Started Labbook 2021.

December 15, 2020

Found an ancient Workshop presentation, before the first RoboCup.
Found an interesting HRI study. from the US Army Research Laboratory, but this platform was based on the Torque Game Engine (TGE).

December 11, 2020

Looked at the robot Spot's API.
Communication with Spot goes secured HTTPS connection with the gRPC protocol on top.
The multiple Coordination frames can be transformed with SE3Pose transformations.
The robot can have multiple clients but only one can control the robot, so you have to request the lease.
Spot has additionally Choreography, which is not a visual programming tool but a way to teach him new moves.
They also have a Ricoh Theta payload example.

Nice blog on why autonomous navigation is difficult.
Kyber Photonics will start building solid-state LiDAR.
For the DARPA subterrain challenge the Australian CSIRO team found own cave, including a drone scout with a rotating LiDAR.

Watched Joris Sijs' presentation on knowledge graphs combined with Spot on Grakn Labs channel.

Interesting BARN dataset with different navigation challenges, presented at the SSRR 2020.

The HDPR paper of Boukas was published at i-SAIRAS 2016.
The repository given as reference [8] in HDPR paper doesn't exist anymore, but is possibly replaced by HPDR package set.

To put Marlon's work in perspective, see the visual reconstruction in Fig. 5.6 of Paul Timothy Furgale's PhD thesis. The reconstruction is widly inaccurate at this scale, although locally more than good enough.
In Robert A. Hewitt's or Paul T. Furgale's no reconstructed maps are given, the only map I could find was Figure 11 of HDPR: A Mobile Testbed for Current and Future Rover Technologies. Yet, the main focus of this paper is the accuracy of the RTK GPS.

Should try ROSLab, which uses Docker files and Jupyter labs to start ROS packages.

Interesting article on multiple-vehicle drone simulation from Auterion, the control-software of the Dutch medicical drone of Aya.

I have found this Matlab script to remove the gravity vector from the raw accelaration data.
The IMU sensor from the Katwijk beach dataset is a STIM300

The LIRS-UST lab from Kazan uses several Kitti and EuRoC datasets for their ProSlam implementation.
Their World Construction Tool is not available for download (yet).

The theory section of Robert A. Hewitt's PhD thesis is based on Paul Timothy Furgale's PhD thesis.

Interesting post on Simulation of Drones by Simulation, also interesting for search-and-rescue.
Sebastian Scherer has several intersting publications on sensor fusion, which are not only applicable to Drones.

NASA's A-PUFFER has a nice design for rescue mission (and also has Zed stereo camera).

The AI-robotics lab from the Eskisehir Osmangazi University has published dataset with ramps, measured with a P3AT with a Asus Xtion Pro RGB-D sensor. The Xtion seems to be already a standard sensor in Gazebo.
The robot description pacakge of the book 'Mastering ROS' also contains a pan-tilt.urdf, which would have been handy for Paul's project.