in the training dataset biases the model to prefer recognizing majority class data over minority class data. Furthermore, the lack of diversity in minority class samples makes it difficult to find a good representation. In this paper, we propose an effective data augmentation method, referred to as bilateral mixup augmentation, which can improve the performance of long-tailed visual recognition. The bilateral mixup augmentation combines two samples generated by a uniform sampler and a re-balanced sampler and augments the training dataset to enhance the representation learning for minority classes. We also reduce the classifier bias using class-wise temperature scaling, which scales the logits differently per class in the training phase. We apply both ideas to the dual-branch network (DBN) framework, presenting a new model, named dual-branch network with bilateral mixup (DBN-Mix). Experiments on popular long-tailed visual recognition datasets show that DBN-Mix improves performance significantly over baseline and that the proposed method achieves state-of-the-art performance in some categories of benchmarks.

Jungho Kim, Changwon Kang, Dongyoung Lee et al.

In this paper, we introduce ProtoOcc, a novel 3D occupancy prediction model designed to predict the occupancy states and semantic classes of 3D voxels through a deep semantic understanding of scenes. ProtoOcc consists of two main components: the Dual Branch Encoder (DBE) and the Prototype Query Decoder (PQD). The DBE produces a new 3D voxel representation by combining 3D voxel and BEV representations across multiple scales through a dual branch structure. This design enhances both performance and computational efficiency by providing a large receptive field for the BEV representation while maintaining a smaller receptive field for the voxel representation. The PQD introduces Prototype Queries to accelerate the decoding process. Scene-Adaptive Prototypes are derived from the 3D voxel features of input sample, while Scene-Agnostic Prototypes are computed by applying Scene-Adaptive Prototypes to an Exponential Moving Average during the training phase. By using these prototype-based queries for decoding, we can directly predict 3D occupancy in a single step, eliminating the need for iterative Transformer decoding. Additionally, we propose the Robust Prototype Learning, which injects noise into prototype generation process and trains the model to denoise during the training phase. ProtoOcc achieves state-of-the-art performance with 45.02% mIoU on the Occ3D-nuScenes benchmark. For single-frame method, it reaches 39.56% mIoU with an inference speed of 12.83 FPS on an NVIDIA RTX 3090. Our code can be found at https://github.com/SPA-junghokim/ProtoOcc.

Jun Ho Lee, Jae Soon Baik, Tae Hwan Hwang et al.

There are inevitably many mislabeled data in real-world datasets. Because deep neural networks (DNNs) have an enormous capacity to memorize noisy labels, a robust training scheme is required to prevent labeling errors from degrading the generalization performance of DNNs. Current state-of-the-art methods present a co-training scheme that trains dual networks using samples associated with small losses. In practice, however, training two networks simultaneously can burden computing resources. In this study, we propose a simple yet effective robust training scheme that operates by training only a single network. During training, the proposed method generates temporal self-ensemble by sampling intermediate network parameters from the weight trajectory formed by stochastic gradient descent optimization. The loss sum evaluated with these self-ensembles is used to identify incorrectly labeled samples. In parallel, our method generates multi-view predictions by transforming an input data into various forms and considers their agreement to identify incorrectly labeled samples. By combining the aforementioned metrics, we present the proposed {\it self-ensemble-based robust training} (SRT) method, which can filter the samples with noisy labels to reduce their influence on training. Experiments on widely-used public datasets demonstrate that the proposed method achieves a state-of-the-art performance in some categories without training the dual networks.

Kyoung Ok Yang, Junho Koh, Jun Won Choi

Various types of sensors have been considered to develop human action recognition (HAR) models. Robust HAR performance can be achieved by fusing multimodal data acquired by different sensors. In this paper, we introduce a new multimodal fusion architecture, referred to as Unified Contrastive Fusion Transformer (UCFFormer) designed to integrate data with diverse distributions to enhance HAR performance. Based on the embedding features extracted from each modality, UCFFormer employs the Unified Transformer to capture the inter-dependency among embeddings in both time and modality domains. We present the Factorized Time-Modality Attention to perform self-attention efficiently for the Unified Transformer. UCFFormer also incorporates contrastive learning to reduce the discrepancy in feature distributions across various modalities, thus generating semantically aligned features for information fusion. Performance evaluation conducted on two popular datasets, UTD-MHAD and NTU RGB+D, demonstrates that UCFFormer achieves state-of-the-art performance, outperforming competing methods by considerable margins.

Jae Soon Baik, In Young Yoon, Kun Hoon Kim et al.

