Pytrafficutils - Process traffic trajectories, extract surrogate safety metrics (SSMs), represent traffic agents.

A personal collection of tools frequently used for traffic analysis and simulation:

• Represent traffic agents (moving or fixed, road users or obstacles, cars or bicycles) and the trajectories of their kinematic states.
• Provide 2D polygons representing the footprints of different kinds of traffic agents (rectangls, diamonds).
• Calculate the Surrogate Safety Metrics (SSMs) Time-To-Collsision (TTC) and Post-Encroachment Time (PET).

Disclaimer

The package is under development. It may contain bugs and sections of unused or insensible code. Major changes to this package are planned for the time to come. A proper API documentation is still missing. Refer to the docstrings for API information.

Installation

1. Clone this repository.

   git clone  https://github.com/chris-konrad/pytrafficutils.git
   

2. Install the package and it's dependencies. Refer to pyproject.toml for an overview of the dependencies.

   cd ./pytrafficutils
   pip install . 
   

Modules

This package has several modules to define traffic agents, their trajectories, their footprints, and to process SSMs

agents

A module defining traffic agents. A traffic agent can be moving or stationary and has a footprint defining it`s 2D shape. Agents hold their trajectory in time, can return their state at a requested time and extrapolate their states from a requested point in time. Several agents are available:

pytrafficutils.agents.Agent() Agent base class. Not designed to create instances.

pytrafficutils.agents.MovingAgent(Agent) Class for agents that move at least during a part of their time in the traffic environment.

pytrafficutils.agents.Bicycle(MovingAgent) Class for moving bicycles / cyclists with a diamond footprint. Moving agents can linearly extrapolate their state in time.

pytrafficutils.agents.Car(MovingAgent) Class for moving cars with a rectangular footprint.

pytrafficutils.agents.StationaryAgent(Agent) Class for agents that do not move during their entire time in the traffic environment. For example, a parked vehicle or fixed obstacle.

pytrafficutils.agents.RectObstacle(StationaryAgent) Class for fixed rectangular obstacles.

footprints

A module defining the footprints of traffic agents as 2D shapes. Footprints are defined as a polygon (set of nodes relative to the center/reference point/anchor). Footprints can be translated / rotated to different positions and support plotting. Several footprints are available:

pytrafficutils.footprints.Footprint() Footprint base class. Not designed to create instances.

pytrafficutils.footprints.RectangularFootprint(Footprint) A footprint shaped like a rectangle for use as cars / trucks / busses.

pytrafficutils.footprints.DiamondFootprint(Footprint) A footprint shaped like a diamond / rhombus for use as bicycles.

ssm

A module to evaluate surrogate safety metrics (SSMs) using the pytrafficutils.ssm.SurrogateSafetyProcessor() class. This class takes a list of traffic agents pytrafficutils.agents.Agent() and evaluates SSMs on the trajectories of their kinematic states. Respects footprints where appropriate. Uses matrix operations where possible for computational efficiency. Currently supports the calculation of:

• Time-To-Collision (TTC): Calculates the TTC using linear extrapolation of the velocity (i.e., speed and heading). Reports the minimum TTC and classifies interactions as conflicts using (multiple) configurable TTC thresholds.

Warning

Currently, only TTC is supported by pytrafficutils.ssm.SurrogateSafetyProcessor(). The Post-Encroachment Time (PET) can be calculated using the function pytrafficutils.ssm.pet() for specific trajectory pairs, but this is not yet integrated with SurrogateSafetyProcessor() and does not consider footprints.

Authors

• Christoph M. Konrad, [email protected]

License

This package is licensed under the terms of the MIT license.