“Incremental End-to-End Learning for Lateral Control in Autonomous Driving” has been accepted for IEEE Access. The paper can be found at https://ieeexplore.ieee.org/document/9737528.
Category: News
“Markerless Vision-based Gait Analyses” paper accepted
“Ridon Vehicle” paper accepted
SAE MobilityForward Challenge: AI Mini-Challenge
SAE MobilityForward Challenge: AI Mini-Challenge sponsored by Ford Motor Company was held on October 25-26, 2021.
Congrats to Assistant Professor Jaerock Kwon and students Huda Hussaini, Mintu Mariya Joy, Richa Chachra and Shouryan Nikam — members of a UM-Dearborn team that racked up several third place finishes in a recent SAE Mobility Forward AI Mini Challenge. The competition challenges universities to apply novel data science techniques to emerging technology issues relevant to the mobility industry. The UM-Dearborn team’s project identified socio-economic disparities in transportation for Pennsylvania residents and outlined options for innovative services that could make transportation access more equitable.
UM-Dearborn’s Mauto team won 3rd place in all three categories. I am very honored to be a faculty advisor of the team.
Awards
- Solutions Stakeholders Report – 3rd place – $100 – University of Michigan – Dearborn
- Stakeholder Presentation – 3rd place – $100 – University of Michigan – Dearborn
- Showcase Booth – 3rd place – $100 – University of Michigan – Dearborn
Team members
- Members: Huda Hussaini, Mintu Mariya Joy, Richa Chachra, Shouryan Nikam
- Mentors: Jaerock Kwon, Songan Zhang (Ford), Eric Tseng (Ford)
Job Posting
Prof. Kwon is looking for an MS student for conducting research on human motion inference. Read the following page for more details.
Send an email to jrkwon@umich.edu stating your research interest with your CV. Knowledge of machine learning and data analysis is required.
- Application deadline: Oct 31, 2021
- Appointment type: hourly based for 2 to 3 months
New Bimians
We have new lab members. Welcome aboard!
Andrew Soong, M.S. student in Mechanical Engineering, Ann Arbor Aws Khalil, Ph.D. student in Electrical and Computer Engineering Jesudara Omidokun, M.S. student in Electrical and Computer Engineering
Joint Conference on AI in Smart Cities
On April 7, the Joint Conference on AI in Smart Cities hosted by KSCEE, KOTAA, and KOCSEA was successfully held in virtual.
- KSCEE: the Korea-American Society of Civil and Environmental Engineers (KSCEE)
- KOTAA: Korean Transportation Association in America (KOTAA)
- KOCSEA: the Korean Computer Scientists and Engineers Association of America
As conference co-char and president of KOCSEA, I am so happy that there was no technical difficulties during the conference. I also belive that parcipants like the conference.
Job Posting for 2021 SURE
Objective:
The Bio-Inspired Machine Intelligence (BIMI) lab (https://bimi.jrkwon.com) is looking for one undergraduate student who is interested in research funded by the 2021 Summer Undergraduate Research Experience (SURE) program.
Research Summary:
Driver Style Transfer for Autonomous Driving
People drive a vehicle differently. Yet, self-driving has a pre-defined style designed by manufacturers. It can be very uncomfortable for a driver who is aggressive in driving when a self-driving vehicle drives very cautiously. This can also be true vise versa.
To provide an individualized autonomous driving experience, we will collect several driving styles from human drivers in a simulated environment. Then we will extract features from the driving data that can represent driving styles.
To validate our approach, we will create a driving model from human driving data and the model will be deployed to a controller system of a simulated autonomous vehicle.
Through this research, we will be able to provide an individualized autonomous driving experience for a driver who can feel the autonomous vehicle drives just like how the driver usually drives.
Responsibilities:
- Read papers to understand the problems.
- Participate in data collection, processing, and management.
- Implement a neural network using Keras and/or PyTorch.
- Prepare technical report.
Qualifications:
- Required experiences:
- Python Programming skill
- Literacy in Linux commands
- Machine learning toolkit: Keras and/or PyTorch
- Preferred
- Git and GitHub(or GitLab)
- Robotic Operating Systems (ROS)
- scikit-learn
- pandas
Program Requirements:
- The student must be a registered UM-Dearborn undergraduate student with an expected graduation date no sooner than December 2021.
- The student will receive a stipend of $3,200 to be paid in installments throughout the project period (May-August). Students will receive payments of $1,000 the first week of June, July, and August, with a final payment of $200 after the Showcase in September.
- Note, if the student receives financial aid, they should check with the Financial Aid Office to determine if this stipend will impact their financial aid package.
- The student should not receive duplicate support while participating in this program; supplementing with external funding during the duration of the SURE program is NOT ALLOWED.
- Student participants and faculty mentors must attend the SURE Kickoff on Tuesday, May 18th at 4 pm.
- The student is required to attend the SURE Professional Development Seminars taking place on the following dates: 6/3, 6/17, 7/8, 7/22, 8/5, and 8/19 (all sessions run from 1:00 pm-2:00 pm; a schedule will be provided at the Kickoff event).
- The student is expected to present their work at the SURE Project Showcase in the fall.
Contact Information
Please contact Prof. Jaerock Kwon at jrkwon@umich.edu.
Data Viewer for OSCAR
I have just completed implementing a data viewer for the OSCAR (http://github.com/jrkwon/oscar).
It can show the labeled steering angles and corresponding predictions to visually perform a neural network’s inference.
KSEA-UKC 2020
Houcem presented his work in a technical symposium during UKC 2020.
