Handover/Takeover

Installation

Run the following command:

pip install numpy pandas matplotlib scikit-learn casadi

Python Simulations

python plot.py

Files:

1. lane_change.py and double_lane_change.py are the main files performing lane change and double lane change scenario in a dummy python environment
2. plot.py contains the plotting code and integrating the above two files. There are different arguments like lane change, double lane change that can be set. For more information, see the argparse parameters.

CARLA simulations

source your carla simulator using:

bash ./CarlaUE4.sh

open another terminal:

cd carla_simulator
python blend_control.py

Files:

1. autonomous_control.py contains the source code for controlling the carla vehicle using MPC controller
2. mpc.py contains the MPC controller code
3. manual_steer.py contains the source code for controlling the carla vechile manually
4. wheel_config.ini contains the logitech steering wheel mapping information
5. blend_control.py contains the main code that's required for collecting and storing the data along with the visualizations. There are different arguments with proper description added in the file. For more information, see the argparse parameters.

---

TO DOs:

1. in blend_control.py ensure that the reference path is getting displayed in the carla simulator properly
2. record the data using CARLA simulations and store into the desired format. desired recorded behaviour is in recordings folder. have to stabilize the human control a bit more.
3. take data of 15-20 licensed drivers for stronger impact:

• 3 runs for simulator familarity using manual_control.py
• 3 runs for desired double lane change behaviour in blend_control.py
• 5 runs for actual data recording step using blend_control.py

5. compare different transition metrics - check plot.py to see how to compare different transitions and improve analyze.py
6. start writing paper with the results