A data-driven and application-aware approach to sensory system calibration in an autonomous vehicle

Nowicki, Michał

doi:10.1016/j.measurement.2022.111002

Cited by 4 publications

(1 citation statement)

References 44 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Finally, we evaluated the performance of the different neural models applied for keypoints extraction directly for the objective of camera pose accuracy. The ground truth trajectories were provided by DGPS (2 cm of accuracy), while the camera and DGPS receiver were calibrated using an optimisationbased procedure developed specifically for this project [34]. The 3-D model of the charging station was obtained using a geodetic laser scanning system.…”

Section: A Evaluation Proceduresmentioning

confidence: 99%

Geometry-Aware Keypoint Network: Accurate Prediction of Point Features in Challenging Scenario

Nowak¹,

Skrzypczyński²

2022

Annals of Computer Science and Information Systems

View full text Add to dashboard Cite

In this paper, we consider a challenging scenario of localising a camera with respect to a charging station for electric buses. In this application, we face a number of problems, including a substantial scale change as the bus approaches the station, and the need to detect keypoints on a weakly textured object in a wide range of lighting and weather conditions. Therefore, we use a deep convolutional neural network to detect the features, while retaining a conventional procedure for pose estimation with 2D-to-3D associations. We leverage here the backbone of HRNet, a state-of-the-art network used for detection of feature points in human pose recognition, and we further improve the solution adding constraints that stem from the known scene geometry. We incorporate the reprojection-based geometric priors in a novel loss function for HRNet training and use the object geometry to construct sanity checks in postprocessing. Moreover, we demonstrate that our Geometry-Aware Keypoint Network yields feasible estimates of the geometric uncertainty of point features. The proposed architecture and solutions are tested on a large dataset of images and trajectories collected with a real city bus and charging station under varying environmental conditions.

show abstract

Section: A Evaluation Proceduresmentioning

confidence: 99%

Geometry-Aware Keypoint Network: Accurate Prediction of Point Features in Challenging Scenario

Nowak¹,

Skrzypczyński²

2022

Annals of Computer Science and Information Systems

View full text Add to dashboard Cite

show abstract

Robust Depth-Aided RGBD-Inertial Odometry for Indoor Localization

Zhao

Wang

et al. 2023

Measurement

View full text Add to dashboard Cite

Sensor-Based Prognostic Health Management of Advanced Driver Assistance System for Autonomous Vehicles: A Recent Survey

et al. 2022

View full text Add to dashboard Cite

Recently, the advanced driver assistance system (ADAS) of autonomous vehicles (AVs) has offered substantial benefits to drivers. Improvement of passenger safety is one of the key factors for evolving AVs. An automated system provided by the ADAS in autonomous vehicles is a salient feature for passenger safety in modern vehicles. With an increasing number of electronic control units and a combination of multiple sensors, there are now sufficient computing aptitudes in the car to support ADAS deployment. An ADAS is composed of various sensors: radio detection and ranging (RADAR), cameras, ultrasonic sensors, and LiDAR. However, continual use of multiple sensors and actuators of the ADAS can lead to failure of AV sensors. Thus, prognostic health management (PHM) of ADAS is important for smooth and continuous operation of AVs. The PHM of AVs has recently been introduced and is still progressing. There is a lack of surveys available related to sensor-based PHM of AVs in the literature. Therefore, the objective of the current study was to identify sensor-based PHM, emphasizing different fault identification and isolation (FDI) techniques with challenges and gaps existing in this field.

show abstract

A data-driven and application-aware approach to sensory system calibration in an autonomous vehicle

Cited by 4 publications

References 44 publications

Geometry-Aware Keypoint Network: Accurate Prediction of Point Features in Challenging Scenario

Geometry-Aware Keypoint Network: Accurate Prediction of Point Features in Challenging Scenario

Robust Depth-Aided RGBD-Inertial Odometry for Indoor Localization

Sensor-Based Prognostic Health Management of Advanced Driver Assistance System for Autonomous Vehicles: A Recent Survey

Contact Info

Product

Resources

About