Implementation:CARLA simulator Carla World Get Random Location From Navigation

Knowledge Sources	CARLA Simulator CARLA Documentation
Domains	Navigation Mesh, Pedestrian Simulation
Last Updated	2026-02-15 00:00 GMT

Overview

Concrete tool for sampling random walkable locations from the pedestrian navigation mesh, provided by the CARLA simulator.

Description

The World.get_random_location_from_navigation() method queries the Recast/Detour navigation mesh loaded for the current map and returns a random point on a walkable surface. This is the primary mechanism for generating valid pedestrian spawn locations and walking destinations.

Internally, the method calls into CARLA's Navigation module (carla::nav::Navigation), which interfaces with Detour's dtNavMeshQuery to sample a random polygon on the NavMesh, then selects a random point within that polygon. The returned location is guaranteed to lie on a navigable surface (sidewalk, crosswalk, plaza, etc.).

The method returns None if the navigation mesh is not loaded or if the random sampling fails. This can happen if the map was loaded without its corresponding NavMesh binary file.

Usage

Use this method when you need to:

• Generate spawn positions for pedestrian walkers on sidewalks
• Produce random destination waypoints for pedestrian AI controllers
• Create realistic pedestrian distributions across the map without manually specifying positions

Code Reference

Source Location

• Repository: CARLA
• File: LibCarla/source/carla/nav/Navigation.cpp
• Lines: L1062-1100
• Python Binding: PythonAPI/carla/src/World.cpp, L310

Signature

World.get_random_location_from_navigation() -> carla.Location or None

Import

import carla

I/O Contract

Inputs

Name	Type	Required	Description
(none)	--	--	This method takes no parameters. It samples from the NavMesh associated with the currently loaded map.

Outputs

Name	Type	Description
return	carla.Location or None	A random 3D location (x, y, z) on the navigation mesh. Returns None if the NavMesh is not available or sampling fails.

Usage Examples

Basic Example

import carla
import random

client = carla.Client('localhost', 2000)
client.set_timeout(10.0)
world = client.get_world()

# Sample a random walkable location for a pedestrian
spawn_location = world.get_random_location_from_navigation()
if spawn_location is not None:
    print(f"Pedestrian spawn at: x={spawn_location.x:.2f}, "
          f"y={spawn_location.y:.2f}, z={spawn_location.z:.2f}")

Spawning Multiple Pedestrians

import carla
import random

client = carla.Client('localhost', 2000)
client.set_timeout(10.0)
world = client.get_world()

# Get walker blueprints
walker_blueprints = world.get_blueprint_library().filter('walker.pedestrian.*')
number_of_walkers = 50

# Generate spawn points on the navigation mesh
spawn_points = []
for _ in range(number_of_walkers):
    spawn_loc = world.get_random_location_from_navigation()
    if spawn_loc is not None:
        spawn_points.append(carla.Transform(spawn_loc))

# Batch-spawn pedestrians at the sampled locations
batch = []
for spawn_point in spawn_points:
    walker_bp = random.choice(walker_blueprints)
    if walker_bp.has_attribute('is_invincible'):
        walker_bp.set_attribute('is_invincible', 'false')
    batch.append(carla.command.SpawnActor(walker_bp, spawn_point))

results = client.apply_batch_sync(batch, do_tick=True)

walker_ids = []
for result in results:
    if not result.error:
        walker_ids.append(result.actor_id)

print(f"Spawned {len(walker_ids)} pedestrians on sidewalks")

Generating Walking Destinations

import carla

client = carla.Client('localhost', 2000)
client.set_timeout(10.0)
world = client.get_world()

# Assume walkers and their AI controllers are already spawned
# walker_controllers is a list of WalkerAIController actors

for controller in walker_controllers:
    controller.start()
    # Use get_random_location_from_navigation as the walking destination
    destination = world.get_random_location_from_navigation()
    if destination is not None:
        controller.go_to_location(destination)
    controller.set_max_speed(1.4)  # Average walking speed in m/s

Robust Sampling with Retry

import carla

client = carla.Client('localhost', 2000)
client.set_timeout(10.0)
world = client.get_world()

def get_valid_spawn_locations(world, count, max_attempts_per_location=10):
    """Sample 'count' valid locations from the NavMesh with retry logic."""
    locations = []
    for _ in range(count):
        for attempt in range(max_attempts_per_location):
            loc = world.get_random_location_from_navigation()
            if loc is not None:
                locations.append(loc)
                break
        else:
            print("Warning: Could not sample a valid NavMesh location after "
                  f"{max_attempts_per_location} attempts")
    return locations

# Sample 100 valid spawn locations
spawn_locations = get_valid_spawn_locations(world, count=100)
print(f"Sampled {len(spawn_locations)} valid pedestrian locations")

Related Pages

Implements Principle

• Principle:CARLA_simulator_Carla_Pedestrian_Navigation_Mesh

Uses Heuristic

• Heuristic:CARLA_simulator_Carla_Walker_Spawn_Vertical_Offset