Deep neural networks have demonstrated remarkable advancements in various fields using large, well-annotated datasets. However, real-world data often exhibit long-tailed distributions and label noise, significantly degrading generalization performance. Recent studies addressing these issues have focused on noisy sample selection methods that estimate the centroid of each class based on high-confidence samples within each target class. The performance of these methods is limited because they use only the training samples within each class for class centroid estimation, making the quality of centroids susceptible to long-tailed distributions and noisy labels. In this study, we present a robust training framework called Distribution-aware Sample Selection and Contrastive Learning (DaSC). Specifically, DaSC introduces a Distribution-aware Class Centroid Estimation (DaCC) to generate enhanced class centroids. DaCC performs weighted averaging of the features from all samples, with weights determined based on model predictions. Additionally, we propose a confidence-aware contrastive learning strategy to obtain balanced and robust representations. The training samples are categorized into high-confidence and low-confidence samples. Our method then applies Semi-supervised Balanced Contrastive Loss (SBCL) using high-confidence samples, leveraging reliable label information to mitigate class bias. For the low-confidence samples, our method computes Mixup-enhanced Instance Discrimination Loss (MIDL) to improve their representations in a self-supervised manner. Our experimental results on CIFAR and real-world noisy-label datasets demonstrate the superior performance of the proposed DaSC compared to previous approaches.

Soo Won Seo, KyungChae Lee, Hyungchan Cho et al.

Human-Object Interaction (HOI) detection aims to localize human-object pairs and classify their interactions from a single image, a task that demands strong visual understanding and nuanced contextual reasoning. Recent approaches have leveraged Vision-Language Models (VLMs) to introduce semantic priors, significantly improving HOI detection performance. However, existing methods often fail to fully capitalize on the diverse contextual cues distributed across the entire scene. To overcome these limitations, we propose the Instance-centric Context Mining Network (InCoM-Net)-a novel framework that effectively integrates rich semantic knowledge extracted from VLMs with instance-specific features produced by an object detector. This design enables deeper interaction reasoning by modeling relationships not only within each detected instance but also across instances and their surrounding scene context. InCoM-Net comprises two core components: Instancecentric Context Refinement (ICR), which separately extracts intra-instance, inter-instance, and global contextual cues from VLM-derived features, and Progressive Context Aggregation (ProCA), which iteratively fuses these multicontext features with instance-level detector features to support high-level HOI reasoning. Extensive experiments on the HICO-DET and V-COCO benchmarks show that InCoM-Net achieves state-of-the-art performance, surpassing previous HOI detection methods. Code is available at https://github.com/nowuss/InCoM-Net.

Jae Soon Baik, In Young Yoon, Jun Won Choi

Modern deep learning has achieved great success in various fields. However, it requires the labeling of huge amounts of data, which is expensive and labor-intensive. Active learning (AL), which identifies the most informative samples to be labeled, is becoming increasingly important to maximize the efficiency of the training process. The existing AL methods mostly use only a single final fixed model for acquiring the samples to be labeled. This strategy may not be good enough in that the structural uncertainty of a model for given training data is not considered to acquire the samples. In this study, we propose a novel acquisition criterion based on temporal self-ensemble generated by conventional stochastic gradient descent (SGD) optimization. These self-ensemble models are obtained by capturing the intermediate network weights obtained through SGD iterations. Our acquisition function relies on a consistency measure between the student and teacher models. The student models are given a fixed number of temporal self-ensemble models, and the teacher model is constructed by averaging the weights of the student models. Using the proposed acquisition criterion, we present an AL algorithm, namely student-teacher consistency-based AL (ST-CoNAL). Experiments conducted for image classification tasks on CIFAR-10, CIFAR-100, Caltech-256, and Tiny ImageNet datasets demonstrate that the proposed ST-CoNAL achieves significantly better performance than the existing acquisition methods. Furthermore, extensive experiments show the robustness and effectiveness of our methods.

Yecheol Kim, Junho Lee, Changsoo Park et al.

3D object detection is crucial for applications like autonomous driving and robotics. However, in real-world environments, variations in sensor data distribution due to sensor upgrades, weather changes, and geographic differences can adversely affect detection performance. Semi-Supervised Domain Adaptation (SSDA) aims to mitigate these challenges by transferring knowledge from a source domain, abundant in labeled data, to a target domain where labels are scarce. This paper presents a new SSDA method referred to as Target-Oriented Domain Augmentation (TODA) specifically tailored for LiDAR-based 3D object detection. TODA efficiently utilizes all available data, including labeled data in the source domain, and both labeled data and unlabeled data in the target domain to enhance domain adaptation performance. TODA consists of two stages: TargetMix and AdvMix. TargetMix employs mixing augmentation accounting for LiDAR sensor characteristics to facilitate feature alignment between the source-domain and target-domain. AdvMix applies point-wise adversarial augmentation with mixing augmentation, which perturbs the unlabeled data to align the features within both labeled and unlabeled data in the target domain. Our experiments conducted on the challenging domain adaptation tasks demonstrate that TODA outperforms existing domain adaptation techniques designed for 3D object detection by significant margins. The code is available at: https://github.com/rasd3/TODA.

Seok Hwan Lee, Taein Son, Soo Won Seo et al.

