A map includes both the 3D model of a town and its road definition. A map's road definition is based on an OpenDRIVE file, a standarized, annotated road definition format. The way the [OpenDRIVE standard 1.4](http://www.opendrive.org/docs/OpenDRIVEFormatSpecRev1.4H.pdf) defines roads, lanes, junctions, etc. determines the functionality of the Python API and the reasoning behind decisions made.
__To change the map, the world has to change too__. The simulation will be recreated from scratch. You can either restart with the same map in a new world or you can change both the map and the world:
Traffic signs defined in the OpenDRIVE file are translated to CARLA as landmark objects that can be queried from the API. The following methods and classes can be used to manipulate and work with landmark objects:
- __[`carla.Landmark`](https://carla.readthedocs.io/en/latest/python_api/#carla.Landmark)__ objects represent OpenDRIVE signals. The attributes and methods of this class describe the landmark and its area of influence.
- [`carla.LandmarkOrientation`](https://carla.readthedocs.io/en/latest/python_api/#carla.LandmarkOrientation) states the orientation of the landmark with regard to the road's geometry definition.
- [`carla.LandmarkType`](https://carla.readthedocs.io/en/latest/python_api/#carla.LandmarkType) contains common landmark types to facilitate translation to OpenDRIVE types.
- __[`carla.Waypoint`](https://carla.readthedocs.io/en/latest/python_api/#carla.Waypoint)__ can get landmarks located a certain distance ahead of it. The landmark type to get can be specified.
- __[`carla.Map`](https://carla.readthedocs.io/en/latest/python_api/#carla.Map)__ retrieves sets of landmarks. It can return all landmarks in the map, or those which have a common ID, type or group.
- __[`carla.World`](https://carla.readthedocs.io/en/latest/python_api/#carla.World)__ acts as intermediary between landmarks and the `carla.TrafficSign` and `carla.TrafficLight` that represent them in the simulation.
A [`carla.Waypoint`](python_api.md#carla.Waypoint) is a 3D-directed point in the CARLA world corresponding to an OpenDRIVE lane. Everything related to waypoints happens on the client-side; communication with the server is only needed once to get the [map object](python_api.md#carlamap) containing the waypoint information.
Each waypoint contains a [`carla.Transform`](python_api.md#carla.Transform) which states its location on the map and the orientation of the lane containing it. The variables `road_id`,`section_id`,`lane_id` and `s` correspond to the OpenDRIVE road. The `id` of the waypoint is constructed from a hash combination of these four values.
!!! Note
Waypoints closer than __2cm within the same road__ share the same `id`.
A waypoint holds information about the __lane__ containing it. This information includes the lane's left and right __lane markings__, a boolean to determine if it's inside a junction, the lane type, width, and lane changing permissions.
The lane types defined by [OpenDRIVE standard 1.4](http://www.opendrive.org/docs/OpenDRIVEFormatSpecRev1.4H.pdf) are translated to the API in [`carla.LaneType`](python_api.md#carla.LaneType) as a series of enum values.
The lane markings surrounding a lane are accessed through [`carla.LaneMarking`](python_api.md#carla.LaneMarking). Lane markings are defined by a series of variables:
- __color:__ [`carla.LaneMarkingColor`](python_api.md#carla.LaneMarkingColor) are enum values that define the marking's color.
- __lane_change:__ [`carla.LaneChange`](python_api.md#carla.LaneChange) states if the lane permits turning left, right, both or none.
- __type:__ [`carla.LaneMarkingType`](python_api.md#carla.LaneMarkingType) are enum values that define the type of marking according to the OpenDRIVE standard.
A [`carla.Junction`](python_api.md#carla.Junction) represents an OpenDRIVE junction. This class encompasses a junction with a bounding box to identify lanes or vehicles within it.
The `carla.Junction` class contains the `get_waypoints` method which returns a pair of waypoints for every lane within the junction. Each pair is located at the start and end points of the junction boundaries.
Every object on a CARLA map has a set of associated variables which can be found [here][env_obj]. Included in these variables is a [unique ID][env_obj_id] that can be used to [toggle][toggle_env_obj] that object's visibility on the map. You can use the Python API to [fetch][fetch_env_obj] the IDs of each environment object based on their [semantic tag][semantic_tag]:
Navigation in CARLA is managed via the Waypoint API, a combination of methods from [`carla.Waypoint`](python_api.md#carla.Waypoint) and [`carla.Map`](python_api.md#carla.Map).
