Controller and Vehicle Model

Vehicle Model:

• Bicycle model with states [X,Y,psi,vx,vy,omega]
• Nonlinear tire model (pacejka-like)
• Acceleration dependent normal forces
• Inputs: Steering angle, rotational speed of front and rear axle

Controller A:

• Exact Input/Output Linearization
• Output: Deviation of front decoupling point at J/(l_R*m) from the reference trajectory
• Relative degree: 2,2

Controller B:

• Exact Input/Output Linearization
• Output: Deviation of rear decoupling point at -J/(l_F*m) from the reference trajectory
• Relative degree: 3,3

Definition of Test Cases

• Test 01 - Nominal
• Test 02 - Initial Condition Response: Y(0) = -0.2m, psi(0)=-2°
• Test 03 - Low Friction: mu0=0.6
• Test 04 - Low, Unkown Friction: mu0=0.6
• Test 05 - Mismatched Parameters: m, J, L increased by 30%

Szenario I - Lane Change

Szenario II - Double Lane Change

Downloads

IV_2013_.m_-_002_-_06.12.2012.zip (Source Code)

Manual_Comparison_of_Trajectory_Tracking_Controllers_for_Emergency_Situations.pdf

In order to construct maneuver automata for autonomous driving, we provide a set of motion primitives in an XML format consisting of

• id
• initial_curvature
• initial_velocity
• orientation
• successors (id of possible next motion primitives)
• x-coordinates of the trajectory
• y-coordinates of the trajectory

Download: zip file

We provide SpaceEx models of an electro-mechanical brake with conformance monitor. More information of these models can be obtained from the publication "G Frehse, A Hamann, S Quinton, M Woehrle: Formal analysis of timing effects on closed-loop properties of control software. In: Real-Time Systems Symposium (RTSS), 2014 IEEE, 53-62". Link: pdf file.

Download: zip file

This data set is obtained from recordings of the twizy vehicle. We provide accelerations (lateral, longitudinal), speed, yaw, as well as control commands leading to these values. The sources for the recordings are the vehicle's CAN bus, DGPS and IMU.

Download: zip-file

This data has been used to perform trace conformance as presented in this deliverable. The data has been obtained from the multi-body Dynacar simulator.

Download: zip-file

This data set has been used to validate the controller of the Twizy vehicle at Tecnalia using the Dynacar simulator.

Download: zip-file

We provide an energy management benchmark problem for a smart grid where electrical energy is supplied to a load via local power production from a solar PhotoVoltaic (PV) installation. The smart grid is connected with the main grid, which can eventually provide the energy needed for balancing demand and generation. The goal is to set the battery energy flow so as to keep the energy exchange with the main grid as close as possible to a nominal profile, within certified bounds, avoiding the fluctuations caused by the local PV energy production. Some energy production profiles of the PV
installation and environmental data on irradiation and temperature are available for the design of the energy management strategy, together with a hybrid model for the battery and the electrical load profile. We describe a data-driven solution, pointing out its limits and providing some hint on possible direction for improvement.

Download: zip-file