RWR4013

Digital Twins for Smart Cities

Build a practical city digital twin: computer vision → simulation → optimization → decision

Computer Vision Traffic Simulation Optimization KPIs (LOS / Delay / Emissions) Planning Decisions

Instructor: Dr. Ahmad Mohammadi · RoadwayVR

Preview this course

Preview

Course workflow preview: Data → Model → Connect → Decide

Course Overview

How the course is structured and what it prepares you to do.

Course overview

This 12-week course introduces digital twins for smart cities through a practical, project-based workflow aligned with four common layers/phases of a city digital twin: (1) Data Layer (Computer Vision) to collect field traffic data from video (detection → tracking → counts/speeds), (2) Model Layer (Simulation) to build a digital network and traffic model in SUMO (network → demand → signals), (3) Connection Layer (Optimization) to evaluate system performance using KPIs (delay, emissions, level of service) and test scenarios (e.g., signal timings), and (4) Service Layer (Decision) to select the most effective scenario and communicate findings through a written report, a presentation, and a project showcase (demo video + website).

Teaching Philosophy (UDL)

  • Multiple means of representation: Concepts are delivered through lecture slides, hands-on slides, video tutorials, and self-paced website resources.
  • Multiple means of engagement: Two support channels are provided (Office Hours + Community Learning Platform) to reduce barriers and keep students supported.
  • Multiple means of expression: Students demonstrate learning through participation, transportation news brief, weekly deliverables, assignments, a paper-based midterm with written reasoning, and a design project with professional deliverables (report + presentation + showcase).

Course Outline

What you’ll learn and how you’ll be evaluated.

Course Learning Outcomes

  • Develop foundational knowledge in digital twins for smart cities (computer vision, simulation, optimization, and visualization for planning)
  • Collect and process real-world traffic data from video using object detection and tracking to produce counts and speed estimates
  • Build digital road networks and traffic demand models using open-source tools (SUMO, QGIS) and standard modeling practices
  • Test and compare transportation scenarios (e.g., signal timing alternatives) and evaluate performance using KPIs (delay, emissions, level of service)
  • Diagnose physical–digital mismatches, propose calibration actions, and justify validation decisions using evidence
  • Communicate findings through professional reports, visualizations, presentations, and a project showcase (demo video + website)

Assessment

  • Class participation (weekly) (5%)
  • In-class deliverables (weekly) (15%)
  • Transportation News Brief presentation (once per student) (10%)
  • Assignments (throughout term) (10%)
  • Midterm examination (paper-based) (Week 8) (25%)
  • Design project (cumulative) (35%)

Topics

This 12-week course is organized around weekly lecture concepts and hands-on labs. Each week includes lecture slides, hands-on slides, and video tutorials.