Video action detection (VAD) is a formidable vision task that involves the localization and classification of actions within the spatial and temporal dimensions of a video clip. Among the myriad VAD architectures, two-stage VAD methods utilize a pre-trained person detector to extract the region of interest features, subsequently employing these features for action detection. However, the performance of two-stage VAD methods has been limited as they depend solely on localized actor features to infer action semantics. In this study, we propose a new two-stage VAD framework called Joint Actor-scene context Relation modeling based on Visual Semantics (JARViS), which effectively consolidates cross-modal action semantics distributed globally across spatial and temporal dimensions using Transformer attention. JARViS employs a person detector to produce densely sampled actor features from a keyframe. Concurrently, it uses a video backbone to create spatio-temporal scene features from a video clip. Finally, the fine-grained interactions between actors and scenes are modeled through a Unified Action-Scene Context Transformer to directly output the final set of actions in parallel. Our experimental results demonstrate that JARViS outperforms existing methods by significant margins and achieves state-of-the-art performance on three popular VAD datasets, including AVA, UCF101-24, and JHMDB51-21.

Konyul Park, Daehun Kim, Jiyong Oh et al.

Reliable off-road navigation requires accurate estimation of traversable regions and robust perception under diverse terrain and sensing conditions. However, existing datasets lack both scalability and multi-modality, which limits progress in 3D traversability prediction. In this work, we introduce STONE, a large-scale multi-modal dataset for off-road navigation. STONE provides (1) trajectory-guided 3D traversability maps generated by a fully automated, annotation-free pipeline, and (2) comprehensive surround-view sensing with synchronized 128-channel LiDAR, six RGB cameras, and three 4D imaging radars. The dataset covers a wide range of environments and conditions, including day and night, grasslands, farmlands, construction sites, and lakes. Our auto-labeling pipeline reconstructs dense terrain surfaces from LiDAR scans, extracts geometric attributes such as slope, elevation, and roughness, and assigns traversability labels beyond the robot's trajectory using a Mahalanobis-distance-based criterion. This design enables scalable, geometry-aware ground-truth construction without manual annotation. Finally, we establish a benchmark for voxel-level 3D traversability prediction and provide strong baselines under both single-modal and multi-modal settings. STONE is available at: https://konyul.github.io/STONE-dataset/

Geonho Bang, Kwangjin Choi, Jisong Kim et al.

The inherent noisy and sparse characteristics of radar data pose challenges in finding effective representations for 3D object detection. In this paper, we propose RadarDistill, a novel knowledge distillation (KD) method, which can improve the representation of radar data by leveraging LiDAR data. RadarDistill successfully transfers desirable characteristics of LiDAR features into radar features using three key components: Cross-Modality Alignment (CMA), Activation-based Feature Distillation (AFD), and Proposal-based Feature Distillation (PFD). CMA enhances the density of radar features by employing multiple layers of dilation operations, effectively addressing the challenge of inefficient knowledge transfer from LiDAR to radar. AFD selectively transfers knowledge based on regions of the LiDAR features, with a specific focus on areas where activation intensity exceeds a predefined threshold. PFD similarly guides the radar network to selectively mimic features from the LiDAR network within the object proposals. Our comparative analyses conducted on the nuScenes datasets demonstrate that RadarDistill achieves state-of-the-art (SOTA) performance for radar-only object detection task, recording 20.5% in mAP and 43.7% in NDS. Also, RadarDistill significantly improves the performance of the camera-radar fusion model.

Jisong Kim, Minjae Seong, Jun Won Choi

Accurate and robust 3D object detection is a critical component in autonomous vehicles and robotics. While recent radar-camera fusion methods have made significant progress by fusing information in the bird's-eye view (BEV) representation, they often struggle to effectively capture the motion of dynamic objects, leading to limited performance in real-world scenarios. In this paper, we introduce CRT-Fusion, a novel framework that integrates temporal information into radar-camera fusion to address this challenge. Our approach comprises three key modules: Multi-View Fusion (MVF), Motion Feature Estimator (MFE), and Motion Guided Temporal Fusion (MGTF). The MVF module fuses radar and image features within both the camera view and bird's-eye view, thereby generating a more precise unified BEV representation. The MFE module conducts two simultaneous tasks: estimation of pixel-wise velocity information and BEV segmentation. Based on the velocity and the occupancy score map obtained from the MFE module, the MGTF module aligns and fuses feature maps across multiple timestamps in a recurrent manner. By considering the motion of dynamic objects, CRT-Fusion can produce robust BEV feature maps, thereby improving detection accuracy and robustness. Extensive evaluations on the challenging nuScenes dataset demonstrate that CRT-Fusion achieves state-of-the-art performance for radar-camera-based 3D object detection. Our approach outperforms the previous best method in terms of NDS by +1.7%, while also surpassing the leading approach in mAP by +1.4%. These significant improvements in both metrics showcase the effectiveness of our proposed fusion strategy in enhancing the reliability and accuracy of 3D object detection.

Konyul Park, Yecheol Kim, Daehun Kim et al.

