Smooth-Trajectron++: Augmenting the Trajectron++ behaviour prediction model with smooth attention

Understanding traffic participants’ behaviour is crucial for predicting their future trajectories, enabling autonomous vehicles to better assess the environment and consequently anticipate possible dangerous situations at an early stage. While the integration of cognitive processes and machine learning models has demonstrated promise in various...

An Investigation of the Medical Ultrasound Image Sparse Spaces used for the Model-Based Imaging

In this thesis, we investigate a sparse basis for ultrasound images, so that we can use sparse regularization in imaging. Actually, there are few previous researches explicitly demonstrating that medical ultrasound images can be sparsified for some dictionary. We consider various orthogonal transforms such as wavelet transforms, cosine...

Direction-of-Arrival Estimation using an Unsynchronized Array Of Acoustic Vector Sensors

Direction-Of-Arrival (DOA) estimation of acoustic signals is of great interest in various applications including battlefield acoustic and noise localization. Acoustic sensors are employed in an array configuration to estimate DOAs based on the time differences of arrival DOAs. However, the acoustic sensors in the network have all their own Data...

Reward system design for incorporating control performance

Reinforcement learning (RL) is a machine learning technique whereby the controller learns the control law by optimizing the received cumulative amount of reward. A reward is an instantaneous evaluation of the applied action at the current state, given by reward function. However in theory the reward function is assumed to be given, in practice...

Nonlinear fuzzy observer for payload estimation in physical human-robot interaction

Robotics have developed at a very high rate in the last few years. Industries are becoming capable of automating repetitive procedures without pause, and the production speed and quality are improving everyday. The main drawback of robotics in industry is the inability to deal with slight changes in an environment or in certain object parameters...

Filtering techniques for orbital debris conjunction analysis

The steadily growing amount of orbital debris increases the probability and amount of collisions between two objects in orbit about the Earth. These collisions in turn create even more debris, and it is therefore important to keep track of future conjunctions. The U.S. Space Surveillance Network (SSN) uses ground- and space-based sensors to...