src package¶

Subpackages¶

Submodules¶

src.const module¶

src.ngsim_trajdata module¶

class src.ngsim_trajdata.NGSIMTrajdata(file_path: str)[source]¶

Bases: object

src.ngsim_trajdata.car_df_index(trajdata: src.ngsim_trajdata.NGSIMTrajdata, carid: int, frame: int)[source]¶

given frame and carid, find index of car in trajdata Returns 0 if it does not exist

src.ngsim_trajdata.carid_set(trajdata: src.ngsim_trajdata.NGSIMTrajdata)[source]¶

get all the vehicle id :param trajdata: NGSIMTrajdata :return: all the vehicle id

src.ngsim_trajdata.carsinframe(trajdata: src.ngsim_trajdata.NGSIMTrajdata, frame: int)[source]¶

get the vehicles in a given frame :param trajdata: NGSIMTrajdata :param frame: frame index :return: vehicles in the frame, [] if no such frame

src.ngsim_trajdata.first_carid(trajdata: src.ngsim_trajdata.NGSIMTrajdata, frame: int)[source]¶

get the first vehicle’s id in a given frame :param trajdata: NGSIMTrajdata :param frame: frame :return: the first vehicle’s id in a given frame

src.ngsim_trajdata.get_frame_range(trajdata: src.ngsim_trajdata.NGSIMTrajdata, carid: int)[source]¶

get the frame range for a car given the car id :param trajdata: NGSIMTrajdata :param carid: car id :return: frame range

src.ngsim_trajdata.iscarinframe(trajdata: src.ngsim_trajdata.NGSIMTrajdata, carid: int, frame: int)[source]¶

check if a car is in a frame given the car id and the frame :param trajdata: NGSIMTrajdata :param carid: car id :param frame: frame index :return: true if in the frame, false otherwise

src.ngsim_trajdata.load_ngsim_trajdata(filepath: str)[source]¶

load NGSIMTrajdata from a file given the file path :param filepath: file path :return: the loaded NGSIMTrajdata object

src.ngsim_trajdata.nth_carid(trajdata: src.ngsim_trajdata.NGSIMTrajdata, frame: int, n: int)[source]¶

get the nth vehicle’s id in a given frame :param trajdata: NGSIMTrajdata :param frame: frame :param n: n :return: the nth vehicle’s id in a given frame

src.roadway module¶

class src.roadway.NGSIMRoadway(name: str, boundaries: list, centerlines: list)[source]¶

Bases: object

class src.roadway.RoadwayInputParams(filepath_boundaries: str, filepath_centerlines: str)[source]¶

Bases: object

src.roadway.convert_curves_feet_to_meters(roadWay: src.Roadway.roadway.Roadway)[source]¶

convert all feet unit to meters :param roadWay: Roadway object :return: converted Roadway object

src.roadway.get_segid(lane: list)[source]¶

Parameters

lane – list of CurvePt

Returns

get the segment id of corresponding lane

src.roadway.integrate(centerline_fn: str, boundary_fn: str, dist_threshold_lane_connect: float = 2.0, desired_distance_between_curve_samples: float = 1.0)[source]¶

Parameters

• centerline_path – center line file path

• boundary_path – boundary file path

• dist_threshold_lane_connect – [m]

• desired_distance_between_curve_samples – [m]

Returns

src.roadway.read_boundaries(filepath_boundaries: str)[source]¶

load boundaries from a file given the file path :param filepath_boundaries: the file path of the boundary :return: the loaded boundary

src.roadway.read_centerlines(filepath_centerlines: str)[source]¶

load centerlines from a file given the file path :param filepath_centerlines: the file path of the center line :return: the loaded center line

src.roadway.read_roadway(input_params: src.roadway.RoadwayInputParams)[source]¶

load roadway :param input_params: RoadwayInputParams :return: the loaded NGSIMRoadway object

src.splines module¶

src.splines.arclength_1(spline_coeffs_x, spline_coeffs_y, t_min: float = 0.0, t_max: float = 1.0, n_intervals: int = 100)[source]¶

src.splines.arclength_2(spline_coeffs_x, spline_coeffs_y, t_min: float = 0.0, t_max: float = None, n_intervals_per_segment: int = 100)[source]¶

src.splines.calc_curve_length_1(spline_coeffs_x, spline_coeffs_y, n_intervals: int = 100)[source]¶

src.splines.calc_curve_length_2(spline_coeffs_x, spline_coeffs_y, n_intervals_per_segment: int = 100)[source]¶

src.splines.calc_curve_param_given_arclen(spline_coeffs_x, spline_coeffs_y, s_arr, max_iterations: int = 50, curve_length: float = None, epsilon: float = 0.0001, n_intervals_in_arclen: int = 100)[source]¶

src.splines.fit_cubic_spline(pts, open: bool = True)[source]¶

src.splines.sample_spline_1(spline_coeffs, t: float)[source]¶

src.splines.sample_spline_2(spline_coeffs, t_arr)[source]¶

src.splines.sample_spline_curvature_1(spline_coeffs_x, spline_coeffs_y, t: float)[source]¶

src.splines.sample_spline_curvature_2(spline_coeffs_x, spline_coeffs_y, t_arr)[source]¶

src.splines.sample_spline_derivative(spline_coeffs, t: float)[source]¶

src.splines.sample_spline_derivative2(spline_coeffs, t: float)[source]¶

src.splines.sample_spline_derivative_of_curvature_1(spline_coeffs_x, spline_coeffs_y, t, stepsize=0.0001)[source]¶

src.splines.sample_spline_derivative_of_curvature_2(spline_coeffs_x, spline_coeffs_y, t_arr, stepsize=0.0001)[source]¶

src.splines.sample_spline_speed_1(spline_coeffs_x, spline_coeffs_y, t: float)[source]¶

src.splines.sample_spline_speed_2(spline_coeffs_x, spline_coeffs_y, t: float)[source]¶

src.splines.sample_spline_theta_1(spline_coeffs_x, spline_coeffs_y, t: float, stepsize=0.0001)[source]¶

src.splines.sample_spline_theta_2(spline_coeffs_x, spline_coeffs_y, t_arr)[source]¶

src.trajdata module¶

class src.trajdata.FilterTrajectoryResult(trajdata: src.ngsim_trajdata.NGSIMTrajdata, carid: int)[source]¶