Modern autonomous driving perception systems utilize complementary multi-modal sensors, such as LiDAR and cameras. Although sensor fusion architectures enhance performance in challenging environments, they still suffer significant performance drops under severe sensor failures, such as LiDAR beam reduction, LiDAR drop, limited field of view, camera drop, and occlusion. This limitation stems from inter-modality dependencies in current sensor fusion frameworks. In this study, we introduce an efficient and robust LiDAR-camera 3D object detector, referred to as MoME, which can achieve robust performance through a mixture of experts approach. Our MoME fully decouples modality dependencies using three parallel expert decoders, which use camera features, LiDAR features, or a combination of both to decode object queries, respectively. We propose Multi-Expert Decoding (MED) framework, where each query is decoded selectively using one of three expert decoders. MoME utilizes an Adaptive Query Router (AQR) to select the most appropriate expert decoder for each query based on the quality of camera and LiDAR features. This ensures that each query is processed by the best-suited expert, resulting in robust performance across diverse sensor failure scenarios. We evaluated the performance of MoME on the nuScenes-R benchmark. Our MoME achieved state-of-the-art performance in extreme weather and sensor failure conditions, significantly outperforming the existing models across various sensor failure scenarios.

Minjae Seong, Jisong Kim, Geonho Bang et al.

Accurate 3D perception is essential for understanding the environment in autonomous driving. Recent advancements in 3D semantic occupancy prediction have leveraged camera-LiDAR fusion to improve robustness and accuracy. However, current methods allocate computational resources uniformly across all voxels, leading to inefficiency, and they also fail to adequately address occlusions, resulting in reduced accuracy in challenging scenarios. We propose MR-Occ, a novel approach for camera-LiDAR fusion-based 3D semantic occupancy prediction, addressing these challenges through three key components: Hierarchical Voxel Feature Refinement (HVFR), Multi-scale Occupancy Decoder (MOD), and Pixel to Voxel Fusion Network (PVF-Net). HVFR improves performance by enhancing features for critical voxels, reducing computational cost. MOD introduces an `occluded' class to better handle regions obscured from sensor view, improving accuracy. PVF-Net leverages densified LiDAR features to effectively fuse camera and LiDAR data through a deformable attention mechanism. Extensive experiments demonstrate that MR-Occ achieves state-of-the-art performance on the nuScenes-Occupancy dataset, surpassing previous approaches by +5.2% in IoU and +5.3% in mIoU while using fewer parameters and FLOPs. Moreover, MR-Occ demonstrates superior performance on the SemanticKITTI dataset, further validating its effectiveness and generalizability across diverse 3D semantic occupancy benchmarks.

Konyul Park, Yecheol Kim, Junho Koh et al.

Developing high-performance, real-time architectures for LiDAR-based 3D object detectors is essential for the successful commercialization of autonomous vehicles. Pillar-based methods stand out as a practical choice for onboard deployment due to their computational efficiency. However, despite their efficiency, these methods can sometimes underperform compared to alternative point encoding techniques such as Voxel-encoding or PointNet++. We argue that current pillar-based methods have not sufficiently captured the fine-grained distributions of LiDAR points within each pillar structure. Consequently, there exists considerable room for improvement in pillar feature encoding. In this paper, we introduce a novel pillar encoding architecture referred to as Fine-Grained Pillar Feature Encoding (FG-PFE). FG-PFE utilizes Spatio-Temporal Virtual (STV) grids to capture the distribution of point clouds within each pillar across vertical, temporal, and horizontal dimensions. Through STV grids, points within each pillar are individually encoded using Vertical PFE (V-PFE), Temporal PFE (T-PFE), and Horizontal PFE (H-PFE). These encoded features are then aggregated through an Attentive Pillar Aggregation method. Our experiments conducted on the nuScenes dataset demonstrate that FG-PFE achieves significant performance improvements over baseline models such as PointPillar, CenterPoint-Pillar, and PillarNet, with only a minor increase in computational overhead.

Jisong Kim, Geonho Bang, Kwangjin Choi et al.

In this paper, we present a novel point generation model, referred to as Pillar-based Point Generation Network (PillarGen), which facilitates the transformation of point clouds from one domain into another. PillarGen can produce synthetic point clouds with enhanced density and quality based on the provided input point clouds. The PillarGen model performs the following three steps: 1) pillar encoding, 2) Occupied Pillar Prediction (OPP), and 3) Pillar to Point Generation (PPG). The input point clouds are encoded using a pillar grid structure to generate pillar features. Then, OPP determines the active pillars used for point generation and predicts the center of points and the number of points to be generated for each active pillar. PPG generates the synthetic points for each active pillar based on the information provided by OPP. We evaluate the performance of PillarGen using our proprietary radar dataset, focusing on enhancing the density and quality of short-range radar data using the long-range radar data as supervision. Our experiments demonstrate that PillarGen outperforms traditional point upsampling methods in quantitative and qualitative measures. We also confirm that when PillarGen is incorporated into bird's eye view object detection, a significant improvement in detection accuracy is achieved.

Taein Son, Soo Won Seo, Jisong Kim et al.

