A fresh attempt at overviewing the map model docs

This commit is contained in:
Dustin Carlino 2020-10-11 12:54:17 -07:00
parent b9efdaef20
commit 4d549f80df
6 changed files with 277 additions and 132 deletions

View File

@ -16,12 +16,13 @@
- [API](dev/api.md)
- [Testing](dev/testing.md)
- [Map model](map/README.md)
- [Details](map/details.md)
- [Importing](map/importing/README.md)
- [convert_osm](map/importing/convert_osm.md)
- [Road/intersection geometry](map/importing/geometry.md)
- [The rest](map/importing/rest.md)
- [Misc](map/importing/misc.md)
- [Live edits](map/edits.md)
- [Misc](map/misc.md)
- [Exporting](map/platform.md)
- [Traffic simulation](trafficsim/README.md)
- [Discrete event simulation](trafficsim/discrete_event.md)

View File

@ -1,105 +1,130 @@
# Map model
A/B Street builds a rich representation of a city map using OpenStreetMap (OSM)
and other sources. This chapter describes how.
A/B Street transforms OpenStreetMap (OSM) data into a detailed geometric and
semantic representation of the world for traffic simulation. This chapter
describes that map model, with the hopes that it'll be useful for purposes
beyond this project.
TODO: Integrate pictures from
[these slides](https://docs.google.com/presentation/d/1cF7qFtjAzkXL_r62CjxBvgQnLvuQ9I2WTE2iX_5tMCY/edit?usp=sharing).
## Overview
[This recorded presentation](https://youtu.be/chYd5I-5oyc?t=439) covers some of
this.
A `Map` covers everything inside some hand-drawn boundary, usually scoped to a
city or a few of a city's districts. Unlike OSM, it doesn't cover the entire
world; it only has areas specifically extracted for some purpose.
## The map
A map consists of many objects. Mainly, there are roads, broken down into
individual lanes, and intersections. A road is a single segment connecting
exactly two intersections (as opposed to OSM, where a single "way" may span many
intersections). Lanes within a road have a specific type, which dictates their
direction of travel (or lack of travel, like on-street parking) and uses.
Sidewalks are represented as bidirectional lanes. Roads connect at
intersections, which contain an explicit set of turns, each linking a source
lane to a destination lane.
A single city is broken down into different pieces...
Maps also contain parking lots and buildings, which connect to the nearest
driveable lane and a sidewalk. Maps have water and park areas, only used for
drawing. They also represent public transit stops and routes.
A/B Street comes with a few maps, each defined by a bounding/clipping polygon
for some portion of Seattle. Each map has these objects:
## How is a map used?
- **Roads**: A single road connects two intersections, carrying OSM metadata and
containing some child lanes.
- **Lanes**: An individual lane of traffic. Driving (any vehicle), bus-only, and
bike-only lanes have a direction. On-street parking lanes don't allow any
movement, and they have some number of parking spots. Sidewalks are
bidirectional.
- **Intersections**: An intersection has references to all of the incoming and
outgoing lanes. Most intersections have a stop sign or traffic signal policy
controlling movement through it.
- **Border** intersections on the edge of the map are special places where
agents may appear or disappear.
- **Turns**: A turn connects one lane to another, via some intersection.
(Sidewalks are bidirectional, so specifying the intersection is necessary to
distinguish crosswalks at each end of a sidewalk.)
- **Buildings**: A building has a position, OSM metadata, and a **front path**
connecting the edge of the building to the nearest sidewalk. Most trips in A/B
Street begin and end at buildings. Some buildings also contain a number of
off-street parking spots.
- **Area**: An area has geometry and OSM metadata and represents a body of
water, forest, park, etc. They're just used for drawing.
- **Bus stop**: A bus stop is placed some distance along a sidewalk, with a
pointer to the position on the adjacent driving or bus lane where a bus stops
