LAB: Learnable Activation Binarizer for Binary Neural Networks

Binary Neural Networks (BNNs) are receiving an upsurge of attention for bringing power-hungry deep learning towards edge devices. The traditional wisdom in this space is to employ sign(.) for binarizing featuremaps. We argue and illustrate that sign(.) is a uniqueness bottleneck, limiting information propagation throughout the network. To...

Robust Bluetooth Low Power Communication

This work presents the first robust connection mode for Bluetooth Low Energy on an energy harvesting sensor.

The negative latency haptic glove: Making Tactile Internet tangible

The field of Tactile Internet (TI) is transporting physical interactions over vast distances as if they were near. This teleoperation method increases the quality of tele-engineering, telesurgery and many other fields. However, it requires the strict end to end latency constraints of a few ms, which is still out of reach. This thesis looks at...

Multi-­Robot Exploration and Planning in Limited Communication Environments

Distributed cooperative robots can be highly beneficial in mapping disaster environments and assisting with search and rescue operations. In most situations such environments only allow for only limited communication between robots. This thesis reports on simulation experiments conducted to test the impact of having only partial communication...

Improving the Predictability of Event Chains in ROS 2

ROS 2 is a well-known software framework for the development of robotic applications. The open-source project enjoys active interest and participation from both academia and industry in applications spanning from autonomous vehicles to industrial robotics. ROS 2's success is underpinned by its disposition for collaborative development as well as...

Digital Control of Multi-channel RF Energy Harvesters

This thesis tries to match the rectifier block of an RF energy harvester to a specific frequency. This specific frequency has the most power and is searched and selected by a smart algorithm. <br/>The algorithm tunes a varactor by applying a proper voltage to set it on different channels.<br/>Then the best channel is sampled and compared with...