Localization and Mapping for Mobile Robots


Real-Time Mapping

Student: Joey Wilson

For a robot to act, it must have a robust model of the world which adapts in real-time to its surroundings. In this project, we study and develop algorithms for mapping with high levels of scene understanding, to enable complex autonomy in self-driving cars and more. We also generate ROS packages for our software, and test on real robots in off-road and on-road driving environments.


Bridging Bayesian Methods and Deep Learning for Semantic Mapping

Student: Joey Wilson

Robotic perception is at a crossroads between classical, hand-crafted methods, and modern, implicit architectures. While probabilistic methods offer quantifiable uncertainty and the ability to transfer between data sets due to robust algorithms, they are generally slow and hand-crafted. In contrast, modern deep learning methods trade more efficient and optimized results for less reliability when exposed to new data due to operating in an implicit space. We propose to bridge the gap between Bayesian and deep learning-based mapping for real-time, robust mapping with quantifiable uncertainty and optimized performance in an algorithm called Convolutional Bayesian Kernel Inference (ConvBKI).

Software: https://github.com/UMich-CURLY/NeuralBKI

Semantic map

ConvBKI – Semantic Map

Uncertainty map

ConvBKI – Uncertainty Map


Dynamic Mapping in a Unified Framework

Student: Joey Wilson

Dynamic semantic map

Whereas most modern methods propose to segment foreground from background, we model all objects within a single map for downstream tasks. Using velocity cues, we can embed a higher level of scene understanding in our map to reason about how objects are moving, thereby removing artifacts from dynamic objects.

Software: https://github.com/UMich-CURLY/3DMapping