Video Action Detection (VAD) entails localizing and categorizing action instances within videos, which inherently consist of diverse information sources such as audio, visual cues, and surrounding scene contexts. Leveraging this multi-modal information effectively for VAD poses a significant challenge, as the model must identify action-relevant cues with precision. In this study, we introduce a novel multi-modal VAD architecture, referred to as the Joint Actor-centric Visual, Audio, Language Encoder (JoVALE). JoVALE is the first VAD method to integrate audio and visual features with scene descriptive context sourced from large-capacity image captioning models. At the heart of JoVALE is the actor-centric aggregation of audio, visual, and scene descriptive information, enabling adaptive integration of crucial features for recognizing each actor's actions. We have developed a Transformer-based architecture, the Actor-centric Multi-modal Fusion Network, specifically designed to capture the dynamic interactions among actors and their multi-modal contexts. Our evaluation on three prominent VAD benchmarks, including AVA, UCF101-24, and JHMDB51-21, demonstrates that incorporating multi-modal information significantly enhances performance, setting new state-of-the-art performances in the field.

Junyong Yun, Jungho Kim, ByungHyun Lee et al.

Imitation learning (IL) is widely used for motion planning in autonomous driving due to its data efficiency and access to real-world driving data. For safe and robust real-world driving, IL-based planning requires capturing the complex driving contexts inherent in real-world data and enabling context-adaptive decision-making, rather than relying solely on expert trajectory imitation. In this paper, we propose CarPLAN, a novel IL-based motion planning framework that explicitly enhances driving context understanding and enables adaptive planning across diverse traffic scenarios. Our contributions are twofold: We introduce Displacement-Aware Predictive Encoding (DPE) to improve the model's spatial awareness by predicting future displacement vectors between the Autonomous Vehicle (AV) and surrounding scene elements. This allows the planner to account for relational spacing when generating trajectories. In addition to the standard imitation loss, we incorporate an augmented loss term that captures displacement prediction errors, ensuring planning decisions consider relative distances from other agents. To improve the model's ability to handle diverse driving contexts, we propose Context-Adaptive Multi-Expert Decoder (CMD), which leverages the Mixture of Experts (MoE) framework. CMD dynamically selects the most suitable expert decoders based on scene structure at each Transformer layer, enabling adaptive and context-aware planning in dynamic environments. We evaluate CarPLAN on the nuPlan benchmark and demonstrate state-of-the-art performance across all closed-loop simulation metrics. In particular, CarPLAN exhibits robust performance on challenging scenarios such as Test14-Hard, validating its effectiveness in complex driving conditions. Additional experiments on the Waymax benchmark further demonstrate its generalization capability across different benchmark settings.

Jungho Kim, Jiyong Oh, Seunghoon Yu et al.

The end-to-end (E2E) paradigm, which maps sensor inputs directly to driving decisions, has recently attracted significant attention due to its unified modeling capability and scalability. However, ensuring safety in this unified framework remains one of the most critical challenges. In this work, we propose SafeDrive, an E2E planning framework designed to perform explicit and interpretable safety reasoning through a trajectory-conditioned Sparse World Model. SafeDrive comprises two complementary networks: the Sparse World Network (SWNet) and the Fine-grained Reasoning Network (FRNet). SWNet constructs trajectory-conditioned sparse worlds that simulate the future behaviors of critical dynamic agents and road entities, providing interaction-centric representations for downstream reasoning. FRNet then evaluates agent-specific collision risks and temporal adherence to drivable regions, enabling precise identification of safety-critical events across future timesteps. SafeDrive achieves state-of-the-art performance on both open-loop and closed-loop benchmarks. On NAVSIM, it records a PDMS of 91.6 and an EPDMS of 87.5, with only 61 collisions out of 12,146 scenarios (0.5%). On Bench2Drive, SafeDrive attains a 66.8% driving score.

Geonho Bang, Minjae Seong, Jisong Kim et al.

Radar-camera fusion methods have emerged as a cost-effective approach for 3D object detection but still lag behind LiDAR-based methods in performance. Recent works have focused on employing temporal fusion and Knowledge Distillation (KD) strategies to overcome these limitations. However, existing approaches have not sufficiently accounted for uncertainties arising from object motion or sensor-specific errors inherent in radar and camera modalities. In this work, we propose RCTDistill, a novel cross-modal KD method based on temporal fusion, comprising three key modules: Range-Azimuth Knowledge Distillation (RAKD), Temporal Knowledge Distillation (TKD), and Region-Decoupled Knowledge Distillation (RDKD). RAKD is designed to consider the inherent errors in the range and azimuth directions, enabling effective knowledge transfer from LiDAR features to refine inaccurate BEV representations. TKD mitigates temporal misalignment caused by dynamic objects by aligning historical radar-camera BEV features with current LiDAR representations. RDKD enhances feature discrimination by distilling relational knowledge from the teacher model, allowing the student to differentiate foreground and background features. RCTDistill achieves state-of-the-art radar-camera fusion performance on both the nuScenes and View-of-Delft (VoD) datasets, with the fastest inference speed of 26.2 FPS.

Changwon Kang, Jisong Kim, Hongjae Shin et al.

