Multi-sensor data fusion and quality evaluation in challenging weather environment

A key factor to deliver reliable intelligent vehicles is the proper exploitation of data gathered from the different sensors equipping the vehicle. To do so, a data fusion algorithm is applied. However, the reliability of the sensors can change (for example due to weather conditions), therefore, a solution to evaluate the quality of the data gathered from each sensor must be investigated, to automatically adapt the data fusion algorithm and avoid the use of less reliable sensors or credible data or information these sensors provide directly or they derive through processing techniques. Sensors are becoming more and more complex and are no longer specific to only one type of measure or characteristic/feature. They can be used to measure directly or indirectly (through processing) different characteristics/features (such as velocity, position, size, etc.) of the observed environment. Each sensor category is better adapted to capture a specific characteristic/feature but in a meantime, the data they provide can also be exploited to identify (in a less reliable/credible way) other features. T

Faculty Supervisor:

Henry Leung

Student:

Abdessattar Hayouni

Partner:

Thales solutions numériques Inc

Discipline:

Engineering - computer / electrical

Sector:

Transportation and warehousing

University:

University of Calgary