Formulating and solving the multi-calendar resumable and non-resumable Naval Surface Ship Work Period Problem

This project deals with the development of a new mathematical model and a fast solution method to optimize the multi-calendar naval surface ship refit scheduling problem with resumable and non-resumable operations. The Naval Surface Ship Work Period Problem (NSWPP) is a highly complex resource-constrained project scheduling problem (RCPSP) with many work orders that are equivalent to small projects.

Integrity Monitoring of Motion Estimation and Hazard Detection Algorithms in Environmentally-Impacted Scenarios

Imagine some of the difficult driving conditions experienced by vehicle operators. In these conditions, the sun might be blindingly bright, or the snow might obfuscate what is going on around the vehicle. Surprisingly, the sensors used by autonomous vehicles to understand the environment they are in suffer from similar effects. As a field, robotics has yet to tackle integrity monitoring of the sensors used in autonomous applications.

Optimal and heuristic optimization methods for large-scale naval refit operations

This project deals with the development of new mathematical models and solution methods to optimize naval surface ship refit operations. The Naval Surface Ship Work Period Problem (NSWPP) is a highly complex resource-constrained project scheduling problem (RCPSP) with many work orders that are equivalent to small projects.

Research into Convolutional Neural Network (CNN) Explainability

Machine Learning is advancing at an astounding rate. It is powered by complex models such as deep neural networks (DNNs). These models have a wide range of real-world applications, in fields like Computer Vision, Natural Language Processing, Information Retrieval and others. But Machine Learning is not without some serious limitations and drawbacks. The most serious one is the lack of transparency in their inferences, which works against relying completely in these models and leaves users with little understanding of how particular decisions are made.

Automatic Adjustment of Photometric Camera Parameters to Improve Visual Motion Estimation

Cameras are a fundamental component of modern robotic systems. As robots have become relied upon for safety-critical tasks, the need for robust sensing is apparent. Cameras have a major limitation, compared to other sensors such as LIDAR, in high-dynamic-range environments where lighting conditions rapidly change. These changes can cause visual navigation algorithms to struggle and, in some cases, fail in instances where images become severely under- or overexposed.

Constrained Batch Estimation for Train Positioning with Inertial Sensors

Similar to current efforts in the automotive industry, there is a substantial interest in developing fully autonomous trains. One of the key steps towards enabling autonomous operation is being able to accurately estimate train position and speed. In addition to this, better estimates will also increase safety and reduce distance between trains, allowing more frequent trains in peak hours. The current project deals with a method of estimating the velocity and position without using GPS measurements, which is the standard method.

Integrated Configurable Power Input/Output Systems for Avionic Applications

Thales Canada develops control systems for avionics applications, which operate in harsh environments that may compromise the functionality of very high density chips. The company needs to develop a generic power interface for different avionics applications with a high level of criticality. However, such circuitry requires a lot of space on printed circuit boards when implemented as discrete components.

Constrained Kalman filtering for train position estimation

Just like for the automotive industry, there is growing interest in the development of fully autonomous trains. One of the key steps in the creation of a fully autonomous solution is optaining an accurate estimate of the train position and velocity. Accurate estimates are critical component of the train safety during operation and better estimates allow more trains to operate safely on the same track. The current project deals with trains operating in areas without GPS coverage, such as subways, and so accurate position measurements cannot be obtained as frequently.

Assessment of deep learning for analyzing radar signals in maritime environment

The proposed internships aim at investigating the relevance of deep learning (DL) techniques for target detection in radar data processing. More specifically, we are looking to demonstrate the feasibility of DL techniques to deal with unusual types of data (i.e., radar data) in situations where an well performing processing with classical techniques is a challenge (e.g., detection of objects in noisy scenes from a maritime environment caused by the interference produced by the reflection of the radar waves on the sea).

Anomaly Detection in Land Vehicle Traffic Activity

This project’s objective is to develop a capability to detect and describe anomalous situations in ground vehicle traffic. Anomalous situations are described as substantial/important changes from the traffic frequently observed for a particular route and/or time. In this sense, anomaly can be quantitatively measured by the degree of predictability of current traffic given historical observations. In the use case of interest, information from traffic will be captured from a GMTI sensor performing recurrent surveillances (1-3 hours per day, multiple days per week) over the same area.