Automated identification of the relationships between traffic actors and surrounding objects, in order to describe their behavior and predict their intentions, has become the focus of increasing attention in the field of autonomous driving. Therefore, in this work, we propose a Road Scene Graphs-Graph Convolutional Network (RSG-GCN) as a novel, graphbased model for predicting the topological graph structure of a given traffic scene. The status of the actors and HD map information are integrated as prior knowledge, allowing the edges linking the actor nodes to capture potential semantic relationships, such as "vehicle approaching pedestrian" and "pedestrian waiting at intersection". To train this model, we created our own RSG dataset, as well as a relational dataset and benchmark derived from nuScenes. Our extensive range of experiments demonstrate that our model can more accurately predict semantic relationships and behavior in a given traffic scene than other popular traffic scene prediction models. In particular, regarding the use of HD map prior knowledge, we found that the resulting increase in accuracy significantly outweighs performance loss caused by the increase in graph size. The downstream applications of RSG include traffic scene retrieval and synthetic traffic scene generation, which are briefly described.
Robot dynamic modeling and parameter identification are essential for many analyses. Highfidelity multi-body dynamics simulators can model the robot's dynamic behavior, but they can't identify the robot non-linear dynamic model, which is needed for controller design. This study proposes a three-step machine learning framework for obtaining the dynamic equations of serial manipulators from data. This framework consists of three steps. Initially, A library of candidate functions is first constructed, together with a data set based on robot unforced response. Secondly, the best models that can represent dynamic systems for each candidate function utilizing the training data set are then obtained using the SINDy-PI algorithm and Akaike Information Criterion (AIC). Through the MSE and test data, those best models will be reduced to get the functions that best describe the dynamic system. Finally, the dynamic equations that characterize the system are derived using the SINDy algorithm, including the applied force or torque. The technique was tested on three case studies -double pendulum, two-link of KUKA robot, and two-link of Stanford robot. The framework correctly determines the structure of the dynamic system and simultaneously accurately identified its parameters. The framework was able to deal with an ill-conditioned system of equations that arises for complex robot configuration.
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.