def __init__(self, state, groups=None, obstacles=None, config_file=None):
self.config = DefaultConfig()
if config_file:
self.config.load_config(config_file)
# TODO: load obstacles from config
self.scene_config = self.config.sub_config("scene")
# initiate obstacles
self.env = EnvState(obstacles, self.config("resolution", 10.0))
# initiate agents
self.peds = PedState(state, groups, self.config)
# construct forces
self.forces = self.make_forces(self.config)
def __init__(self, state, groups=None, obstacles=None, fires=None, exits=None, border=None, config_file=None):
self.config = DefaultConfig()
if config_file:
self.config.load_config(config_file)
# TODO: load obstacles from config
self.scene_config = self.config.sub_config("scene")
# initiate obstacles
self.env = EnvState(self.scene_config, obstacles, fires, exits, self.config("resolution", 10.0))
# initiate agents
self.peds = PedState(self, state, groups, border, self.scene_config)
# if the agents have knowledge about the positions of emergency exits,
# set closest emergency exit as target
if self.scene_config("exit_knowledge") :
self.peds.set_exits_as_targets()
else :
# We only want to consider one fire right now..
if fires is not None and len(fires) <= 1 :
self.peds.set_fire_target()
# construct forces
self.forces = self.make_forces(self.config)
class Simulator:
"""Simulate social force model.
...
Attributes
----------
state : np.ndarray [n, 6] or [n, 7]
Each entry represents a pedestrian state, (x, y, v_x, v_y, d_x, d_y, [tau])
obstacles : np.ndarray
Environmental obstacles
groups : List of Lists
Group members are denoted by their indices in the state
config : Dict
Loaded from a toml config file
max_speeds : np.ndarray
Maximum speed of pedestrians
forces : List
Forces to factor in during navigation
Methods
---------
capped_velocity(desired_velcity)
Scale down a desired velocity to its capped speed
step()
Make one step
"""
def __init__(self, state, groups=None, obstacles=None, fires=None, exits=None, border=None, config_file=None):
self.config = DefaultConfig()
if config_file:
self.config.load_config(config_file)
# TODO: load obstacles from config
self.scene_config = self.config.sub_config("scene")
# initiate obstacles
self.env = EnvState(self.scene_config, obstacles, fires, exits, self.config("resolution", 10.0))
# initiate agents
self.peds = PedState(self, state, groups, border, self.scene_config)
# if the agents have knowledge about the positions of emergency exits,
# set closest emergency exit as target
if self.scene_config("exit_knowledge") :
self.peds.set_exits_as_targets()
else :
# We only want to consider one fire right now..
if fires is not None and len(fires) <= 1 :
self.peds.set_fire_target()
# construct forces
self.forces = self.make_forces(self.config)
def make_forces(self, force_configs):
"""Construct forces"""
force_list = [
forces.DesiredForce(),
forces.SocialForce(),
forces.ObstacleForce(),
forces.FireForce(),
forces.PedRepulsiveForce(),
forces.SpaceRepulsiveForce(),
]
group_forces = [
forces.GroupCoherenceForceAlt(),
forces.GroupRepulsiveForce(),
forces.GroupGazeForceAlt(),
]
if self.scene_config("enable_group"):
force_list += group_forces
# initiate forces
for force in force_list:
force.init(self, force_configs)
return force_list
def compute_forces(self):
"""compute forces"""
return sum(map(lambda x: x.get_force(), self.forces))
def get_states(self):
"""Expose whole state"""
return self.peds.get_states()
def get_length(self):
"""Get simulation length"""
return len(self.get_states()[0])
def get_obstacles(self):
return self.env.obstacles
def get_fires(self) :
return self.env.fires
def get_exits(self) :
return self.env.exits
def get_smoke_radius(self) :
return self.env.smoke_radius
def set_smoke_radius(self, new_radius) :
self.env.smoke_radius = new_radius
def get_escaped(self) :
return np.array(self.peds.escaped)
def get_dead(self) :
return np.array(self.peds.dead)
def get_health(self) :
return np.array(self.peds.av_health)
def get_panic(self) :
return np.array(self.peds.av_panic)
def step_once(self):
"""step once"""
self.peds.step(self.compute_forces())
def step(self, n=1):
"""Step n time"""
N = n
finished = False
for _ in range(n):
self.step_once()
if not finished :
if self.peds.get_nr_escaped() + self.peds.get_nr_dead() == self.peds.get_nr_peds() :
print(str(self.peds.get_nr_escaped()) + " escaped and " + str(self.peds.get_nr_dead()) + " died of " + str(self.peds.get_nr_peds()) + " in total after " + str(_) + " steps.")
finished = True
N = _
return N
def step_until_end(self):
all_escaped_or_dead = False
n = 1
N = 1
while not all_escaped_or_dead :
self.step_once()
if self.peds.get_nr_escaped() == self.peds.get_nr_peds() :
print("All people escaped after " + str(n) + " steps!")
all_escaped_or_dead = True
N = n
return N
if self.peds.get_nr_dead() == self.peds.get_nr_peds() :
print("All people died after " + str(n) + " steps!")
all_escaped_or_dead = True
N = n
return N
if self.peds.get_nr_escaped() + self.peds.get_nr_dead() == self.peds.get_nr_peds() :
print(str(self.peds.get_nr_escaped()) + " escaped and " + str(self.peds.get_nr_dead()) + " died of " + str(self.peds.get_nr_peds()) + " in total after " + str(n) + " steps.")
all_escaped_or_dead = True
N = n
return N
n += 1
return n
class Simulator:
"""Simulate social force model.
...
Attributes
----------
state : np.ndarray [n, 6] or [n, 7]
Each entry represents a pedestrian state, (x, y, v_x, v_y, d_x, d_y, [tau])
obstacles : np.ndarray
Environmental obstacles
groups : List of Lists
Group members are denoted by their indices in the state
config : Dict
Loaded from a toml config file
max_speeds : np.ndarray
Maximum speed of pedestrians
forces : List
Forces to factor in during navigation
Methods
---------
capped_velocity(desired_velcity)
Scale down a desired velocity to its capped speed
step()
Make one step
"""
def __init__(self, state, groups=None, obstacles=None, config_file=None):
self.config = DefaultConfig()
if config_file:
self.config.load_config(config_file)
# TODO: load obstacles from config
self.scene_config = self.config.sub_config("scene")
# initiate obstacles
self.env = EnvState(obstacles, self.config("resolution", 10.0))
# initiate agents
self.peds = PedState(state, groups, self.config)
# construct forces
self.forces = self.make_forces(self.config)
def make_forces(self, force_configs):
"""Construct forces"""
force_list = [
forces.DesiredForce(),
forces.SocialForce(),
forces.ObstacleForce(),
# forces.PedRepulsiveForce(),
# forces.SpaceRepulsiveForce(),
]
group_forces = [
forces.GroupCoherenceForceAlt(),
forces.GroupRepulsiveForce(),
forces.GroupGazeForceAlt(),
]
if self.scene_config("enable_group"):
force_list += group_forces
# initiate forces
for force in force_list:
force.init(self, force_configs)
return force_list
def compute_forces(self):
"""compute forces"""
return sum(map(lambda x: x.get_force(), self.forces))
def get_states(self):
"""Expose whole state"""
return self.peds.get_states()
def get_length(self):
"""Get simulation length"""
return len(self.get_states()[0])
def get_obstacles(self):
return self.env.obstacles
def step_once(self):
"""step once"""
self.peds.step(self.compute_forces())
def step(self, n=1):
"""Step n time"""
for _ in range(n):
self.step_once()
return self