trajectory_supervisor.supervisor_modules.supmod_RSS.src.supmod_RSS module

class trajectory_supervisor.supervisor_modules.supmod_RSS.src.supmod_RSS.SupModRSS(supmod_config_path: str, veh_params: dict, localgg: numpy.ndarray)[source]

Bases: object

Class handling the safety rating based on the RSS principle presented by Shalev Shwartz.

NOTE: This SupMod is a drafted version and in BETA state, handle with care.

Note

For further details:

  1. Shalev-Shwartz, S. Shammah, and A. Shashua, “On a Formal Model of Safe and Scalable Self-driving Cars,” 2017.

Authors
Created on

15.05.2019

Init the RSS SupMod.

Parameters

  • localgg – maximum g-g-values at a certain position on the map (currently, only the max a used)

  • supmod_config_path – path pointing to the config file hosting relevant parameters

  • veh_params – dict of vehicle parameters; must hold the following keys: veh_width - width of the ego-vehicle [in m] veh_length - length of the ego-vehicle [in m]

calc_score(ego_pos: numpy.ndarray, ego_heading: float, ego_curv: float, ego_vel: float, objects: dict)[source]

This function uses kinematic data of the ego as well as the object vehicle and use the data to calculate the safety of the vehicle. Since the curvature of the object vehicle is needed but not provided, the path of the object vehicle is used as the input of the function in order to calculate the curvature of the vehicle. Due to the noise, the calculated curvature might not be that accurate, so there’s also a flag which gives an option to use the calculated curvature or not.

Parameters

  • ego_pos – numpy array holding global x and y of the ego-vehicle

  • ego_heading – heading of the ego-vehicle in the global coordinate frame

  • ego_curv – curvature of the path at the current position

  • ego_vel – absolute velocity of the ego-vehicle

  • objects – dict of objects, each holding a dict (with at least: ‘X’, ‘Y’, ‘theta’, ‘v_x’)

Returns

  • safety_rss - the binary result of the safety assessment(true: safe, false: not safe)

  • safety_parameters - a descriptive dict holding more detailed log / debugging information

update_map(bound_l: numpy.ndarray, bound_r: numpy.ndarray) → None[source]

Update the internal map representation.

Parameters

  • bound_l – left bound of the track

  • bound_r – right bound of the track