Bases: object

src.trajdata.convert(tdraw: src.ngsim_trajdata.NGSIMTrajdata, roadway: src.Roadway.roadway.Roadway)[source]¶

Parameters

• tdraw – trajectory raw data

• roadway – roadway class

Returns

ListRecord(), a preprocessed and integrated version of tdraw and roadway

src.trajdata.convert_raw_ngsim_to_trajdatas()[source]¶

convert the raw trajectory data and roadway to a integrated version and save to a file :return: no return

src.trajdata.copy(trajdata: src.ngsim_trajdata.NGSIMTrajdata, ftr: src.trajdata.FilterTrajectoryResult)[source]¶

src.trajdata.filter_given_trajectory(trajdata: src.ngsim_trajdata.NGSIMTrajdata, carid: int)[source]¶

Parameters

• trajdata – trajectory data file, NGSIMTrajdata object

• carid – the id of the car that we want to filter

Returns

a smoothed trajectory

src.trajdata.filter_trajectory(ftr: src.trajdata.FilterTrajectoryResult, v: src.trajectory_smoothing.VehicleSystem = <src.trajectory_smoothing.VehicleSystem object>)[source]¶

src.trajdata.get_corresponding_roadway(filename: str)[source]¶

Parameters

filename – the path of the saved roadway file

Returns

the roadway class object

src.trajdata.load_ngsim_trajdata(filepath: str, autofilter: bool = True)[source]¶

Parameters

• filepath – the path of the raw trajectory data file

• autofilter – indicates do the filter or not

Returns

the filtered(or not) data class

src.trajdata.load_trajdata(filepath: str)[source]¶

load trajdata from the file given the file path :param filepath: file path :return: ListRecord object

src.trajdata.symmetric_exponential_moving_average(arr: list, T: float, dt: float = 0.1)[source]¶

src.trajectory_smoothing module¶

src.trajectory_smoothing.EKF(v: src.trajectory_smoothing.VehicleSystem, mu: list, Cov, u: list, o: list)[source]¶

Parameters

• v – VehicleSystem

• mu – mean of belief at time t-1

• Cov – cov of belief at time t-1

• u – next applied control

• o – observation for time t

Returns

mu_next np.array, Cov_next np.array

class src.trajectory_smoothing.SimulationResults(x_arr, z_arr, u_arr, mu_arr, Cov_arr)[source]¶

Bases: object

class src.trajectory_smoothing.VehicleSystem(process_noise: float = 0.077, observation_noise: float = 16.7, control_noise_accel: float = 16.7, control_noise_turnrate: float = 0.46)[source]¶

Bases: object

used to do trajectory smoothing

src.trajectory_smoothing.compute_dynamics_jacobian(v: src.trajectory_smoothing.VehicleSystem, x: list, u: list)[source]¶

Computes the dynamics Jacobian Inputs:

• ν is the vehicle

• x is the vehicle state [x,y,θ,v]ᵀ

• u is the control [a,ω]

src.trajectory_smoothing.compute_observation_jacobian(v: src.trajectory_smoothing.VehicleSystem, x: list)[source]¶

Computes the observation Jacobian (H matrix) Inputs:

• ν is the vehicle concrete type

• x is the state estimate [x,y,θ,v]ᵀ

src.trajectory_smoothing.draw_obs_noise(v: src.trajectory_smoothing.VehicleSystem)[source]¶

src.trajectory_smoothing.draw_proc_noise(v: src.trajectory_smoothing.VehicleSystem)[source]¶

src.trajectory_smoothing.get_control_noise_in_control_space(v: src.trajectory_smoothing.VehicleSystem, u: list)[source]¶

To derive the covariance of the additional motion noise, we first determine the covariance matrix of the noise in control space Inputs:

• ν is the vehicle concrete type

• u is the control [a,ω]ᵀ

src.trajectory_smoothing.get_observation_noise_covariance(v: src.trajectory_smoothing.VehicleSystem)[source]¶

src.trajectory_smoothing.get_process_noise_covariance(v: src.trajectory_smoothing.VehicleSystem)[source]¶

src.trajectory_smoothing.get_transform_control_noise_to_state_space(v: src.trajectory_smoothing.VehicleSystem, u: list, x: list)[source]¶

To derive the covariance of the additional motion noise, we transform the covariance of noise in the control space by a linear approximation to the derivative of the motion function with respect to the motion parameters Inputs:

• ν is the vehicle concrete type

• u is the control [a,ω]ᵀ

• x is the state estimate [x,y,θ,v]ᵀ

src.trajectory_smoothing.observe(v: src.trajectory_smoothing.VehicleSystem, x: list)[source]¶

Vehicle observation, returns a saturated observation Inputs:

• ν is the vehicle concrete type

• x is the state estimate [x,y,θ,v]ᵀ

src.trajectory_smoothing.step(v: src.trajectory_smoothing.VehicleSystem, x: list, u: list)[source]¶

Vehicle dynamics, return the new state Inputs:

• ν is the vehicle concrete type

• x is the state estimate [x,y,θ,v]ᵀ

• u is the control [a,ω]ᵀ

Module contents¶