augcog / roar Goto Github PK
View Code? Open in Web Editor NEWRobot Open Autonomous Racing (ROAR)
Home Page: https://augcog.github.io/ROAR/
License: Apache License 2.0
Robot Open Autonomous Racing (ROAR)
Home Page: https://augcog.github.io/ROAR/
License: Apache License 2.0
Is your feature request related to a problem? Please describe.
make sure that even if I don't have aruidno plugged in, the program is still going to run without panicking
os:win10
describe:
1.python .\runner_ios.py -m vr --auto False
2.Xbox controls the car to reverse
result:
exited unexpectedly
Describe the bug
Running python3 runner_jetson.py
and wait for about 1 minute, its going to freeze. It seems like the program is still runing in the background, but is just blocked by something.
To Reproduce
Steps to reproduce the behavior:
python3 runner_jetson.py
on Jetson NanoExpected behavior
It keeps going until agent is finished
Describe the bug
In carla_bridge.py:
convert_xx_from_agent_to_source functions have issue:
Lines 182 to 197 in b7cd322
are not inverse functions of
Lines 29 to 58 in b7cd322
To Reproduce
convert_transform_from_agent_to_source
convert_transform_from_source_to_agent
Expected behavior
The original agent transform should be the same thing as the new agent transform (the one converted to carla and converted back)
Desktop (please complete the following information):
Is your feature request related to a problem? Please describe.
The code currently is using Multi-Threadding. However, according to multiple sources
https://www.guru99.com/difference-between-multiprocessing-and-multithreading.html
https://stackoverflow.com/questions/3044580/multiprocessing-vs-threading-python
It seems like the proper solution should be using multi processing, to avoid race condition, and etc.
Might be useful to get the right parallelism right.
Describe the solution you'd like
Utilize multiprocessing in Jetson modules, and also ensure that proper Agent classes have the ability to run in parallel
Describe the bug
There is unknown object caused collision at a corner in ROAR easy map, but it didn't report collision.
Could you check the ROAR easy map?
Thanks,
To Reproduce
Steps to reproduce the behavior:
Expected behavior
Expect the car can run near the corner in ROAR easy map without spin over or report collision.
Screenshots
Attached the video
Desktop (please complete the following information):
Describe the bug
The car does not behave the same way when running runner_sim.py vs. runner_competition_evaluator.py.
To Reproduce
Steps to reproduce the behavior:
Expected behavior
The car speeds up more during runner_sim.py.
Screenshots
If applicable, add screenshots to help explain your problem.
Desktop (please complete the following information):
Additional context
Add any other context about the problem here.
Describe the bug
Ground Plane detection module does not work on actual realsense camera
To Reproduce
Steps to reproduce the behavior:
OccupancyMapAgent
in jetson_runner.py
ground_plane_detector.py
, the lines for displaying the image is enabledExpected behavior
a cv2 windows should pop up, as do in the simulator
Is your feature request related to a problem? Please describe.
More aggressive/better PID values
current car doesn't move smoothly, that's probably due to a combination of PID and look ahead values. Need to do more experiment to find an optimal value set.
Is your feature request related to a problem? Please describe.
the car running on the Carla can’t move backward. It will cause when the users operate the car collision with the 'wall', the car will can't move. The user have to restart the runner_sim.py .
To Reproduce
Steps to reproduce the behavior:
python runner_sim.py
gives the error No such file or directory: 'ROAR_Sim/configurations/occu_map_config.json'
Describe the bug
So after some rolling back, i figured out that it is indeed some bad translation between left-hand and right hand coordinate system that messes things up. However, I am stucked at how to make it right. The commit that is working (in the left hand coordinate system) is this one:https://github.com/augcog/ROAR/tree/0bde066531e559e291fc8e395b9c3a53de8f014f/ROAR
Make sure to checkout both the ROAR_Sim and ROAR on that commit.
I've also raised an issue on Github.
I believe the core issue is that we've got many points where we need to notice the coordinate1. Translating data from ROAR Agent to Carla2. Translating data from Carla to ROAR Agent3. Reading in txt waypoint file (since the waypoint file is recorded by the ROAR agent, therefore it will be impacted if the ROAR agent is in the left or right coordinate system).4. Specifically this line that does the transformation in visualization:https://github.com/augcog/ROAR/blob/main/ROAR/visualization_module/visualizer.py#L141
To Reproduce
Steps to reproduce the behavior:
Just run the agent with next waypoint visulazation, it is not going to draw a straight line
np.clip() got applied to Kp, Ki, Kd raw coefficients in find_k_values function in PIDController class.
PID values got clipped where the minimum is 0 and the maximum is 1.
Steps to reproduce the behavior:
Expected behavior
PID values can go over 1, for example k_p = 10000 is valid
Additional context
np.clip is used in 33 places across 11 files. Need to be checked.
Describe the bug
wapoint file is not accurate due to movement of base station. Need to recollect waypoints
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.