Comments (7)
If the velocity is in m/sec, then how big is the arena in meters? In other words, at a velocity of 1 m/s, how long would it take for an agent to travel from one end of the arena to the other (coordinates (0,40) to (40,40) in the 40x40 arena).
I guess I didn't state the reference question clearly. How about this-- if I get a velocity observation of (0.0, 0.0, 1.0) while the agent is facing toward the upper left corner of the arena, does that mean it is moving toward the upper left corner (that would mean we are using the agent's reference frame) or does it mean it is moving directly towards the top of the arena (that would mean we are using the arena's (or the world's) reference frame).
from animalai-olympics.
What is a "row" in this context?
from animalai-olympics.
The arena is 40 rows x 40 columns, so a velocity might be expressed as rows/sec.
from animalai-olympics.
Hello,
We're using the Unity convention which means that:
- lengths are expressed in "units" which in our case we'll associate to 1 metre (there are no rows/columns, they are continuous values for all 3 axes)
- speeds are therefore expressed in m/s (framerate and machine independent)
Regarding the referential, each arena has its own referential for coordinates, speed is referential independent and expressed in m/s always. All arenas are static so there is no relative/absolute speed.
I would have thought that an agent in the wild would be able to observe its velocity in its own frame but not in the world frame.
not sure I understand your point, we assume perception of velocity is the same for all agents.
Also, just to be sure, actions such as forward and backward are implemented as momentary accelerations in the agent's frame, is that correct?
correct
from animalai-olympics.
Also, just to be sure, actions such as forward and backward are implemented as momentary accelerations in the agent's frame, is that correct? I surmised that by seeing the effects of multiple forward inputs-- the velocity increases with each one and then gradually slows to zero after the inputs end.
More detail would be nice for this point, for instance; from Unity's ML-Agent's documentation? I would like to know what is the motion model of an agent?
from animalai-olympics.
If the velocity is in m/sec, then how big is the arena in meters? In other words, at a velocity of 1 m/s, how long would it take for an agent to travel from one end of the arena to the other (coordinates (0,40) to (40,40) in the 40x40 arena).
The arena is 40 metres by 40 metres, it would therefore take 40 seconds in your example, sorry for the lack of clarity!
I guess I didn't state the reference question clearly. How about this-- if I get a velocity observation of (0.0, 0.0, 1.0) while the agent is facing toward the upper left corner of the arena, does that mean it is moving toward the upper left corner (that would mean we are using the agent's reference frame) or does it mean it is moving directly towards the top of the arena (that would mean we are using the arena's (or the world's) reference frame).
OK I now understand your question, the speed at the moment is in the world's referential which as you mention doesn't make much sense from an animal perspective. We'll change that and make it agent's referential.
More detail would be nice for this point, for instance; from Unity's ML-Agent's documentation? I would like to know what is the motion model of an agent?
I will write a documentation about that and the above point which will itself point to the Unity documentation for the physics function we're using. We'll put that in the final release next week.
In a nutshell rotation is immediate by a fixed positive/negative angle and forward/backward motion is a constant force applied every time you send a non zero action (see here and here for the exact Physics function used)
from animalai-olympics.
Thanks, that clears it up.
from animalai-olympics.
Related Issues (20)
- Website is down HOT 1
- what's the angle of view of the agent? HOT 2
- [beginner's question] how much computing resources are required to train a reasonable model? HOT 2
- Get agent x, y, z position during training HOT 1
- Question about example config files HOT 2
- Can i get the velocity of the agent? HOT 2
- Docker and AWS HOT 5
- Gym Wrapper multi-agent HOT 2
- overtrained? HOT 8
- Question about internal memory category HOT 1
- Instructions for final submissions HOT 2
- Env directory is ignored in .gitignore HOT 1
- Windows ENV cant donwload HOT 1
- Online-world suggestion HOT 1
- Google Colab Run error HOT 1
- Question about the way about curriculum HOT 2
- Windows environment download link HOT 1
- Modify environment assets / make new environment? HOT 1
- Tensorflow 2 compatability HOT 2
- CylinderTunnel and CylinderTunnelTransparent items not showing in the arena
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from animalai-olympics.