Search

Detecting object locations and semantic classes in an image, such as traffic signs, traffic lights, and guide signs, is the crucial problem for autonomous driving, known as object detection. However, stable object detection in complex real-world environments, such as urban environments, is still challenging because of clutter, time of day, blur etc., even with modern deep convolutional neural networks (DCNNs). On the other hand, a high definition (HD) map is a pre-built information resource for autonomous driving tasks, especially for controls. Besides controls, HD map utilization for detection tasks has been gaining attention in recent years, enabling us to stabilize detection even in complex real-world environments. However, it is challenging to use object information from an HD map as detection directly because the self-localization error affects the transformed object locations on the image coordinate system from the HD map's coordinate system. This paper explores incorporating HD map information into deep feature maps of a DCNN-based model. Two proposed modules implicitly make the feature extraction efficient and stable by utilizing HD map information. As a result of the experiments, the proposed module improved a modern model for challenging images of the urban area Shinjuku by 37 % in mAP, even in self-localization errors. [ABSTRACT FROM AUTHOR]

Deep reinforcement learning (DRL) has been successfully adopted in many tasks, such as autonomous driving and gaming, to achieve or surpass human-level performance. This paper proposes a DRL-based trajectory planner for automated parking systems (APS). A thorough review of literature in this field is presented. A simulation study is conducted to investigate the trajectory planning performance of the parking agent for: (i) different neural-network architectures; (ii) different training set-ups; (iii) efficacy of human-demonstration. Real-time capability of the proposed planner on various embedded hardware platforms is also discussed by the paper, showing promising performance. Insights of the use of DRL for APS are concluded at the end of the paper. [ABSTRACT FROM AUTHOR]

Autonomous driving approaches today are mainly based on perception-planning-action modular pipelines and the End2End paradigm respectively. The End2End paradigm is a strategy that directly maps raw sensor data to vehicle control actions. This strategy is very promising and appealing because complex module design and cumbersome data labeling are avoided. Since this approach lacks a degree of interpretability, safety and practicability. we propose an autonomous driving approach based on trajectory learning using deep neural networks in this paper. In comparison to End2End algorithm, it is found that the trajectory learning algorithm performs better in autonomous driving. As for trajectory learning algorithm, the CNN_Raw-RNN network structure is established, which is verified to be more effective than the original CNN_LSTM network structure. Besides, we propose an autonomous driving architecture of a pilot and copilot combination. The pilot is responsible for trajectory prediction via imitation learning with labeled driving trajectories, while the copilot is a safety module that is employed to verify the effectiveness of the vehicle trajectory by the results of the semantic segmentation auxiliary task. The proposed autonomous driving architecture is verified with a real car on urban roads without manual intervention within 40 km. [ABSTRACT FROM AUTHOR]

The smart regenerative braking system in electric vehicles implements automatic control of the regeneration torque of motor to improve driver's comfort and energy efficiency. To apply this system, the accurate prediction of the vehicle deceleration states is the preliminary step to reflect the driver's behaviors. In this paper, we proposed a vehicle deceleration prediction model via deep neural network, which consists of a sequential recurrent neural network model with long-short term memory cell and a two-layer conventional neural network model. This model accommodates the physical constraint to designate the vehicle stop location in front of the traffic signals. The model is trained by vehicle experiment data with three drivers through the hyper-parameter optimization method. Using this model, the deceleration characteristics are characterized by two explicit parameters such that deceleration point, maximum point according to the initial slope and the shape of the braking profile. Using these two parameters as clustering variables through a K-means clustering method, the deceleration types are classified. These deceleration types to the input to the prediction model results in higher prediction accuracy of the vehicle states. The driving style of the three drivers at braking situations is analyzed according to the deceleration types as well. [ABSTRACT FROM AUTHOR]