def init_planner(self, init_plan_scheme):
r_r = self.traj.reward(self.reward.reward_th, fw=tt)
if config.PREDICT_HUMAN_IGNORES_ROBOT:
# Predict that human's reward is independent of the robot's trajectory
# (i.e. human ignores the robot).
r_h = self.traj_h.reward(self.reward_h_ignore_robot.reward_th,
fw=tt)
else:
# Predict that human's reward is not independent of the robot's
# trajectory (i.e. the robot's prediction of the human will rely
# on how it chooses to initialize the human's belief of the robot's
# plan).
r_h = self.traj_h.reward(self.reward_h.reward_th, fw=tt)
self.optimizer = PredictReactMaximizer(
r_h,
self.traj_h,
r_r,
self.traj,
use_second_order=self.use_second_order,
init_plan_scheme=init_plan_scheme)
self.planner = TwoCarPlanner(self.traj,
self.traj_h,
self.human,
self.optimizer,
self.bounds,
name=self.name)
def init_planner(self, init_plan_scheme):
r_h = self.traj_h.reward(self.reward_h.reward_torch, fw=torch)
r_r = self.traj.reward(self.reward.reward_torch, fw=torch)
self.optimizer = ILQRMaximizer(self.reward_h, self.traj_h, self.reward,
self.traj, self.dyn)
self.planner = TwoCarPlanner(self.traj,
self.traj_h,
self.human,
self.optimizer,
self.bounds,
name=self.name)
def init_planner(self, init_plan_scheme):
r_h = self.traj_h.reward(self.reward_h.reward_th, fw=tt)
r_r = self.traj.reward(self.reward.reward_th, fw=tt)
self.optimizer = IteratedBestResponseMaximizer\
(
r_h, self.traj_h, r_r, self.traj,
init_plan_scheme=init_plan_scheme)
self.planner = TwoCarPlanner(self.traj,
self.traj_h,
self.human,
self.optimizer,
self.bounds,
name=self.name)
def init_planner(self, init_plan_scheme):
r_r = self.traj.reward(self.reward.reward_th, fw=tt)
r_h = self.traj_h.reward(self.reward_h.reward_th, fw=tt)
# self.optimizer = NestedMaximizer(
# r_h, self.traj_h, r_r, self.traj,
# use_second_order=self.use_second_order)
self.optimizer = NestedMaximizer(
r_h,
self.traj_h,
r_r,
self.traj,
use_second_order=self.use_second_order,
init_plan_scheme=init_plan_scheme)
self.planner = TwoCarPlanner(self.traj,
self.traj_h,
self.human,
self.optimizer,
self.bounds,
name=self.name)
def init_planner(self, init_plan_scheme):
self.traj_truck = None
r_r = self.traj.reward(self.reward.reward_th, fw=tt)
r_h = self.traj_h.reward(self.reward_h.reward_th, fw=tt)
self.optimizer = HierarchicalMaximizer(
r_h,
self.traj_h,
r_r,
self.traj,
mat_name=self.mat_name,
n=self.strat_dim,
proj=self.proj.proj_th,
traj_truck=self.traj_truck,
use_second_order=self.use_second_order,
init_plan_scheme=init_plan_scheme)
self.planner = TwoCarPlanner(self.traj,
self.traj_h,
self.human,
self.optimizer,
self.bounds,
name=self.name)