Dense, volumetric maps are essential to enable robot navigation and interaction with the environment. To achieve low latency, dense maps are typically computed onboard the robot, often on computationally constrained hardware. Previous works leave a gap between CPU-based systems for robotic mapping which, due to computation constraints, limit map resolution or scale, and GPU-based reconstruction systems which omit features that are critical to robotic path planning, such as computation of the Euclidean Signed Distance Field (ESDF). We introduce a library, nvblox, that aims to fill this gap, by GPU-accelerating robotic volumetric mapping. Nvblox delivers a significant performance improvement over the state of the art, achieving up to a 177× speed-up in surface reconstruction, and up to a 31× improvement in distance field computation, and is available open-source1.

Domain adaptation is an important task to enable learning when labels are scarce. While most works focus only on the image modality, there are many important multi-modal datasets. In order to leverage multi-modality for domain adaptation, we propose cross-modal learning, where we enforce consistency between the predictions of two modalities via mutual mimicking. We constrain our network to make correct predictions on labeled data and consistent predictions across modalities on unlabeled target-domain data. Experiments in unsupervised and semi-supervised domain adaptation settings prove the effectiveness of this novel domain adaptation strategy. Specifically, we evaluate on the task of 3D semantic segmentation from either the 2D image, the 3D point cloud or from both. We leverage recent driving datasets to produce a wide variety of domain adaptation scenarios including changes in scene layout, lighting, sensor setup and weather, as well as the synthetic-to-real setup. Our method significantly improves over previous uni-modal adaptation baselines on all adaption scenarios. Code will be made available upon publication.

Unsupervised Domain Adaptation (UDA) is crucial to tackle the lack of annotations in a new domain. There are many multi-modal datasets, but most UDA approaches are uni-modal. In this work, we explore how to learn from multi-modality and propose cross-modal UDA (xMUDA) where we assume the presence of 2D images and 3D point clouds for 3D semantic segmentation. This is challenging as the two input spaces are heterogeneous and can be impacted differently by domain shift. In xMUDA, modalities learn from each other through mutual mimicking, disentangled from the segmentation objective, to prevent the stronger modality from adopting false predictions from the weaker one. We evaluate on new UDA scenarios including day-to-night, country-to-country and dataset-to-dataset, leveraging recent autonomous driving datasets. xMUDA brings large improvements over uni-modal UDA on all tested scenarios, and is complementary to state-of-the-art UDA techniques. Code is available at https://github.com/valeoai/xmuda.

Reinforcement Learning (RL) aims at learning an optimal behavior policy from its own experiments and not rule-based control methods. However, there is no RL algorithm yet capable of handling a task as difficult as urban driving. We present a novel technique, coined implicit affordances, to effectively leverage RL for urban driving thus including lane keeping, pedestrians and vehicles avoidance, and traffic light detection. To our knowledge we are the first to present a successful RL agent handling such a complex task especially regarding the traffic light detection. Furthermore, we have demonstrated the effectiveness of our method by winning the Camera Only track of the CARLA challenge.

Autonomous vehicles require clear road markings and a high-level quality of infrastructure. This paper addresses road course detection problem in non cooperative environments (i.e. absence or poor quality of road-marking, working zones, etc.). To cope with visual lane detection challenges in these difficult scenarios, we propose a virtual lane generation system to provide a comfortable and safe ride. Based on Lidar sensor, the dual-level virtual lane system consists of the combination of two blocks: the first constructs virtual lanes based on independent road-borders detection, while the second level uses dynamic objects detection and their trajectories in order to estimate the lane parameters. Furthermore, the system is able to anticipate road infrastructures thanks to the independent detection of road borders. Thus we are able to manage difficult use cases such as bifurcations and exit lanes without cartography. The performance is tested through extensive experiments with Cruise4U Valeo self-driving cars on highway and beltway roads. Experimental results demonstrate the accurate and robust performance of the proposed system.

Convolutional neural networks are designed for dense data, but vision data is often sparse (stereo depth, point clouds, pen stroke, etc.). We present a method to handle sparse depth data with optional dense RGB, and accomplish depth completion and semantic segmentation changing only the last layer. Our proposal efficiently learns sparse features without the need of an additional validity mask. We show how to ensure network robustness to varying input sparsities. Our method even works with densities as low as 0.8% (8 layer lidar), and outperforms all published state-of-the-art on the Kitti depth completion benchmark.

Convolutional neural networks are commonly used to control the steering angle for autonomous cars. Most of the time, multiple long range cameras are used to generate lateral failure cases. In this paper we present a novel model to generate this data and label augmentation using only one short range fisheye camera. We present our simulator and how it can be used as a consistent metric for lateral end-to-end control evaluation. Experiments are conducted on a custom dataset corresponding to more than 10000 km and 200 hours of open road driving. Finally we evaluate this model on real world driving scenarios, open road and a custom test track with challenging obstacle avoidance and sharp turns. In our simulator based on real-world videos, the final model was capable of more than 99% autonomy on urban road.