#450 Digital Twin for Driving

Principal Investigator
Erick Guerra
Start Date
July 1, 2023
End Date
June 30, 2024
Project Type
Research Applied
Grant Program
US DOT BIL, Safety21, 2023 - 2028 (4811)
Grant Cycle
Safety21 : 23-24


Digital tools for mapping are now an integral tool for navigation. Waze, Google maps have leveraged Satellite and Google Street imagery which allow us to have an accurate representation of any set of GPS coordinates. This technology evolved just a couple of months ago with the release in May 2023 of dynamic models of the Earth. Companies like Cesium, a Google backed up startup based in Philadelphia is offering Unity and Unreal models of the Earth, with an Application Programming Interface (API) that can be leveraged for any dynamic Geospatial application. While a number of drone applications have been developed, no solid test has been developed for a driving application.
We propose with this project to test the Unity and Unreal API of the Earth to assess the drivability of these Virtual Reality models. We will partner with Jitsik LLC, a small startup in Mixed Reality who has been paving the way in this direction. Jitsik recently established a proof of concept for driving, by integrating two scenarios taken from Google Earth. The first scenario is a mountainous road just outside Denver, Colorado. The second scenario is a downtown neighborhood of Denver, Colorado. These proofs of concept were quite successful in the sense that they were easy to deploy, with very little coding. On the downside, the simulated Denver downtown neighborhood offered a rough terrain. It is indeed the result of imperfections from the satellite imagery. 
We propose with this project to explore the drivability of this Unity/Unreal Earth API. We propose to target the Roosevelt Blvd in Philadelphia, which has already been the subject of academic research (Erick Guerra). The Roosevelt Blvd lies in an urban area, but often feels like a highway. We propose to leverage the Digital Twin of the Roosevelt Blvd and develop filtering techniques to remove existing traffic, smooth the road. Once the environment has been cleaned up, we will add Virtual Traffic lights and stops to make the scenario ready for a Human Subject Research Project. We anticipate some techniques developed through this research project will become part of a toolkit that can be applied to any neighborhood in the world to create a Digital Twin of that neighborhood.
This project is a collaboration between the University of Pennsylvania Engineering Department and Jitsik LLC, a startup in Mixed Reality, led by Dr. Helen Loeb. Helen has been working in driving simulation for over a decade.


Strategic Description / RD&T
This project is a Research Project at its Core. Should we be successful in establishing the value of the Earth 3D database for driving, we anticipate additional education and translational research projects will follow. 

This project addresses the Safety goal of the US DOT Research, Development and Technology Strategic Plan (listed page 14) by targeting a critical artery in Philadelphia. In the previous 5 years, the Boulevard had accounted for 14% of the City of Philadelphia’s total traffic fatalities. The techniques developed in this project will translate into a toolkit that can be reused to create a digital twin of any other road in the US.

This project addresses the Transformation goal of the US DOT Research, Development and Technology Strategic Plan (listed page 50). It is indeed Data Driven as it leverages advanced data collection and data processing capabilities to create accurate, credible, and accessible information for decision making on infrastructure. This project leverages satellite and drone imagery of the existing US roads and aims at refining this model to prepare it for commercial use in safety and education.

This project has a strong component in Technology Transfer and Deployment  as it precisely assesses the suitability of GeoSpatial computational models made available in Spring 2023 for practical applications. These applications will range from University studies, to planning by local authorities to education of young novice drivers. The models developed will also have a value for Urban Planning and Design.

Deployment Plan
Detailed plan for deployment
Our project benefits from the Proof of Concept developed by Jitsik in Spring 2023. Several youtube videos are available online. The results from this proof of concept showed that neighborhoods that are void of trees or high building present a much better mathematical resolution for the satellite imagery. A quick test of the outskirts of Denver showed that this neighborhood would require little smoothing. Other neighborhoods such as downton Denver would require more work.
Our focus is the Roosevelt Blvd in Philadelphia. While it is an urban area, there are no problematic high buildings. The area is fairly open. While many trees border the road, we anticipate they will not create major problem to the research project. We anticipate for this project.

Quarter 1: Establish working collaboration with Cesium, including the development of a Cesium EcoSystem grant (Cesium Ecosystem Grant Application – Cesium). Partner with the Computer Graphics Department of the University of Philadelphia to recruit students with a graphic computational background. Explore the Roosevelt Blvd through Google3D Earth and Cesium Technology. Physically drive the Roosevelt Blvd to anticipate issues in the Digital Twin Deployment Process.

Quarter 2: Develop Filtering techniques to create a smooth surface for driving. Edit all existing vehicles and obstructions out of the environment/

Quarter 3: Render Traffic lights and stop signs into the simulation.

Quarter 4: Use Jitsik Technology to validate the new development. We will use the MetaQuest Pro.
Expected Outcomes/Impacts
Anticipated Outcomes/Impacts
This project is very ambitious as it aims to make satellite imagery of the earth drivable for urban planners, for driving safety and for education. Drivers may want to experiment driving on the left in the UK before renting a vehicle. Drivers may want to visualize a merge of highways before actually being there. Google released 3D visual directions for driving through its API. This project goes one step further by making these models drivable. 
Expected Outputs
Anticipated Outputs
We anticipate the development of a toolkit that can be reused for any neighborhood in the US, so that drivers can experience a simulated experience in the environment they live. The project will improve safety by paving the ground to Human Subject Research so that dangerous roadways can get better designed.
A search of “driving”, “earth”, “satellite” in the TRID database led to 7 references. Four references are anterior to 2005 when satellite positioning was limited by the Department of Defense GPS Selective Availability.  These references are not relevant to our project. For the three references.
In 2018, Gao et al published “GPS Modeling for Vehicle Intelligent Driving Simulation”. This article establishes a new GPS algorithm based on the functional characteristics of GPS. It does not address driving applications and is irrelevant to our application.
In 2020, Wang et al published “Development of Unmanned Roller and Its Application in Highway Engineering”. This article targets the development of a Highly Automated Vehicle for construction sites. Its use of GPS targets self-driving. It does not address the smoothness of the terrain, our research focus.
In 2021, Goswami et al published “Intersection of Automotive and Satellite Connectivity: Use Cases and Exploration of a Hybrid Model”. This paper explores a potential intersection of automotive and satellite connectivity. It is focused on connectivity and communication and does not address the quality of 3D mapping for driving applications. Our work is unique and pushes the limit of what can be accomplished with satellite imagery driving.

Individuals Involved

Email Name Affiliation Role Position
erickg@upenn.edu Guerra, Erick University of Pennsylvania PI Faculty - Tenured
LoebH@email.chop.edu Loeb, Helen Jitsik LLC Co-PI Other


Amount of UTC Funds Awarded
Total Project Budget (from all funding sources)


Type Name Uploaded
Data Management Plan Data_management_plan_-_DigitalTwinDriving.pdf Nov. 9, 2023, 10:40 a.m.
Publication Loeb_SAE_2024_Cesium_paper.pdf March 27, 2024, 7:36 a.m.
Progress Report 450_Progress_Report_2024-03-31 March 27, 2024, 11:19 a.m.

Match Sources

No match sources!


Name Type
JITSIK LLC Deployment & Equity Partner Deployment & Equity Partner
Delaware Valley Regional Planning Commission Deployment & Equity Partner Deployment & Equity Partner
PA Safe Roads PAC Equity Partner Equity Partner