You cannot select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

83 lines
3.3 KiB
Python

7 months ago
from agent.Base_Agent import Base_Agent
from behaviors.custom.Walk.Env import Env
from math_ops.Math_Ops import Math_Ops as M
from math_ops.Neural_Network import run_mlp
import numpy as np
import pickle
class Walk():
def __init__(self, base_agent : Base_Agent) -> None:
self.world = base_agent.world
self.description = "Omnidirectional RL walk"
self.auto_head = True
self.env = Env(base_agent)
self.last_executed = 0
with open(M.get_active_directory([
"/behaviors/custom/Walk/walk_R0.pkl",
"/behaviors/custom/Walk/walk_R1_R3.pkl",
"/behaviors/custom/Walk/walk_R2.pkl",
"/behaviors/custom/Walk/walk_R1_R3.pkl",
"/behaviors/custom/Walk/walk_R4.pkl"
][self.world.robot.type]), 'rb') as f:
self.model = pickle.load(f)
def execute(self, reset, target_2d, is_target_absolute, orientation, is_orientation_absolute, distance):
'''
Parameters
----------
target_2d : array_like
2D target in absolute or relative coordinates (use is_target_absolute to specify)
is_target_absolute : bool
True if target_2d is in absolute coordinates, False if relative to robot's torso
orientation : float
absolute or relative orientation of torso, in degrees
set to None to go towards the target (is_orientation_absolute is ignored)
is_orientation_absolute : bool
True if orientation is relative to the field, False if relative to the robot's torso
distance : float
distance to final target [0,0.5] (influences walk speed when approaching the final target)
set to None to consider target_2d the final target
'''
r = self.world.robot
#------------------------ 0. Override reset (since some behaviors use this as a sub-behavior)
if reset and self.world.time_local_ms - self.last_executed == 20:
reset = False
self.last_executed = self.world.time_local_ms
#------------------------ 1. Define walk parameters
if is_target_absolute: # convert to target relative to (head position + torso orientation)
raw_target = target_2d - r.loc_head_position[:2]
self.env.walk_rel_target = M.rotate_2d_vec(raw_target, -r.imu_torso_orientation)
else:
self.env.walk_rel_target = target_2d
if distance is None:
self.env.walk_distance = np.linalg.norm(self.env.walk_rel_target)
else:
self.env.walk_distance = distance # MAX_LINEAR_DIST = 0.5
# Relative orientation values are decreased to avoid overshoot
if orientation is None:
self.env.walk_rel_orientation = M.vector_angle(self.env.walk_rel_target) * 0.3
elif is_orientation_absolute:
self.env.walk_rel_orientation = M.normalize_deg( orientation - r.imu_torso_orientation )
else:
self.env.walk_rel_orientation = orientation * 0.3
#------------------------ 2. Execute behavior
obs = self.env.observe(reset)
action = run_mlp(obs, self.model)
self.env.execute(action)
return False
def is_ready(self):
''' Returns True if Walk Behavior is ready to start under current game/robot conditions '''
return True