for pick-up.
- **Bus route**: A bus route has a name and a list of stops that buses will
cycle between. In the future, they'll include information about the
frequency/schedule of the route.
- **Parking lot**: A parking lot is connected to a road, has a shape, and has
some internal driving "aisles." The number and position of individual parking
spots is auto-generated.
Unlike some GIS systems, maps don't use any kind of database -- they're just a
file, anywhere from 1 to ~500MB (depending on the size of their boundary). Once
loaded into memory, different objects from the map can be accessed directly,
along with a large API to perform various queries.
## Coordinate system
Most of the map's API is read-only; once built, a map doesn't change until
user-created edits are applied.
A/B Street converts (longitude, latitude) coordinates into a simpler form.
The pipeline to import a map from OSM data (and also optional supplementary,
city-specific data) is complex and may take a few minutes to run, but it happens
once offline. Applications using maps just read the final file.
- An (x, y) point starts with the top-left of the bounding polygon as the
origin. Note this is screen drawing order, not a Cartesian plane (with Y
increasing upwards) -- so angle calculations account for this.
- The (x, y) values are f64's trimmed to a few decimal places, with way more
precision than is really needed. These might become actual fixed-point
integers later, but for now, a `Pt2D` skirts around Rust's limits on f64's by
guaranteeing no NaN's or infinities and thus providing the full `Eq` trait.
- A few places in map conversion compare points using different thresholds,
usually below 1 meter. Ideally these epsilon comparisons could be eliminated
in favor of a fixed-point integer representation, but for now, explicit
thresholds are useful.
## Features
## Invariants
Why use A/B Street's map model instead of processing OSM directly?
Ideally, the finalized maps would satisfy a list of invariants, simplifying the
traffic simulation and drawing code built on top. But the input data is quite
messy and for now, most of these aren't quite guaranteed to be true.
TODO: Order these better. For each one, show before/after pictures
- Some minimum length for lanes and turns. Very small lanes can't be drawn, tend
to break intersection polygons, and may lead to gridlocked traffic.
- Some guarantees that positions along adjacent lanes actually match up, even
though different lanes on the same road may have different lengths. Examples
include the position of a bus stop on the sidewalk and bus lane matching up.
- Additionally, parking lanes without an adjacent driving lane or bus stops
without any driving or bus lanes make no sense and should never occur.
- Connectivity -- any sidewalk should be reachable from any other, and most
driving lanes should be accessible from any others. There are exceptions due
to border intersections -- if a car spawns on a highway along the border of
the map, it may be forced to disappear on the opposite border of the map, if
the highway happens to not have any exits within the map boundary.
### Area clipping
## Connectivity
Bodies of water, forests, parks, and other areas are represented in OSM as
relations, requiring the user to stitch together multiple polylines in undefined
orders and handle inner holes. A/B Street maps handle all of that, and also clip
the area's polygon to the boundary of the entire map -- including coastlines.
For a single mode, each lane is connected to two intersections. Turns connect
two lanes. There are no turns between sidewalks and driving/bike/bus lanes.
### Road and intersection geometry
All buildings and parking lots have driveways. This must connect to a sidewalk,
allowing pedestrians to enter/exit that object. The driveway OPTIONALLY connects
to the nearest driveable lane. This allows cars to enter/exit that object for
parking.
OSM represents roads as a polyline of the physical center of the road. A/B
Street infers the number and type of lanes from OSM metadata, then creates
individual lanes of appropriate width, each with a center-line and polygon for
geometry. At intersections, the roads and lanes are "trimmed back" to avoid
overlapping, and the "common area" becomes the intersection's polygon. This
heuristic process is reasonably robust to complex shapes, with special treatment
of highway on/off-ramps, although it does still have some bugs.
Public transit stops are located somewhere on a sidewalk. They're associated
with a driveable position where the bus or train stops. In the future, this will
need to account for dedicated surface-level platforms and for underground
transit stations, likely associated with a building.
### Turns
There's a concept of "parking blackholes." If you treat every road as
bidirectional without access restrictions, then the graph is connected. But the
more detailed view has to factor in one-way roads and things near the map
border. These blackholes influence where cars will try to look for parking
(since we don't want them entering a blackhole and getting stuck) and also, for
temporary/unintentional reasons, where pedestrian<->bicycle transitions will
happen.
At each intersection, A/B Street infers all legal movements between vehicle
lanes and sidewalks. This process makes use of OSM metadata about turn lanes,
inferring reasonable defaults for multi-lane roads. OSM turn restriction
relations, which may span a sequence of several roads to describe U-turns around
complex intersections, are also used.
### Parking lots
OSM models parking lots as areas along with the driveable aisles. Usually the
capacity of a lot isn't tagged. A/B Street automatically fills paring lots with
individual stalls along the aisles, estimating the capacity just from this
geometry.
### Stop signs
At unsignalized intersections, A/B Street infers which roads have to stop, and
which have right-of-way.
### Traffic signals
OSM has no way to describe how traffic signals are configured. A/B Street models
fixed-timer signals, automatically inferring the number of phases, their
duration, and the movements that are prioritized and permitted during each
phase.
### Pathfinding
A/B Street can determine routes along lanes and turns for vehicles and
pedestrians. These routes obey OSM's turn restriction relations that span
multiple road segments. They also avoid roads that're tagged as not allowing
through-traffic, depending on the route's origin and destination and vehicle
type. The pathfinding optionally makes use of contraction hierarchies to greatly
speed up query performance, at the cost of a slower offline importing process.
### Bridge z-ordering
OSM tags bridges and tunnels, but the roads that happen to pass underneath
bridges aren't mapped. A/B Street detects these and represents the z-order for
drawing.
### Buildings
Similar to areas, A/B Street consolidates the geometry of OSM buildings, which
may be split into multiple polygons. Each building is also associated with the
nearest driveable lane and sidewalk, and metadata is used to infer a land-use
(like residential and commercial) and commercial amenities available.
### Experimental: public transit
A/B Street uses bus stops and route relations from OSM to build a model of
public transit routes. OSM makes few guarantees about how the specifics of the
route are specified, but A/B Street produces specific paths, handling clipping
to the map boundary.
... All of this isn't the case yet, but it's a WIP!
### Experimental: separated cyclepaths, tramways, and walking paths
Some cyclepaths, tram lines, and footpaths in OSM are tagged as separate ways,
with no association to a "main" road. Sometimes this is true -- they're
independent trails that only occasionally cross roads. But often they run
alongside a road. A/B Street attempts to detect these and "snap" them to the
main road as extra lanes.
... But this doesn't work yet at all.

105
book/src/map/details.md Normal file
View File

@ -0,0 +1,105 @@
# Map model details
A/B Street builds a rich representation of a city map using OpenStreetMap (OSM)
and other sources. This chapter describes how.
TODO: Integrate pictures from
[these slides](https://docs.google.com/presentation/d/1cF7qFtjAzkXL_r62CjxBvgQnLvuQ9I2WTE2iX_5tMCY/edit?usp=sharing).
[This recorded presentation](https://youtu.be/chYd5I-5oyc?t=439) covers some of
this.
## The map
A single city is broken down into different pieces...
A/B Street comes with a few maps, each defined by a bounding/clipping polygon
for some portion of Seattle. Each map has these objects:
- **Roads**: A single road connects two intersections, carrying OSM metadata and
containing some child lanes.
- **Lanes**: An individual lane of traffic. Driving (any vehicle), bus-only, and
bike-only lanes have a direction. On-street parking lanes don't allow any
movement, and they have some number of parking spots. Sidewalks are
bidirectional.
- **Intersections**: An intersection has references to all of the incoming and
outgoing lanes. Most intersections have a stop sign or traffic signal policy
controlling movement through it.
- **Border** intersections on the edge of the map are special places where
agents may appear or disappear.
- **Turns**: A turn connects one lane to another, via some intersection.
(Sidewalks are bidirectional, so specifying the intersection is necessary to
distinguish crosswalks at each end of a sidewalk.)
- **Buildings**: A building has a position, OSM metadata, and a **front path**
connecting the edge of the building to the nearest sidewalk. Most trips in A/B
Street begin and end at buildings. Some buildings also contain a number of
off-street parking spots.
- **Area**: An area has geometry and OSM metadata and represents a body of
water, forest, park, etc. They're just used for drawing.
- **Bus stop**: A bus stop is placed some distance along a sidewalk, with a
pointer to the position on the adjacent driving or bus lane where a bus stops
for pick-up.
- **Bus route**: A bus route has a name and a list of stops that buses will
cycle between. In the future, they'll include information about the
frequency/schedule of the route.
- **Parking lot**: A parking lot is connected to a road, has a shape, and has
some internal driving "aisles." The number and position of individual parking
spots is auto-generated.
## Coordinate system
A/B Street converts (longitude, latitude) coordinates into a simpler form.
- An (x, y) point starts with the top-left of the bounding polygon as the
origin. Note this is screen drawing order, not a Cartesian plane (with Y
increasing upwards) -- so angle calculations account for this.
- The (x, y) values are f64's trimmed to a few decimal places, with way more
precision than is really needed. These might become actual fixed-point
integers later, but for now, a `Pt2D` skirts around Rust's limits on f64's by
guaranteeing no NaN's or infinities and thus providing the full `Eq` trait.
- A few places in map conversion compare points using different thresholds,
usually below 1 meter. Ideally these epsilon comparisons could be eliminated
in favor of a fixed-point integer representation, but for now, explicit
thresholds are useful.
## Invariants
Ideally, the finalized maps would satisfy a list of invariants, simplifying the
traffic simulation and drawing code built on top. But the input data is quite
messy and for now, most of these aren't quite guaranteed to be true.
- Some minimum length for lanes and turns. Very small lanes can't be drawn, tend
to break intersection polygons, and may lead to gridlocked traffic.
- Some guarantees that positions along adjacent lanes actually match up, even
though different lanes on the same road may have different lengths. Examples
include the position of a bus stop on the sidewalk and bus lane matching up.
- Additionally, parking lanes without an adjacent driving lane or bus stops
without any driving or bus lanes make no sense and should never occur.
- Connectivity -- any sidewalk should be reachable from any other, and most
driving lanes should be accessible from any others. There are exceptions due
to border intersections -- if a car spawns on a highway along the border of
the map, it may be forced to disappear on the opposite border of the map, if
the highway happens to not have any exits within the map boundary.
## Connectivity
For a single mode, each lane is connected to two intersections. Turns connect
two lanes. There are no turns between sidewalks and driving/bike/bus lanes.
All buildings and parking lots have driveways. This must connect to a sidewalk,
allowing pedestrians to enter/exit that object. The driveway OPTIONALLY connects
to the nearest driveable lane. This allows cars to enter/exit that object for
parking.
Public transit stops are located somewhere on a sidewalk. They're associated
with a driveable position where the bus or train stops. In the future, this will
need to account for dedicated surface-level platforms and for underground
transit stations, likely associated with a building.
There's a concept of "parking blackholes." If you treat every road as
bidirectional without access restrictions, then the graph is connected. But the
more detailed view has to factor in one-way roads and things near the map
border. These blackholes influence where cars will try to look for parking
(since we don't want them entering a blackhole and getting stuck) and also, for
temporary/unintentional reasons, where pedestrian<->bicycle transitions will
happen.

View File

@ -1,6 +1,23 @@
# Importing
Overview of the process. The importer tool.
This chapter describes the process of transforming OSM extracts into A/B
Street's map model. The steps are:
Don't be afraid of how complicated this seems. It started simple -- just bring
in OSM roads, chop into pieces, generate turns.
1. A large .osm file is clipped to a hand-drawn boundary region, using
`osmconvert`
2. The `convert_osm` crate reads the clipped `.osm`, and a bunch of optional
supplementary files, and produces a `RawMap`
3. Part of the `map_model` crate transforms the `RawMap` into the final `Map`
4. Other applications read and use the `Map` file
The `importer` crate orchestrates these steps, along with automatically
downloading any missing input data.
The rest of these sections describe each step in a bit more detail. Keeping the
docs up-to-date is hard; the best reference is the code, which is hopefully
organized clearly.
Don't be afraid of how complicated this pipeline seems -- each step is
relatively simple. If it helps, imagine how this started -- just chop up OSM
ways into road segments, infer lanes for each road, and infer turns between the
lanes.

View File

@ -1,46 +1,43 @@
# From OSM to RawMap (`convert_osm` crate)
The first phase of map building reads in data from OSM files and a few others,
producing a serialized `RawMap`. Importing all maps (one for each pre-defined
bounding polygon) takes a few minutes. Players don't see this cost; it only
takes a few seconds to load a serialized map.
producing a serialized `RawMap`.
- `osm.rs`: Read .osm, extracting the points for road-like ways, buildings, and
areas
- Areas usually come from a relation of multiple ways, with the points out of
Only major steps are described; see the code for the rest.
## extract.rs
Read .osm, extracting the points for road-like ways, buildings, and areas
- Areas usually come from a relation of multiple ways, with the points out of
order. Gluing all the points together fails when the .osm has some ways
clipped out. In that case, try to trace along the map boundary if the
partial area intersects the boundary in a clear way. Otherwise, just use a
straight line to try to close off the polygon.
- Also read traffic signal locations and turn restrictions between OSM ways
- `split_ways.rs`: Split OSM ways into road segments
- OSM ways cross many intersections, so treat points with multiple ways and
the points at the beginning and end of a way as intersections, then split
the way into road segments between two intersections.
- This phase remembers which road segment is the beginning and end of the OSM
clipped out. In that case, try to trace along the map boundary if the partial
area intersects the boundary in a clear way. Otherwise, just use a straight
line to try to close off the polygon.
- Also read traffic signal locations and turn restrictions between OSM ways
## split_ways.rs
Split OSM ways into road segments
- OSM ways cross many intersections, so treat points with multiple ways and the
points at the beginning and end of a way as intersections, then split the way
into road segments between two intersections.
- This phase remembers which road segment is the beginning and end of the OSM
way, for per-lane turn restrictions later
- Apply turn restrictions between roads here. Since OSM ways cross many
intersections, the turn restrictions only apply to one particular road
segment that gets created from the way. Make sure the destination of the
restriction is actually incident to a particular source road.
- `clip.rs`: Clip the map to the boundary polygon
- Osmosis options in `import.sh` preserve ways that cross the boundary
- Trim roads that cross the boundary. There may be cases where a road dips out
- Apply turn restrictions between roads here. Since OSM ways cross many
intersections, the turn restrictions only apply to one particular road segment
that gets created from the way. Make sure the destination of the restriction
is actually incident to a particular source road.
## clip
Clip the map to the boundary polygon
- `osmconvert` options preserve ways that cross the boundary
- Trim roads that cross the boundary. There may be cases where a road dips out
of bounds, then immediately comes back in. Disconnecting it isn't ideal, but
it's better to manually tune the boundary polygon when this happens than try
to preserve lots of out-of-bounds geometry.
- Area polygons are intersected with the boundary polygon using the `clipping`
- Area polygons are intersected with the boundary polygon using the `clipping`
crate
- `lib.rs`: Remove cul-de-sacs (roads that begin and end at the same
intersection), because they mess up parking hints and pathfinding.
- `lib.rs`: Apply parking hints from a King County GIS blockface dataset
- Match each blockface to the nearest edge of a road
- Interpret the metadata to assign on-street parking there or not
- `lib.rs`: Apply offstreet parking hints from a King County GIS dataset
- Match each point to the building containing it, plumbing through the number
of spots
- `lib.rs`: **Disabled**: Apply sidewalk presence hints from a King County GIS
dataset
- Match each sidewalk line to the nearest edge of a road
- Update the road to have a sidewalk on none, one, or both sides
- `lib.rs` using the `srtm` module: Load (extremely poor quality) elevation data