Project: #77 SmartShuttle: Model Based Design and Evaluation of Automated On-Demand Shuttles for Solving the First-Mile and Last-Mile Problem in a Smart City Progress Report - Reporting Period Ending: Sept. 30, 2018 Principal Investigator: Umit Ozguner Status: Completed Start Date: June 19, 2017 End Date: Sept. 30, 2018 Research Type: Applied Grant Type: Research Grant Program: FAST Act - Mobility National (2016 - 2022) Grant Cycle: 2017 Mobility21 UTC Progress Report (Last Updated: Sept. 30, 2018, 8:24 p.m.) % Project Completed to Date: 90 % Grant Award Expended: 93 % Match Expended & Document: 99 USDOT Requirements Accomplishments Year 1 Accomplishments (100% complete): - The unified, scalable and replicable architecture that is being used was prepared. - The hardware-in-the-loop automated driving simulator was prepared/updated for use in this project. - Extensive model-in-the-loop and hardware-in-the-loop (HiL) simulations were conducted. - Two Denso DSRC modems were added to the HiL simulator and used to test communication with other vehicles and instrumented traffic lights. Year 2 Accomplishments (90% complete): - Results (architecture) from first year was applied to our newly donated Dash EV neighborhood electric vehicle, converting it to a research vehicle for automated driving. - Our path following and collision avoidance algorithms were scaled and replicated from our Ford Fusion Hybrid research automated driving vehicle to our Dash EV vehicle. - We used our Ford Fusion Hybrid automated vehicle to collect perception sensor data from the OSU campus and are preparing a labeled dataset. We will collect similar data from the Easton Town Center outdoor shopping area. - We identified a relatively large and mostly empty parking lot for our proof-of-concept demo. We have also used our parking lot and the lawn area behind our garage for some of the scalability and replicability tests. We did some proof-of-concept demos in our parking lot. We will do one more proof-of-concept demo to finish the project. - We re-created the OSU AV test pilot route in our HiL simulator with other vehicles and created a soft version of our automated driving vehicle that shuttles from our lab to our main center along that pilot AV route. This has demonstrated the effectiveness of our HiL simulator environment for testing automated driving. Impacts The SmartShuttle project impact is on demonstrating a scalable and replicable low speed autonomous shuttle solution for smart cities, especially for the smaller companies that dominate that area. The resulting impact is expected to be more widespread use of autonomous shuttles in smart cities and more mobility choices especially for the first-mile and last-mile problem. The project work until now has already had an impact. Prof. Levent Guvenc and Prof. Bilin Aksun Guvenc were members of the Autonomous Electric Vehicles working group of Smart Columbus and informed group members in Columbus and in OSU of the potential benefits and the accompanying problems of autonomous shuttles used in geo-fenced areas as first and last mile solutions. The project results were helpful in making Smart Columbus and OSU leaders understand more about the capabilities of existing autonomous shuttles and make more informed decisions. We shared project results and the resulting expertise with the OSMI (Ohio State Mobility Initiatives) group where Prof. Levent Guvenc was a member. This, among other developments, led recently to Ohio Department of Transportation's DriveOhio smart mobility program which is one of the results of the continuation of the OSMI group effort. After the DriveOhio organization, the State of Ohio started to legally allow autonomous driving on public roads using a governor's executive order. OSU, Smart Columbus, DriveOhio and Jobs Ohio joined hands and have decided to start autonomous shuttle driving in Columbus. The vendor was chosen recently and autonomous shuttle rides will start in 2018 in a geo-fenced area in downtown Columbus. The PIs Guvenc and Aksun-Guvenc informed the OSU Mobility and Smart Columbus management teams on potentials and challenges of current autonomous shuttles using the SmartShuttle project. As a result, the Columbus downtown deployment will be assisted autonomous driving with someone at the driver seat all the time and intervening during critical situations. All of the Smart Columbus project management team visited our Automated Driving Lab during Spring 2018 for a two hour meeting. The project results were shown to them then along with limitations and the state-of-the-art of Av shuttles. The HiL simulator environment for CAV testing in the Automated Driving Lab was seen to be a very useful resource and Smart Columbus wants to use this in evaluating the autonomous shuttle deployment in the downtown area. Several conference presentations were made in the SAE World Congress 2018. Presentations in IEEE ITSC 2018 and IEEE CDC 2018 will take place in November and December. Two SAE World Congress 2019 papers are under preparation with abstracts accepted. One journal paper in the SAE International Journal of Connected and Connected Vehicles has been accepted for publication. Two journal papers are being prepared and will be submitted by December 2018. Other Accepted Conference Publications Acknowledging Project Support (not yet presented): S. Zhu, S.Y. Gelbal, B. Aksun-Guvenc, 2018, "Online Quintic Path Planning of Minimum Curvature Variation with Application in Collision Avoidance," IEEE Intelligent Transportation Systems Conference, Hawaii, accepted. SAE World Congress 2019 Publications Under Preparation (Abstracts Accepted): Gelbal, S.Y., Zhu, S., Li, Xinchen, Cantas, M.R., Aksun-Guvenc, B., Guvenc, L., 2019, “A Unified Approach for Autonomous Vehicle in a Smart City,” WCX19: SAE World Congress Experience, April 9-11, Detroit, Michigan, Session AE 504 Intelligent Transportation Systems, SAE offer/paper number 19AE-0184, abstract accepted. Wang, H., Guvenc, L., 2019, “Discrete-Time Robust PD controlled system with DOB/CDOB compensation for high speed autonomous vehicle path following,” WCX19: SAE World Congress Experience, April 9-11, Detroit, Michigan, Session AE 501 Intelligent Vehicle Initiative, SAE offer/paper number 19AE-0134, abstract accepted. Website pages with stories related to SmartShuttle: Road to smart mobility means merging minds, Chronicle of Higher Education https://www.chronicle.com/paid-article/Road-to-smart-mobility-means/93 Columbus: Driverless Destination https://mae.osu.edu/news/2017/10/columbus-driverless-destination Driving Toward an Automated Future, Together https://car.osu.edu/news/2017/09/driving-toward-automated-future-together Avoiding pedestrian collisions in a future of automated cars https://car.osu.edu/news/2017/09/avoiding-pedestrian-collisions-future-automated-cars YouTube Videos: DASH Low Speed Autonomous Shuttle https://youtu.be/K9dCd4ofYxA Realistic Simulation of Autonomous Shuttle on OSU AV Pilot Test Route https://youtu.be/_yWiZWP0Rag Automatically Labeled Camera Data on OSU AV Pilot Test Route https://youtu.be/E63lKSIFmQI Outcomes New Partners There were no new partners. Innova donated a new Dash EV neighborhood electric vehicle. OSU TTM was able to obtain a state license plate for this vehicle and we are allowed to use it in public roads. We moved the drive-by-wire, sensors and computers from the old Dash vehicle to this new one during year two of the project. Issues There are no technical issues related to this project. We have finished the majority of the project tasks on time. However, we still want to conduct a final proof-of-concept demo, finalize the project paper submissions and use the small budget left on the project. We therefore want a no-cost extension until Dec. 31, 2018. This will enable us to fully use our second year budget and finalize our proof-of-concept demo and final papers being submitted to the SAE World Congress and journal papers under preparation. The proof-of-concept demo will use map building and map matching (instead of our previous SLAM approach) and deep learning for object detection and tracking. There were coordination issues between OSU and CMU due to late signing of sub-contracts that resulted in late approval and delayed transfer of OSU funding. First year project funding was available as year 2 of the project was starting. There were similar delays in the transfer of year 2 funding. We worked hard to make sure that this issue did not affect the technical performance of the project.