The goal of multi-task learning is to learn to conduct multiple tasks simultaneously based on a shared data representation. While this approach can improve learning efficiency, it may also cause performance degradation due to task conflicts that arise when optimizing the model for different objectives. To address this challenge, we introduce MAESTRO, a structured framework designed to generate task-specific features and mitigate feature interference in multi-task 3D perception, including 3D object detection, bird's-eye view (BEV) map segmentation, and 3D occupancy prediction. MAESTRO comprises three components: the Class-wise Prototype Generator (CPG), the Task-Specific Feature Generator (TSFG), and the Scene Prototype Aggregator (SPA). CPG groups class categories into foreground and background groups and generates group-wise prototypes. The foreground and background prototypes are assigned to the 3D object detection task and the map segmentation task, respectively, while both are assigned to the 3D occupancy prediction task. TSFG leverages these prototype groups to retain task-relevant features while suppressing irrelevant features, thereby enhancing the performance for each task. SPA enhances the prototype groups assigned for 3D occupancy prediction by utilizing the information produced by the 3D object detection head and the map segmentation head. Extensive experiments on the nuScenes and Occ3D benchmarks demonstrate that MAESTRO consistently outperforms existing methods across 3D object detection, BEV map segmentation, and 3D occupancy prediction tasks.

Junho Koh, Youngwoo Lee, Jungho Kim et al.

Multi-view camera-based 3D perception can be conducted using bird's eye view (BEV) features obtained through perspective view-to-BEV transformations. Several studies have shown that the performance of these 3D perception methods can be further enhanced by combining sequential BEV features obtained from multiple camera frames. However, even after compensating for the ego-motion of an autonomous agent, the performance gain from temporal aggregation is limited when combining a large number of image frames. This limitation arises due to dynamic changes in BEV features over time caused by object motion. In this paper, we introduce a novel temporal 3D perception method called OnlineBEV, which combines BEV features over time using a recurrent structure. This structure increases the effective number of combined features with minimal memory usage. However, it is critical to spatially align the features over time to maintain strong performance. OnlineBEV employs the Motion-guided BEV Fusion Network (MBFNet) to achieve temporal feature alignment. MBFNet extracts motion features from consecutive BEV frames and dynamically aligns historical BEV features with current ones using these motion features. To enforce temporal feature alignment explicitly, we use Temporal Consistency Learning Loss, which captures discrepancies between historical and target BEV features. Experiments conducted on the nuScenes benchmark demonstrate that OnlineBEV achieves significant performance gains over the current best method, SOLOFusion. OnlineBEV achieves 63.9% NDS on the nuScenes test set, recording state-of-the-art performance in the camera-only 3D object detection task.

Minjae Lee, Seongmin Park, Hyungmin Kim et al.

3D object detection using point cloud (PC) data is essential for perception pipelines of autonomous driving, where efficient encoding is key to meeting stringent resource and latency requirements. PointPillars, a widely adopted bird's-eye view (BEV) encoding, aggregates 3D point cloud data into 2D pillars for fast and accurate 3D object detection. However, the state-of-the-art methods employing PointPillars overlook the inherent sparsity of pillar encoding where only a valid pillar is encoded with a vector of channel elements, missing opportunities for significant computational reduction. Meanwhile, current sparse convolution accelerators are designed to handle only element-wise activation sparsity and do not effectively address the vector sparsity imposed by pillar encoding. In this paper, we propose SPADE, an algorithm-hardware co-design strategy to maximize vector sparsity in pillar-based 3D object detection and accelerate vector-sparse convolution commensurate with the improved sparsity. SPADE consists of three components: (1) a dynamic vector pruning algorithm balancing accuracy and computation savings from vector sparsity, (2) a sparse coordinate management hardware transforming 2D systolic array into a vector-sparse convolution accelerator, and (3) sparsity-aware dataflow optimization tailoring sparse convolution schedules for hardware efficiency. Taped-out with a commercial technology, SPADE saves the amount of computation by 36.3--89.2\% for representative 3D object detection networks and benchmarks, leading to 1.3--10.9$\times$ speedup and 1.5--12.6$\times$ energy savings compared to the ideal dense accelerator design. These sparsity-proportional performance gains equate to 4.1--28.8$\times$ speedup and 90.2--372.3$\times$ energy savings compared to the counterpart server and edge platforms.

Junho Koh, Jaekyum Kim, Jinhyuk Yoo et al.

In this paper, we propose a new joint object detection and tracking (JoDT) framework for 3D object detection and tracking based on camera and LiDAR sensors. The proposed method, referred to as 3D DetecTrack, enables the detector and tracker to cooperate to generate a spatio-temporal representation of the camera and LiDAR data, with which 3D object detection and tracking are then performed. The detector constructs the spatio-temporal features via the weighted temporal aggregation of the spatial features obtained by the camera and LiDAR fusion. Then, the detector reconfigures the initial detection results using information from the tracklets maintained up to the previous time step. Based on the spatio-temporal features generated by the detector, the tracker associates the detected objects with previously tracked objects using a graph neural network (GNN). We devise a fully-connected GNN facilitated by a combination of rule-based edge pruning and attention-based edge gating, which exploits both spatial and temporal object contexts to improve tracking performance. The experiments conducted on both KITTI and nuScenes benchmarks demonstrate that the proposed 3D DetecTrack achieves significant improvements in both detection and tracking performances over baseline methods and achieves state-of-the-art performance among existing methods through collaboration between the detector and tracker.