Week Topic (Lecture + Hands-on) Materials
1
Introduction to Digital Twins for Smart Cities
Lecture: What is a Smart City? What is a Digital Twin? The 4 Stages of a Digital Twin; Course syllabus overview; Course learning outcomes
Hands-on: Fundamentals of computer vision; What is an image? Pixel data; Resolution; What are videos? Convolutional Neural Networks (CNNs); Convolution fundamentals
2
Computer Vision I (Object Detection)
Lecture: Neural networks; CNNs; stages of a CNN; object detection workflow (study area & recording, frame extraction, dataset creation, annotation & classes)
Hands-on: Dataset partitioning (train/val/test); YOLO Trainer; understanding results
3
Computer Vision II (Object Tracking)
Lecture: Detection vs. tracking; bird’s-eye view (homography); multi-object tracking (ByteTrack); Supervision (Roboflow); SimJamComputerVision; SimJamCV workflow
Hands-on: Real-world data collection to support planning; example observed counts & turning movements; detection and tracking using SimJamCV
4
Introduction to Traffic Simulation
Lecture: What is traffic simulation? Purpose; examples; SUMO; network development; vehicle characteristics; dynamics; car-following and lane-changing models
Hands-on: Install SUMO; set environment variables; install Notepad++; SUMO files & UI; create a simple network; add opposite flow; intersections (unsignalized + signalized)
5
Digital Road Network Modelling with GIS
Lecture: Spatial data; GIS and the seven steps of GIS; GIS as a foundation for simulation; GIS software types; data layers for simulation; network elements required
Hands-on: Download/install QGIS; map services; imagery & georeferencing; import a GIS map into simulation
6
Digital Road Network Development
Lecture: Edge vs. junction; junction types; import a GIS map into NetEdit; create a road network on top of a GIS map
Hands-on: Lane connections; right of way; traffic signals; traffic flows (demand)
7
Simulation Calibration
Lecture: Accurate network development; accurate signal timing; movement calibration; volume calibration; speed calibration
Hands-on: Traffic movement/volume/speed calibration
8
Midterm (Paper-Based): Physical vs. Digital World Reasoning
In-class assessment (no computer): A paper-based exam including fundamental concepts and printed figures/tables from computer vision (detection/tracking results) and simulation (volumes/speeds, SUMO snapshots, mismatch plots). Students will:
  • Answer conceptual questions on digital twins and smart cities concepts
  • Answer conceptual questions on real-world data collection using object detection and tracking
  • Answer short questions on traffic simulation and GIS-based network modelling
  • Interpret outputs and identify likely sources of mismatch (computer vision vs. simulation)
  • Propose specific calibration actions (what to change and why)
N/A (paper-based)
9
Optimization
Lecture: Intersection control types; traffic signal technology; signal planning; phasing; signal control types; case studies
Hands-on: Intersection functional/physical areas; KPIs; delay → LOS; optimization
10
Decision I (Communication & Presentation)
Lecture: Select best strategy (KPI interpretation; signal timing interpretation); create a professional presentation (visualization + preparation)
Hands-on: Video demonstration
11
Decision II (Write a Formal Technical Report)
Lecture: Report structure; study goal/objectives; formatting tips; writing style & clarity; referencing
Hands-on: Final project website + demo video + screenshots + KPI tables/plots; final project presentations
12
Presentation Week
Lecture: Student presentations
Hands-on: Student presentations

Prerequisites

No prior computer vision or traffic simulation experience required.

  • Basic understanding of transportation concepts (traffic flow, intersections, signals)
  • Comfort with software installation and learning new tools
  • Willingness to engage with technical material

Recommended background: Urban planning, transportation planning, or related field. Students from other disciplines (policy, design, data science) are welcome with instructor permission.

Student Support & Accessibility

Support channels designed to reduce barriers and help students succeed.

  • Office Hours: By appointment via email (flexible scheduling; virtual meetings available).
  • Discord Community: Continuous Q&A support (target ~24-hour response time), peer learning, archived resources, troubleshooting tips.
  • Accessibility: Students requiring accommodations should contact their Accessibility Services office as early as possible.
  • Technical Support: Required software (SUMO, QGIS) is free and open-source; installation tutorials and troubleshooting guides provided.

Office Hours

Appointment by email or join the community Discord group for ongoing Q&A and peer support.

Prefer Discord for quick questions, study groups, and announcements.

Tip: include your full name, course code (RWR4013), and a 1–2 sentence summary of your question.

Optional Course Materials

Recommended references for computer vision, traffic operations, and digital twins.

  1. University of California, Berkeley. Course Materials: CS180/280A: Intro to computer vision and computational photography (Fall 2025). https://cal-cs180.github.io/fa25/
  2. Wunderlich, K., Vasudevan, M., & Wang, P. (2019). Traffic analysis toolbox (FHWA-HOP-18-036). FHWA. https://ops.fhwa.dot.gov/publications/fhwahop18036/fhwahop18036.pdf
  3. Tao, F., Zhang, M., & Nee, A. Y. C. (2019). Digital twin driven smart manufacturing. Academic Press.

Course Website

Access the latest course page, updates, and public materials here:

https://rwr40.github.io/DigitalTwinsforSmartCities/
Open Course Website

Course Description (PDF)

Download the official course description document.

RWR-4013-Digital-Twins-for-Smart-Cities.pdf
Download PDF

Resources

Software

  • Computer vision tools (object detection + tracking workflows)
  • Traffic simulation tools (SUMO)
  • GIS tools for network building (QGIS / OSM)