The client must initially communicate with the server to retrieve the map object containing the waypoint information. This is only required once, all subsequent queries are performed on the client side.
The Waypoint API exposes methods that allow waypoints to connect to each other and construct a path along a road for vehicles to navigate:
-`next(d)` creates a list of waypoints within an approximate distance, `d`, __in the direction of the lane__. The list contains one waypoint for each possible deviation.
-`previous(d)` creates a list of waypoints waypoint within an approximate distance, `d`, __in the opposite direction of the lane__. The list contains one waypoint for each possible deviation.
-`next_until_lane_end(d)` and `previous_until_lane_start(d)` return a list of waypoints a distance `d` apart. The lists go from the current waypoint to the end and beginning of its lane, respectively.
-`get_right_lane()` and `get_left_lane()` return the equivalent waypoint in an adjacent lane, if one exists. A lane change maneuver can be made by finding the next waypoint to the one on its right/left lane, and moving to it.
The client needs to make a request to the server to get the `.xodr` map file and parse it to a [`carla.Map`](python_api.md#carla.Map) object. This only needs to be done once.
The map object contains __recommended spawn points__ for the creation of vehicles. You can get a list of these spawn points, each one containing a [`carla.Transform`](python_api.md#carlatransform), using the method below. Bear in mind that the spawn points may be occupied already, resulting in failed creation of vehicles due to collisions.
You can get started with waypoints by __[getting](python_api.md#carla.Map.get_waypoint) the closest waypoint__ to a specific location or to a particular `road_id`, `lane_id` and `s` value in the map's OpenDRIVE definition:
The below example shows how to __generate a collection of waypoints__ to visualize the city lanes. This will create waypoints all over the map, for every road and lane. All of them will approximately 2 meters apart:
To __generate a minimal graph of road topology__, use the example below. This will return a list of pairs (tuples) of waypoints. The first element in each pair connects with the second element and both define the start and end points of each lane in the map. More information on this method is found in the [PythonAPI](python_api.md#carla.Map.get_topology).
The example below __converts a `carla.Transform` to geographical latitude and longitude coordinates,__ in the form of a [`carla.GeoLocation`](python_api.md#carla.GeoLocation):
There are eight towns in the CARLA ecosystem and each of those towns have two kinds of map, non-layered and layered. [Layers][layer_api] refer to the grouped objects within a map and consist of the following:
Non-layered maps are shown in the table below (click the town name to see an overhead image of the layout). All of the layers are present at all times and cannot be toggled on or off in these maps. Up until CARLA 0.9.11, these were the only kinds of map available.
| **[Town01](img/Town01.jpg)** | A basic town layout consisting of "T junctions".|
| **[Town02](img/Town02.jpg)** | Similar to **Town01**, but smaller.|
| **[Town03](img/Town03.jpg)** | The most complex town, with a 5-lane junction, a roundabout, unevenness, a tunnel, and more.|
| **[Town04](img/Town04.jpg)** | An infinite loop with a highway and a small town.|
| **[Town05](img/Town05.jpg)** | Squared-grid town with cross junctions and a bridge. It has multiple lanes per direction. Useful to perform lane changes. |
| **[Town06](img/Town06.jpg)** | Long highways with many highway entrances and exits. It also has a [**Michigan left**](<https://en.wikipedia.org/wiki/Michigan_left>). |
| **[Town07](img/Town07.jpg)** | A rural environment with narrow roads, barns and hardly any traffic lights. |
| **[Town10](img/Town10.jpg)** | A city environment with different environments such as an avenue or promenade, and more realistic textures.|
The layout of layered maps is the same as non-layered maps but it is possible to toggle off and on the layers of the map. There is a minimum layout that cannot be toggled off and consists of roads, sidewalks, traffic lights and traffic signs. Layered maps can be identified by the suffix `_Opt`, for example, `Town01_Opt`. With these maps it is possible to [load][load_layer] and [unload][unload_layer] layers via the Python API:
CARLA is designed to be extensible and highly customisable for specialist applications. Therefore, in addition to the many maps and assets already avaiable in CARLA out of the box, it is possible to create and import new maps, road networks and assets to populate bespoke environments in a CARLA simulation. The following documents detail the steps needed to build and integrate custom maps:
