#376 Road Rules definitions in Carla

doc/research/paf24/general/carla_traffic_rules.md

@@ -0,0 +1,172 @@
+# Traffic Rules and Principles in CARLA
+
+## Table of Contents
+
+- [Table of Contents](#table-of-contents)
+- [Introduction](#introduction)
+- [Key Traffic Principles and Complex Interactions](#key-traffic-principles-and-complex-interactions)
+- [Using the Traffic Manager in CARLA](#using-the-traffic-manager-in-carla)
+- [CARLA Maps and Their Traffic Scenarios](#carla-maps-and-their-traffic-scenarios)
+  - [Overview of CARLA Maps](#overview-of-carla-maps)
+  - [Rural and Highway Scenarios](#rural-and-highway-scenarios)
+- [Road Network Creation with ASAM OpenDRIVE®](#road-network-creation-with-asam-opendrive)
+- [Units of Measurement](#units-of-measurement)
+- [References](#references)
+
+## Introduction
+
+CARLA (Car Learning to Act) is an open-source simulator for testing and developing autonomous driving systems. It integrates traffic rules to create a flexible environment, letting users adjust regulations and design custom maps to evaluate driving behaviors without following specific national laws.
+
+## Key Traffic Principles and Complex Interactions
+
+- **Traffic Signals**: Vehicles must stop at red lights and go on green.
+
+- **Stop Signs**: Vehicles must stop at stop signs and yield to pedestrians
+  and other cars.
+
+- **Lane Management**: Cars should remain in their designated lanes.
+
+- **Speed Limits**: Different speed limits in various areas of the map allow
+  testing of speed control and adherence to traffic laws.
+
+- **Pedestrian Rights**: Vehicles must yield to pedestrians at crosswalks,
+  which is vital for pedestrian detection and ensuring safe stops.
+
+- **Intersection Management**: Vehicles must navigate intersections
+  effectively, managing stop lights and stop signs while yielding when
+  necessary to keep traffic moving.
+  *Note: According to the [CARLA documentation](https://carla.readthedocs.io/en/latest/adv_traffic_manager/) the Traffic Manager uses its own priority system at junctions. This is a known problem in the simulation.*
+
+- **Roundabout Navigation**: Vehicles should yield to cars already in the
+  roundabout. This tests their ability to navigate circular roads and merge
+  correctly.
+
+- **Crosswalk Behavior**: Vehicles need to detect and stop for pedestrians
+  at crosswalks, highlighting the importance of pedestrian safety.
+
+- **Bicycle Lane Navigation**: Vehicles should be able to detect and safely
+  navigate around cyclists in traffic.
+
+- **Parking Maneuvers**: Vehicles are tested on their ability to park in
+  tight spaces, simulating real-life parking challenges.
+
+## Using the Traffic Manager in CARLA
+
+The Traffic Manager (TM) in CARLA controls vehicle behavior during simulations. Below are key considerations and customization options:
+
+### General Behavior Patterns
+
+- **Non-goal-oriented Navigation**: Vehicles follow dynamically generated trajectories, choosing paths randomly at junctions. Their navigation is endless.
+- **Speed Adjustment**: Vehicles maintain a target speed of 70% of the speed limit unless explicitly modified.
+- **Junction Priority**: Junction handling does not strictly follow traffic regulations. Instead, the TM applies its own priority system, leading to potential issues, such as vehicles inside a roundabout yielding to incoming traffic. This limitation is being addressed.
+
+### Customizing TM Behavior via Python API
+
+The TM behavior can be tailored using the Python API. Key functionalities include:
+
+#### General
+
+- Create a TM instance connected to a specific port.
+- Retrieve the port where the TM is connected.
+
+#### Safety Conditions
+
+- Set a minimum stopping distance for vehicles (globally or individually).
+- Adjust the desired speed as a percentage of the speed limit.
+- Reset traffic lights programmatically.
+
+#### Collision Management
+
+- Enable or disable collisions between vehicles and specific actors.
+- Configure vehicles to ignore:
+  - Other vehicles
+  - Pedestrians
+  - Traffic lights
+
+#### Lane Management
+
+- Force lane changes, overriding collision considerations.
+- Enable or disable lane changes for specific vehicles.
+
+#### Hybrid Physics Mode
+
+- Enable or disable hybrid physics mode for improved performance.
+- Adjust the radius within which physics is applied.
+
+For more details, refer to the [TM section of the Python API documentation](https://carla.readthedocs.io/en/latest/adv_traffic_manager/#vehicle-behavior-considerations).
+
+## CARLA Maps and Their Traffic Scenarios
+
+CARLA has 12 maps, each designed to test different aspects of self-driving,
+from basic navigation to complex urban and rural situations.
+Users can also create custom maps, allowing for additional testing scenarios
+tailored to specific needs.
+
+### Overview of CARLA Maps
+
+- **Town 1**: A small town with T-junctions and bridges, focusing on managing
+  intersections and staying in lanes.
+
+- **Town 2**: Similar to Town 1, this town includes shops and homes, testing
+  navigation and awareness of the surroundings.
+
+- **Town 3**: A larger city area with roundabouts and underpasses, assessing
+  navigation in roundabouts and speed adjustments.
+
+- **Town 4**: A small town with a figure-eight road, testing navigation in
+  multiple lanes and interactions with pedestrians.
+
+- **Town 5**: An urban area with multilane roads and a highway, focusing on
+  highway driving and commercial strategies.
+
+- **Town 6**: A less populated town with unique junctions, testing navigation
+  in multiple lanes and the use of slip roads.
+
+- **Town 7**: A rural community with no marked roads, which tests navigation
+  and parking in residential areas.
+
+- **Town 8 & Town 9**: Hidden maps used for the Leaderboard challenge.
+
+- **Town 10**: An inner-city area with various junctions and pedestrian
+  activity, testing skills in complex traffic situations.
+
+- **Town 11**: A basic map for testing navigation in open areas with fewer
+  visual guides.
+
+- **Town 12**: A large map based on Amarillo, Texas, featuring urban,
+  residential, and rural areas to test navigation and visual skills across
+  diverse environments.
+
+### Rural and Highway Scenarios
+
+Maps like **Town 7** and **Town 12** focus on rural and highway driving,
+simulating open and high-speed areas. They test long-distance driving,
+speed management, and safe overtaking, requiring drivers to adapt to
+unmarked roads and agricultural settings.
+
+## Road Network Creation with ASAM OpenDRIVE®
+
+Road networks in CARLA are built using **ASAM OpenDRIVE®**, which standardizes
+road descriptions:
+
+- **Hierarchical Structure**: Roads are organized into nodes for better
+  integration.
+
+- **Reference Line**: Each road has a reference line that defines its shape.
+
+- **Interconnectivity**: Roads can connect to each other for realistic
+  traffic flow.
+
+## Units of Measurement
+
+To keep tests consistent, CARLA uses standardized units:
+
+- **Speed**: Measured in meters per second (m/s).
+
+- **Distance**: Measured in meters.
+
+## References
+
+- CARLA Documentation: [CARLA](https://carla.readthedocs.io/en/latest/)
+- Traffic Manager Behavior Considerations: [CARLA TM Documentation](https://carla.readthedocs.io/en/latest/adv_traffic_manager/#vehicle-behavior-considerations)
+- ASAM OpenDRIVE® Standard: [ASAM OpenDRIVE](https://www.asam.net/standards/detail/opendrive/)