Public defence in Mechanical Engineering, M.Sc. Alvari Seppänen

Autonomous vehicle perception and navigation in adverse conditions

Autonomous mobility has gained popularity in recent years due to the promise of safer and more efficient transportation systems. However, multiple challenges hinder the realization of fully autonomous transportation. This thesis addresses challenges related to the perception and navigation of outdoor mobile robots in adverse conditions. These conditions refer to adverse weather and limited communication between a remote operator and the robot. Adverse weather conditions affect the perception systems, such as popular light detection and ranging (LiDAR) sensors, causing specific types of noise to the data. This work aims to denoise this data and thus provide clean data for downstream systems. 

Two deep-learning-based denoising approaches are proposed: a supervised approach that utilizes a spatiotemporal module and a self-supervised multi-echo approach. The supervised method's spatiotemporal module enables efficient data usage and generalization from semi-synthetic to fully real-world data. The self-supervised approach learns by predicting the correlation of data points to their neighbors and utilizes multi-echo point clouds for recovering the points representing solid objects. Experiments show that the models achieved 10% and 23% better results compared to their respective state-of-the-art. 

Another challenge addressed is the navigation when the communication between a remote operator and a semi-autonomous mobile robot is compromised. Semi-autonomous control strategies are proposed to aid the operator when the communication signal limits the system's performance. Experiments with a mobile robot prototype revealed that the strategies improved navigation speed by 12%. 

The deep-learning models can be applied to autonomous vehicles that use LiDAR. The models enable more reliable data acquisition in adverse weather conditions. The semi-autonomous control strategies can be applied to robots that have remote-controlled functions and suffer from communication issues. Overall, the results of this thesis enable more reliable and efficient autonomous vehicles in the studied adverse conditions and build foundations for future research.

Doctoral Student: Alvari Seppänen

Opponent: Professor Martin Törngren, KTH Royal Institute of Technology, Stockholm, Sweden

Custos: Prof. Kari Tammi, Aalto University School of Engineering, Department of Mechanical Engineering

The public defense will be organized  in Lecture Hall 216, Otakaari 4.

The thesis is publicly displayed 10 days prior to the defense in the publication archive Aaltodoc of Aalto University. 

Contact information of doctoral student:

Doctoral theses in the School of Engineering: https://aaltodoc.aalto.fi/handle/123456789/49