Francesco Borrelli

Open (1) Energy Efficient Controls for Connected Automated Vehicles (CAVs)

Open. Apprentices needed for the fall semester. Enter your application on the web beginning August 18th. The deadline to apply is Monday, August 30th at 9 AM.

In this project, we aim to improve the energy performance of connected automated vehicles in real-world scenarios. The energy savings can be obtained by harnessing technologies such as (i) remote computations, (ii) forecasts, (iii) historical data, (iv) automation, and (v) coordination with other vehicles and infrastructure. We have the actual vehicles that we can control longitudinal and latitudinal motions and also technologies that we can utilize vehicle to vehicle (V2V) and vehicle to infrastructure (V2I) communications.

Building a Carla virtual environment: We are currently building a virtual environment that we can perform preliminary experiments before the actual real-world experiments. The task will include setting up Carla simulations and mapping virtual environments for both (1) Energy Efficient Controls for Connected Automated Vehicles and (2) Autonomous Rover for the Solar Field projects. Learning outcomes from this task can be (i) experiences of Carla which is a famous virtual environment software for autonomous vehicle testings and (ii) vehicle dynamics modeling with control strategies (PID and MPC).

Day-to-day supervisor for this project: Eric Choi, Graduate Student

Qualifications: Python is mandatory and ROS background preferably

Weekly Hours: 9-11 hrs

Related website: https://sites.google.com/berkeley.edu/mpcconnectedcars/home

Open (2) Autonomous Rover for a Solar Field

Open. Apprentices needed for the fall semester. Enter your application on the web beginning August 18th. The deadline to apply is Monday, August 30th at 9 AM.

The overall goal of the project is to build and develop autonomous ground rovers for inspection, delivery, and maintenance in the solar panel field. Full-stack technology of autonomous driving will be developed, including low-level steering and traction control, high-level planning, decision making, perception, and control.

Full-stack technology for autonomous driving will be involved, including:
(1) Sensor fusion and state estimation
(2) LiDAR and camera-based perception and detection
(3) Low level motion control
(4) Route planning and trajectory generation

Day-to-day supervisor for this project: Thomas Fork, Graduate Student

Qualifications: Software development, Python Prior experience in robotics / autonomous driving Basic knowledge in at least one of these fields: sensor data processing, computer vision, control theory, feedback control, optimization

Weekly Hours: 9-11 hrs

Open (3) Intelligent modelling and control of vehicles in parking lots

Open. Apprentices needed for the fall semester. Enter your application on the web beginning August 18th. The deadline to apply is Monday, August 30th at 9 AM.

In the project, we aim to bring self-driving technology to tightly-constrained, mixed-autonomy environment such as parking lot. We have collected and annotated human drivers' behavior data in the parking lot with drone. With this data, we are trying to construct the simulated environment, extract the behavior model, and simulate with our intelligent control algorithm.

The task will include:
1) Setting up 2D simulation environment;
2) Model the human drivers' behavior using rule-based controller or data-driven model.
3) Simulate the vehicle motion and generate live parking lot traffic scenarios.

The learning outcome will be:
1) Experience with self-driving dataset and industrial standard;
2) Experience with Python, Robot Operating System (ROS);
3) Experience with control strategy design (PID, LQR, MPC) and deep learning in control (CNN, RL etc)

Day-to-day supervisor for this project: Xu Shen, Graduate Student

Qualifications: Mandatory: Python, Machine Learning Preferred: ROS, System Control, Robotics

Weekly Hours: 9-11 hrs

Related website: http://www.mpc.berkeley.edu/

Open (4) Autonomous Racing with the Berkeley Autonomous Race Car

Open. Apprentices needed for the fall semester. Enter your application on the web beginning August 18th. The deadline to apply is Monday, August 30th at 9 AM.

This project involves work on the perception, prediction, and control stacks of the 1/10th scale Berkeley Autonomous Race Car (BARC) platform. The goal is to perform multi-agent racing on an indoor track with onboard sensing and computation. Students can expect to learn about advanced modeling, planning, and control techniques in a high-performance racing setting.

Students will develop the autonomy stack of and design experiments for an RC race car based platform. Specifically, we are looking for students who are experienced and interested in the areas of perception and state estimation, modeling and prediction of adversarial agents, data-driven system identification of dynamical systems, and learning-based control.

Day-to-day supervisor for this project: Edward Zhu, Graduate Student

Qualifications: Students must be proficient in Python and/or C++ and have prior experience with the Robot Operating System (ROS). We don't require that you be an expert in any area, but students must demonstrate a working knowledge of concepts in their area of interest.

Weekly Hours: 9-11 hrs

Related website: http://www.barc-project.com/