Estimating surrounding objects and obstacles by processing sensor data is essential for safe autonomous driving. Grid-based approaches discretize the environment into grid cells, which implicitly solves the data association between measurement data and the filtered state on this grid representation. Recent approaches estimate, in addition to occupancy probabilities, cell velocity distributions using a low-level particle filter. Measured occupancy can thus be classified as static or dynamic, whereby a subsequent tracking of moving objects can be limited to dynamic cells. However, the data association between those cells and multiple predicted objects that are close to each other remains a challenge. In this work, we propose a new association approach in that context. Our main idea is that particles of the underlying low-level particle filter are linked to those high-level objects, i.e., an object label is attached to each particle. Cells are thus associated to objects by evaluating the particle label distribution of that cell. In addition, a subsequent clustering is performed, in which multiple clusters of an object are extracted and finally checked for plausibility to further increase the robustness. Our approach is evaluated with real sensor data in challenging scenarios with occlusions and dense traffic.
While self-supervised monocular depth estimation in driving scenarios has achieved comparable performance to supervised approaches, violations of the static world assumption can still lead to erroneous depth predictions of traffic participants, posing a potential safety issue. In this paper, we present R4Dyn, a novel set of techniques to use cost-efficient radar data on top of a self-supervised depth estimation framework. In particular, we show how radar can be used during training as weak supervision signal, as well as an extra input to enhance the estimation robustness at inference time. Since automotive radars are readily available, this allows to collect training data from a variety of existing vehicles. Moreover, by filtering and expanding the signal to make it compatible with learning-based approaches, we address radar inherent issues, such as noise and sparsity. With R4Dyn we are able to overcome a major limitation of self-supervised depth estimation, i.e. the prediction of traffic participants. We substantially improve the estimation on dynamic objects, such as cars by 37% on the challenging nuScenes dataset, hence demonstrating that radar is a valuable additional sensor for monocular depth estimation in autonomous vehicles. Additionally, we plan on making the code publicly available.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.