Muhammad Haris

• Implement advanced vision-based visual localization that can cope with long-term changes in a scene. • Integrate computationally efficient algorithms suitable for robots with low-cost embedded hardware (i.e., lawnmowers). • Plan and conduct simulation studies; analyze the implemented concepts. • Maintain a long-term image database with an outdoor robot. • Visualized, analyzed, and documented the scientific results. • Contributed to the development of 6 publications and 2 under review.

• Demonstrated proficiency with visual self-localization, navigation, and outdoor robotics. • Reviewed literature on deep learning for visual localization. • Compiled results of end-to-end learning for localization on custom datasets.

• Garnered experience with live navigation on a service robot. • Implemented a system that exploits the properties of slowness learning to perform robot navigation in the representation space. • Executed system testing on a lawnmower robot in an indoor environment. • Created planning-free robust robot navigation in open spaces and around obstacles.

• Completed condition invariant visual localization using slow feature analysis. • Reviewed and interpreted existing literature on place recognition and spatial learning. • Implemented a hierarchical learning system that estimates a camera's position from an image. • Extended the learning system to incorporate temporal loop-closures for increasing robustness. • Achieved a robust system that learns increasingly invariant spatial representation for long-term outdoor localization.

• Measured and documented respiration rate using thermal video imaging. • Built a prototype using an Android tablet and Flir-One thermal camera. • Coordinated the implementation of the front-end GUI in Android. • Conducted offline signal processing in MATLAB. • Evaluated the system using a conventional reference system.

• Participated in the development of software for displaying energy flows on display in the “12V Dual Storage” series vehicle. • Provided insights to support software architecture development. • Programmed desired functionalities and animation. • Performed tested and integration into the vehicle.

• Facilitated the delivery of software training to 50+ end-users. • Identified, documented, and reported software issues to developers.

Image Processing/Machine Learning

Computer Engineering