Class simulation_framework::utilities::vehicle::controller::VehicleControllerFactory

Definition: autonomy/simulation/utilities/vehicle_controller/vehicle_controller_factory.h (line 17)

Inherits from:

• I_VehicleControllerFactory

Members

• Create

Public functions

Function Create

std::unique_ptr< IVehicleController > simulation_framework::utilities::vehicle::controller::VehicleControllerFactory::Create(const VehicleControllerConfig &cfg, double dt) override

Create a configured vehicle controller instance.

Parameters:

• cfg: should contain path to vehicle JSON configuration or structured field should have VehicleProperties type for structured kind

Parameters:

• dt: nominal controller delta time

Returns:

unique_ptr owning an IVehicleController

This factory reads controller tuning, model parameters from the vehicle configuration and constructs a VehicleController pre-populated with those settings. See also: I_VehicleControllerFactory for VehicleControllerConfig definition

Controller overview (concise):

• The controller minimizes a weighted nonlinear least squares cost between predicted motion state outputs and reference targets.

• Cost: C(x,u) = 0.5 * sum_i W_i (y_i(x,u) - y_i_ref)^2, where W_i = Q_i / (S_i * S_i). The controller tracks x, y, psi, v.

• Tuning groups in the config:

• "tracking" (q_lon, q_lat, q_yaw, q_vel): higher values increase tracking accuracy.

• "sensitivity" (s_*): scaling values; smaller means more sensitive.

• "restrictions" (r_*): restricts input rates (jerk, steering rate).

• Optimization controls:

• "prediction_horizon_steps": how many future steps are optimized (>=2 recommended).

• "max_iterations": solver iteration cap.

Sample vehicle_config (full, for reference): { "vehicle_name": "test_vehicle", "single_track": { "lf": 1.234, "lr": 1.234 }, "limits": { "max_acceleration": 8, "max_steering_angle": 0.654, "min_acceleration": -8, "min_steering_angle": -0.654 }, "controller_tuning": { "restrictions": { "r_jerk": 0.1, "r_steering": 2.0 }, "sensitivity": { "s_jerk": 0.1, "s_lat": 0.1, "s_lon": 0.1, "s_steering": 0.1, "s_vel": 0.1, "s_yaw": 0.1 }, "tracking": { "q_lat": 80, "q_lon": 100, "q_vel": 2000, "q_yaw": 2000 }, "optimisation": { "max_iterations": 50, "prediction_horizon_steps": 5 } } }

Note: Fields are read to populate VehicleProperties; ensure the configuration file contains the sections shown above. "dt" in the API is delta time (seconds) between controller steps.

Parameters:

• const VehicleControllerConfig & cfg
• double dt

Return type: std::unique_ptr< IVehicleController >