Browsing University of Waterloo by Supervisor "Czarnecki, Krzysztof"
Now showing items 21-40 of 46
-
MLOD: A multi-view 3D object detection based on robust feature fusion method
(University of Waterloo, 2019-09-19)This thesis presents Multi-view Labelling Object Detector (MLOD). The detector takes an RGB image and a LIDAR point cloud as input and follows the two-stage object detection framework. A Region Proposal Network (RPN) ... -
Modeling and Reasoning with Multisets and Multirelations in Alloy
(University of Waterloo, 2017-01-17)Multisets and multirelations arise naturally in modeling; however, most modeling languages either have limited or completely lack support for multisets and multirelations. Alloy, for instance, is a lightweight relational ... -
Modeling the Effects of AUTOSAR Overhead on Automotive Application Software Timing and Schedulability
(University of Waterloo, 2016-10-14)AUTOSAR (AUTomotive Open System ARchitecture) provides an open and standardized E/E architecture to support modularity, transferability, reusability and scalability of the various components required to implement a function ... -
Motion Planning and Safety for Autonomous Driving
(University of Waterloo, 2019-12-11)This thesis discusses two different problems in motion planning for autonomous driving. The first is the problem of optimizing a lattice planner control set for any particular autonomous driving task, with the goal of ... -
Out-of-Distribution Detection for LiDAR-based 3D Object Detection
(University of Waterloo, 2022-01-18)3D object detection is an essential part of automated driving, and deep neural networks (DNNs) have achieved state-of-the-art performance for this task. However, deep models are notorious for assigning high confidence ... -
Passenger Response to Driving Style in an Autonomous Vehicle
(University of Waterloo, 2019-09-23)Despite rapid advancements in automated driving systems (ADS), current human-computer interaction research tends to focus more on the safety driver in lower level vehicles. The future of automated driving lies in higher ... -
Perception and Prediction in Multi-Agent Urban Traffic Scenarios for Autonomous Driving
(University of Waterloo, 2023-09-21)In multi-agent urban scenarios, autonomous vehicles navigate an intricate network of interactions with a variety of agents, necessitating advanced perception modeling and trajectory prediction. Research to improve perception ... -
Policy Extraction via Online Q-Value Distillation
(University of Waterloo, 2019-08-27)Recently, deep neural networks have been capable of solving complex control tasks in certain challenging environments. However, these deep learning policies continue to be hard to interpret, explain and verify which limits ... -
Quantitative Analyses of Software Product Lines
(University of Waterloo, 2022-01-18)A software product-line (SPL) is a family of related software systems that are jointly developed and reuse a set of shared assets. Each individual software system in an SPL is called a software product and includes a set ... -
Runtime Restriction of the Operational Design Domain: A Safety Concept for Automated Vehicles
(University of Waterloo, 2018-06-14)Automated vehicles need to operate safely in a wide range of environments and hazards. The complex systems that make up an automated vehicle must also ensure safety in the event of system failures. This thesis proposes an ... -
Safety-Oriented Stability Biases for Continual Learning
(University of Waterloo, 2020-01-24)Continual learning is often confounded by “catastrophic forgetting” that prevents neural networks from learning tasks sequentially. In the case of real world classification systems that are safety-validated prior to ... -
Scenario Modeling and Execution for Simulation Testing of Automated-Driving Systems
(University of Waterloo, 2022-09-28)Automated Driving Systems (ADS) have the potential to significantly impact the future of ground mobility. However, safety assurance is still a major obstacle. Field testing alone is impractical and simulation is required ... -
Sparse2SOAP: Domain Adaptation for LiDAR-Based 3D Object Detection
(University of Waterloo, 2023-05-25)In this work, we propose Sparse2SOAP, an extension of the previous work in Sparse2Dense that uses knowledge distillation in a teacher-student framework to densify 3D features, to enable its uses for cross-domain LiDAR-based ... -
Switching GAN-based Image Filters to Improve Perception for Autonomous Driving
(University of Waterloo, 2019-10-24)Autonomous driving holds the potential to increase human productivity, reduce accidents caused by human errors, allow better utilization of roads, reduce traffic accidents and congestion, free up parking space and provide ... -
Synthesis and Exploration of Multi-Level, Multi-Perspective Architectures of Automotive Embedded System
(University of Waterloo, 2016-08-16)In industry, evaluating candidate architectures of automotive embedded systems is routinely done during the design process. Today's engineers, however, are limited in the number of candidates that they are able to evaluate ... -
Towards a Better Understanding of Variability Evolution
(University of Waterloo, 2016-06-07)Highly-configurable software systems often leverage variability modeling to achieve systematical reuse and mass customization. Although facilitating variability management, variability models do not eliminate the variability ... -
Towards Object Re-identification from Point Clouds for 3D MOT
(University of Waterloo, 2023-04-21)This thesis studies the problem of object re-identification (ReID) in a 3D multi-object tracking (MOT) context, by learning to match pairs of objects from cropped (e.g., using their predicted 3D bounding boxes) point cloud ... -
Towards Pixel-Level OOD Detection for Semantic Segmentation
(University of Waterloo, 2019-08-30)There exists wide research surrounding the detection of out of distribution sample for image classification. Safety critical applications, such as autonomous driving, would benefit from the ability to localise the unusual ... -
Towards Synthetic Dataset Generation for Semantic Segmentation Networks
(University of Waterloo, 2019-09-23)Recent work in semantic segmentation research for autonomous vehicles has shifted towards multimodal techniques. The driving factor behind this is a lack of reliable and ample ground truth annotation data of real-world ... -
Training Reject-Classifiers for Out-of-distribution Detection via Explicit Boundary Sample Generation
(University of Waterloo, 2020-01-24)Discriminatively trained neural classifiers can be trusted only when the input data comes from the training distribution (in-distribution). Therefore, detecting out-of-distribution (OOD) samples is very important to avoid ...