ByeoungDo Kim, Seong Hyeon Park, Seokhwan Lee et al.

In this paper, we address the problem of predicting the future motion of a dynamic agent (called a target agent) given its current and past states as well as the information on its environment. It is paramount to develop a prediction model that can exploit the contextual information in both static and dynamic environments surrounding the target agent and generate diverse trajectory samples that are meaningful in a traffic context. We propose a novel prediction model, referred to as the lane-aware prediction (LaPred) network, which uses the instance-level lane entities extracted from a semantic map to predict the multi-modal future trajectories. For each lane candidate found in the neighborhood of the target agent, LaPred extracts the joint features relating the lane and the trajectories of the neighboring agents. Then, the features for all lane candidates are fused with the attention weights learned through a self-supervised learning task that identifies the lane candidate likely to be followed by the target agent. Using the instance-level lane information, LaPred can produce the trajectories compliant with the surroundings better than 2D raster image-based methods and generate the diverse future trajectories given multiple lane candidates. The experiments conducted on the public nuScenes dataset and Argoverse dataset demonstrate that the proposed LaPred method significantly outperforms the existing prediction models, achieving state-of-the-art performance in the benchmarks.

Sun Hong Lim, Sunwoo Kim, Byonghyo Shim et al.

In this paper, we propose a deep learning-based beam tracking method for millimeter-wave (mmWave)communications. Beam tracking is employed for transmitting the known symbols using the sounding beams and tracking time-varying channels to maintain a reliable communication link. When the pose of a user equipment (UE) device varies rapidly, the mmWave channels also tend to vary fast, which hinders seamless communication. Thus, models that can capture temporal behavior of mmWave channels caused by the motion of the device are required, to cope with this problem. Accordingly, we employa deep neural network to analyze the temporal structure and patterns underlying in the time-varying channels and the signals acquired by inertial sensors. We propose a model based on long short termmemory (LSTM) that predicts the distribution of the future channel behavior based on a sequence of input signals available at the UE. This channel distribution is used to 1) control the sounding beams adaptively for the future channel state and 2) update the channel estimate through the measurement update step under a sequential Bayesian estimation framework. Our experimental results demonstrate that the proposed method achieves a significant performance gain over the conventional beam tracking methods under various mobility scenarios.

Junho Koh, Jaekyum Kim, Younji Shin et al.

In this paper, we propose a new video object detector (VoD) method referred to as temporal feature aggregation and motion-aware VoD (TM-VoD), which produces a joint representation of temporal image sequences and object motion. The proposed TM-VoD aggregates visual feature maps extracted by convolutional neural networks applying the temporal attention gating and spatial feature alignment. This temporal feature aggregation is performed in two stages in a hierarchical fashion. In the first stage, the visual feature maps are fused at a pixel level via gated attention model. In the second stage, the proposed method aggregates the features after aligning the object features using temporal box offset calibration and weights them according to the cosine similarity measure. The proposed TM-VoD also finds the representation of the motion of objects in two successive steps. The pixel-level motion features are first computed based on the incremental changes between the adjacent visual feature maps. Then, box-level motion features are obtained from both the region of interest (RoI)-aligned pixel-level motion features and the sequential changes of the box coordinates. Finally, all these features are concatenated to produce a joint representation of the objects for VoD. The experiments conducted on the ImageNet VID dataset demonstrate that the proposed method outperforms existing VoD methods and achieves a performance comparable to that of state-of-the-art VoDs.

Jin Hyeok Yoo, Yecheol Kim, Jisong Kim et al.

In this paper, we propose a new deep architecture for fusing camera and LiDAR sensors for 3D object detection. Because the camera and LiDAR sensor signals have different characteristics and distributions, fusing these two modalities is expected to improve both the accuracy and robustness of 3D object detection. One of the challenges presented by the fusion of cameras and LiDAR is that the spatial feature maps obtained from each modality are represented by significantly different views in the camera and world coordinates; hence, it is not an easy task to combine two heterogeneous feature maps without loss of information. To address this problem, we propose a method called 3D-CVF that combines the camera and LiDAR features using the cross-view spatial feature fusion strategy. First, the method employs auto-calibrated projection, to transform the 2D camera features to a smooth spatial feature map with the highest correspondence to the LiDAR features in the bird's eye view (BEV) domain. Then, a gated feature fusion network is applied to use the spatial attention maps to mix the camera and LiDAR features appropriately according to the region. Next, camera-LiDAR feature fusion is also achieved in the subsequent proposal refinement stage. The camera feature is used from the 2D camera-view domain via 3D RoI grid pooling and fused with the BEV feature for proposal refinement. Our evaluations, conducted on the KITTI and nuScenes 3D object detection datasets demonstrate that the camera-LiDAR fusion offers significant performance gain over single modality and that the proposed 3D-CVF achieves state-of-the-art performance in the KITTI benchmark.

