Are the coordinates in the data set all relative to the robot?
Opened this issue · 5 comments
Hello @zhlstone,
Yes, all coordinates are relative to the robot. The full description of the dataset can be found in https://arxiv.org/abs/2009.05345 and an example of how to use the dataset in https://arxiv.org/abs/1909.09003.
Let us know if you have any other doubts.
Hello @zhlstone,
Yes, all coordinates are relative to the robot. The full description of the dataset can be found in https://arxiv.org/abs/2009.05345 and an example of how to use the dataset in https://arxiv.org/abs/1909.09003.
Let us know if you have any other doubts.
Thanks very much! Another question is whether you include pedestrian speed information in your later work?
Hello @zhlstone,
Yes, in a later work (https://arxiv.org/pdf/2102.08863.pdf) we include pedestrian speed information. Repository: https://github.com/gnns4hri/sngnnv2.
Hello @ljmanso,
I have a question about how to obtain the heat maps in https://arxiv.org/abs/1909.09003, and I didn't find any visualizing code in this project. Thanks in advance!
Hello @zhlstone,
If you want to generate maps of the static scenarios we have an optimised version in this link: https://github.com/gnns4hri/sngnn2d. With the use of a transpose CNN the generation of the image map is much faster. You just have to run the script test.py in that repository. "python3 test.py". It shouldn't take more than one second per image.
On the other hand, if you want to generate a map for dynamic scenarios where the velocity of the entities is considered, you can use the code in the repository: https://github.com/gnns4hri/sngnnv2. In that repository, there is a README file with the instructions to run the training and testing. Take into account that the testing process (generation of the heat map) in this project is not optimised yet, so it will take a long while proportional to the resolution. You can adjust the resolution as indicated in the README file.
I hope that helps. If you have any other question don't hesitate to ask.