clean up the repo further

README.md

@@ -42,23 +42,24 @@ pip install -r requirements.txt
 
 ## Getting started
 This repository are organized in two parts: gym_crowd/ folder contains the simulation environment and
-crowd_nav/ folder contains codes for training and testing the policies. All the commands 
-below should be run inside the crowd_nav/ folder.
+crowd_nav/ folder contains codes for training and testing the policies. Details of the simulation framework can be found 
+[here](crowd_sim/README.md). Below are the instructions for training and testing policies, and they should be executed 
+inside the crowd_nav/ folder.
 
 
 1. Train a policy.
 ```
-python train.py --policy om-sarl
+python train.py --policy sarl
 ```
 2. Test policies with 500 test cases.
 ```
 python test.py --policy orca --phase test
-python test.py --policy om-sarl --model_dir data/output --phase test
+python test.py --policy sarl --model_dir data/output --phase test
 ```
 3. Run policy for one episode and visualize the result.
 ```
 python test.py --policy orca --phase test --visualize --test_case 0
-python test.py --policy om-sarl --model_dir data/output --phase test --visualize --test_case 0
+python test.py --policy sarl --model_dir data/output --phase test --visualize --test_case 0
 ```
 4. Plot training curve
 ```
@@ -73,48 +74,3 @@ SARL | OM-SARL
 <img src="https://i.imgur.com/rUtAGVP.gif" width="400" />|<img src="https://i.imgur.com/UXhcvZL.gif" width="400" />
 
 
-## Simulation Framework
-### Environment
-The environment contains n+1 agents. N of them are humans controlled by certain unknown
-policy. The other is robot and it's controlled by one known policy.
-The environment is built on top of OpenAI gym library, and has implemented two abstract methods.
-* reset(): the environment will reset positions for all the agents and return observation 
-for robot. Observation for one agent is the observable states of all other agents.
-* step(action): taking action of the robot as input, the environment computes observation
-for each agent and call agent.act(observation) to get actions of agents. Then environment detects
-whether there is a collision between agents. If not, the states of agents will be updated. Then 
-observation, reward, done will be returned.
-
-
-### Agent
-Agent is a base class, and has two derived class of human and robot. Agent class holds
-all the physical properties of an agent, including position, velocity, orientation, policy and etc.
-* visibility: humans are always visible, but robot can be set to be visible or invisible
-* sensor: can be either visual input or coordinate input
-* kinematics: can be either holonomic (move in any direction) or unicycle (has rotation constraints)
-* act(observation): transform observation to state and pass it to policy
-
-
-### Policy
-Policy takes state as input and output an action. Current available policies:
-* ORCA: compute collision-free velocity under the reciprocal assumption
-* CADRL: learn a value network to predict the value of a state and during inference it predicts action for the most important human
-* LSTM-RL: use lstm to encode the human states into one fixed-length vector
-* SARL: use pairwise interaction module to model human-robot interaction and use self-attention to aggregate humans' information
-* OM-SARL: extend SARL by encoding intra-human interaction with a local map
-
-
-### State
-There are multiple definition of states in different cases. The state of an agent representing all
-the knowledge of environments is defined as JointState, and it's different from the state of the whole environment.
-* ObservableState: position, velocity, radius of one agent
-* FullState: position, velocity, radius, goal position, preferred velocity, rotation
-* DualState: concatenation of one agent's full state and one another agent's observable state
-* JoinState: concatenation of one agent's full state and all other agents' observable states 
-
-
-### Action
-There are two types of actions depending on what kinematics constraint the agent has.
-* ActionXY: (vx, vy) if kinematics == 'holonomic'
-* ActionRot: (velocity, rotation angle) if kinematics == 'unicycle'
-

crowd_nav/configs/env.config

@@ -14,8 +14,8 @@ discomfort_penalty_factor = 0.5
 
 
 [sim]
-train_val_sim = mixed
-test_sim = mixed
+train_val_sim = circle_crossing
+test_sim = circle_crossing
 square_width = 10
 circle_radius = 4
 human_num = 5

crowd_nav/policy/policy_factory.py

@@ -1,10 +1,8 @@
 from crowd_sim.envs.policy.policy_factory import policy_factory
 from crowd_nav.policy.cadrl import CADRL
 from crowd_nav.policy.lstm_rl import LstmRL
-from crowd_nav.policy.srl import SRL
 from crowd_nav.policy.sarl import SARL
 
 policy_factory['cadrl'] = CADRL
 policy_factory['lstm_rl'] = LstmRL
-policy_factory['srl'] = SRL
 policy_factory['sarl'] = SARL

crowd_nav/test.py

@@ -5,7 +5,7 @@
 import numpy as np
 import gym
 import os
-from crowd_nav.utils.robot import Robot
+from crowd_sim.envs.utils.robot import Robot
 from crowd_nav.utils.explorer import Explorer
 from crowd_nav.policy.policy_factory import policy_factory
 from crowd_sim.envs.policy.orca import ORCA