Jin Hyeok Yoo, Dongsuk Kum, Jun Won Choi

Convolutional neural network (CNN) has led to significant progress in object detection. In order to detect the objects in various sizes, the object detectors often exploit the hierarchy of the multi-scale feature maps called feature pyramid, which is readily obtained by the CNN architecture. However, the performance of these object detectors is limited since the bottom-level feature maps, which experience fewer convolutional layers, lack the semantic information needed to capture the characteristics of the small objects. In order to address such problem, various methods have been proposed to increase the depth for the bottom-level features used for object detection. While most approaches are based on the generation of additional features through the top-down pathway with lateral connections, our approach directly fuses multi-scale feature maps using bidirectional long short term memory (biLSTM) in effort to generate deeply fused semantics. Then, the resulting semantic information is redistributed to the individual pyramidal feature at each scale through the channel-wise attention model. We integrate our semantic combining and attentive redistribution feature network (ScarfNet) with baseline object detectors, i.e., Faster R-CNN, single-shot multibox detector (SSD) and RetinaNet. Our experiments show that our method outperforms the existing feature pyramid methods as well as the baseline detectors and achieve the state of the art performances in the PASCAL VOC and COCO detection benchmarks.

Jaekyum Kim, Junho Koh, Yecheol Kim et al.

The goal of multi-modal learning is to use complimentary information on the relevant task provided by the multiple modalities to achieve reliable and robust performance. Recently, deep learning has led significant improvement in multi-modal learning by allowing for the information fusion in the intermediate feature levels. This paper addresses a problem of designing robust deep multi-modal learning architecture in the presence of imperfect modalities. We introduce deep fusion architecture for object detection which processes each modality using the separate convolutional neural network (CNN) and constructs the joint feature map by combining the intermediate features from the CNNs. In order to facilitate the robustness to the degraded modalities, we employ the gated information fusion (GIF) network which weights the contribution from each modality according to the input feature maps to be fused. The weights are determined through the convolutional layers followed by a sigmoid function and trained along with the information fusion network in an end-to-end fashion. Our experiments show that the proposed GIF network offers the additional architectural flexibility to achieve robust performance in handling some degraded modalities, and show a significant performance improvement based on Single Shot Detector (SSD) for KITTI dataset using the proposed fusion network and data augmentation schemes.

Seong Hyeon Park, ByeongDo Kim, Chang Mook Kang et al.

In this paper, we propose a deep learning based vehicle trajectory prediction technique which can generate the future trajectory sequence of surrounding vehicles in real time. We employ the encoder-decoder architecture which analyzes the pattern underlying in the past trajectory using the long short-term memory (LSTM) based encoder and generates the future trajectory sequence using the LSTM based decoder. This structure produces the $K$ most likely trajectory candidates over occupancy grid map by employing the beam search technique which keeps the $K$ locally best candidates from the decoder output. The experiments conducted on highway traffic scenarios show that the prediction accuracy of the proposed method is significantly higher than the conventional trajectory prediction techniques.

ByeoungDo Kim, Chang Mook Kang, Seung Hi Lee et al.

In this paper, we propose an efficient vehicle trajectory prediction framework based on recurrent neural network. Basically, the characteristic of the vehicle's trajectory is different from that of regular moving objects since it is affected by various latent factors including road structure, traffic rules, and driver's intention. Previous state of the art approaches use sophisticated vehicle behavior model describing these factors and derive the complex trajectory prediction algorithm, which requires a system designer to conduct intensive model optimization for practical use. Our approach is data-driven and simple to use in that it learns complex behavior of the vehicles from the massive amount of trajectory data through deep neural network model. The proposed trajectory prediction method employs the recurrent neural network called long short-term memory (LSTM) to analyze the temporal behavior and predict the future coordinate of the surrounding vehicles. The proposed scheme feeds the sequence of vehicles' coordinates obtained from sensor measurements to the LSTM and produces the probabilistic information on the future location of the vehicles over occupancy grid map. The experiments conducted using the data collected from highway driving show that the proposed method can produce reasonably good estimate of future trajectory.

Hyunmin Chae, Chang Mook Kang, ByeoungDo Kim et al.

In this paper, we propose a new autonomous braking system based on deep reinforcement learning. The proposed autonomous braking system automatically decides whether to apply the brake at each time step when confronting the risk of collision using the information on the obstacle obtained by the sensors. The problem of designing brake control is formulated as searching for the optimal policy in Markov decision process (MDP) model where the state is given by the relative position of the obstacle and the vehicle's speed, and the action space is defined as whether brake is stepped or not. The policy used for brake control is learned through computer simulations using the deep reinforcement learning method called deep Q-network (DQN). In order to derive desirable braking policy, we propose the reward function which balances the damage imposed to the obstacle in case of accident and the reward achieved when the vehicle runs out of risk as soon as possible. DQN is trained for the scenario where a vehicle is encountered with a pedestrian crossing the urban road. Experiments show that the control agent exhibits desirable control behavior and avoids collision without any mistake in various uncertain environments.