Pedestrian Behavior Prediction for Automated Driving: Requirements, Metrics, and Relevant Features
Authors:
Michael Herman,
Jörg Wagner,
Vishnu Prabhakaran,
Nicolas Möser,
Hanna Ziesche,
Waleed Ahmed,
Lutz Bürkle,
Ernst Kloppenburg,
Claudius Gläser
Abstract:
Automated vehicles require a comprehensive understanding of traffic situations to ensure safe and anticipatory driving. In this context, the prediction of pedestrians is particularly challenging as pedestrian behavior can be influenced by multiple factors. In this paper, we thoroughly analyze the requirements on pedestrian behavior prediction for automated driving via a system-level approach. To t…
▽ More
Automated vehicles require a comprehensive understanding of traffic situations to ensure safe and anticipatory driving. In this context, the prediction of pedestrians is particularly challenging as pedestrian behavior can be influenced by multiple factors. In this paper, we thoroughly analyze the requirements on pedestrian behavior prediction for automated driving via a system-level approach. To this end we investigate real-world pedestrian-vehicle interactions with human drivers. Based on human driving behavior we then derive appropriate reaction patterns of an automated vehicle and determine requirements for the prediction of pedestrians. This includes a novel metric tailored to measure prediction performance from a system-level perspective. The proposed metric is evaluated on a large-scale dataset comprising thousands of real-world pedestrian-vehicle interactions. We furthermore conduct an ablation study to evaluate the importance of different contextual cues and compare these results to ones obtained using established performance metrics for pedestrian prediction. Our results highlight the importance of a system-level approach to pedestrian behavior prediction.
△ Less
Submitted 16 October, 2021; v1 submitted 15 December, 2020;
originally announced December 2020.
Revealing biases in the sampling of ecological interaction networks
Authors:
Marcus A. M. de Aguiar,
Erica A. Newman,
Mathias M. Pires,
Justin D. Yeakel,
David H. Hembry,
Laura Burkle,
Dominique Gravel,
Paulo R. Guimaraes Jr,
Jimmy O'Donnell,
Timothee Poisot,
Marie-Josee Fortin
Abstract:
The structure of ecological interactions is commonly understood through analyses of interaction networks. However, these analyses may be sensitive to sampling biases in both the interactors (the nodes of the network) and interactions (the links between nodes). These issues may affect the accuracy of empirically constructed ecological networks. We explore the properties of sampled ecological networ…
▽ More
The structure of ecological interactions is commonly understood through analyses of interaction networks. However, these analyses may be sensitive to sampling biases in both the interactors (the nodes of the network) and interactions (the links between nodes). These issues may affect the accuracy of empirically constructed ecological networks. We explore the properties of sampled ecological networks by simulating large-scale ecological networks with predetermined topologies, and sampling them with different mathematical procedures. Several types of modular networks were generated, intended to represent a wide variety of communities that vary in size and types of ecological interactions. We sampled these networks with different sampling designs that may be encountered in field experiments. The observed networks generated by each sampling process were analyzed with respect to number and size of components. We show that the sampling effort needed to estimate underlying network properties depends both on the sampling design and on network topology. Networks with random or scale-free modules require more complete sampling compared to networks whose modules are nested or bipartite. Overall, the structure of nested modules was the easiest to detect, regardless of sampling design. Sampling according to species degree was consistently found to be the most accurate strategy to estimate network structure. Conversely, sampling according to module results in an accurate view of certain modules, but fails to provide a global picture of the underlying network. We recommend that these findings are incorporated into the design of projects aiming to characterize large networks of species interactions in the field, to reduce sampling biases. The software scripts developed to construct and sample networks are provided for further explorations of network structure and comparisons to real interaction data.
△ Less
Submitted 3 August, 2017;
originally announced August 2017.
