Webinar Presentation

HTML version of the presentation
Image descriptions are contained in brackets. [ ]

Smart Sensors and Infrastructures for Transportation
(July 29, 2021)

Improving Pedestrian Safety at Intersections Through Advanced LIDAR Sensing Technologies
Presenter: Farzana Chowdhury
Presenter’s Org: University of Texas, Arlington

T3 webinars are brought to you by the Intelligent Transportation Systems (ITS) Professional Capacity Building (PCB) Program of the U.S. Department of Transportation (USDOT)’s ITS Joint Program Office (JPO). References in this webinar to any specific commercial products, processes, or services, or the use of any trade, firm, or corporation name is for the information and convenience of the public, and does not constitute endorsement, recommendation, or favoring by the USDOT.

Slide 1: Improving pedestrian safety at intersections through Advanced LIDAR sensing Technologies

Farzana Chowdhury, Ph.D. student, farzana.chowdhury@mavs.uta.edu

Advisor: Dr. Taylor Li, P.Eng. email: Taylor.Li@uta.edu https://actionlab.uta.edu/

[This slide contains two images: (1) the logo of the University of Texas at Arlington’s Department of Civil Engineering and (2) the logo of the ACTION Laboratory.]

Slide 2: Outline

Categories of LIDAR sensors
Three levels of LIDAR sensor algorithms
A demonstration of pedestrian data collection
Integration with automated traffic signal performance metrics (ATSPM)

Slide 3: What is LIDAR?

LIDAR is a method for measuring distances (ranging) by illuminating the target with laser light and measuring the time the reflection of the light takes to return to the sensor
- Mechanical LIDAR sensors: cove 360 degrees
  - E.g., Velodyne LIDAR (16 lines ~128 lines)
- Quasi-solid-state LiDAR sensors: directional, no rotating parts
  - SCHOTT/Cepton/LSIS/HESAI, etc.
  - Very active market to meet the enormous demand for AVs

[This slide contains three images: (1) a graphic of a car with arrows pointed to a bicycle, a cow, and a truck representing LIDAR technology, (2) a diagram of the LIDAR motor, with parts labeled, and (3) a illustration of the LIDAR sensors.]

Slide 4: Three levels of algorithms for LIDAR sensors

Hardware algorithms:
- To make the LIDAR sensors efficient and reliable to generate raw point clouds. It is the concern of LIDAR sensor manufacturers.

Perception and classification algorithms:
- Cluster the point clouds into objects
- Identify the objects characteristics (types, behaviors, etc.)
- OEM or third-party

Integration algorithms:
- Domain-specific applications
- For instance, we integrate the LIDAR tracking algorithm with real-time traffic signal status at intersections

[This slide contains three images: (1) a building using LIDAR sensors, (2) a road using LIDAR sensors, and (3) real-time traffic signal status at intersections with the label, “The focus of this presentation.”]

Slide 5: New trend of LIDAR sensors and algorithms

More new LIDAR manufacturers are established
- But still very expensive due to the high demand of AV industry

The target markets are more diverse
- AV
- Intelligent intersections
- Toxic gas monitoring

LIDAR manufacturers begin to open their perception software to new applications
- Open an avenue for multimodal smart-city applications

Slide 6: Compare with computer-vision solutions

The technical ceilings are the same:
- Clustering algorithm, machine learning algorithms, etc.

LIDAR is commonly considered more robust
- In extreme weathers (foggy, storm, snowstorm, etc.)

LIDAR does not have concerns of privacy
- Privacy concern is being raised at the national level.
  - E.g., Bill S.1214 - Privacy Bill of Rights Act (116^th congress)

Slide 7: Demonstration: Pedestrian mobility analysis at intersections

Pedestrian information is minimal at intersections
- Through push button, we know some people wishes to cross
  - But we do not know how many people wish to cross and how long they have waited
  - Excessively long waiting time will make pedestrians lose respect to traffic light.
    - It was observed a lot during the experiment (e.g., Jaywalking)

With a LIDAR system with perception algorithms, we track the pedestrian behaviors at intersections to understand
- Waiting time before crossing (ped-delay)
- Perception-reaction time to the onset of WALK
- Crossing speed distribution

Slide 8: Locations: City of Irving, TX

[This slide contains three images: (1) a bird’s eye view photo of Irving, TX with four dots in a diamond shape, with lines connecting them, (2) two people on the side of the road looking at computer screens, and (3) people on the side of the road looking at computer screens with a bucket truck extended to the top of a telephone pole.]

Slide 9: System Architecture

[This slide contains one diagram of a system architecture from the UTA program 1 to UTA program 2, UTA-in-motion platform, then sensors and a graphic user interface.]

Slide 10: Pedestrian tracking process

[This slide contains one photo of a street from the bird’s eye view showing the pedestrian tracking process with three arrows pointing to “Ped Waiting Zone,” “Entering Zone (during Walk),” “Cross Ending Zone.”]

Slide 11: Live Demo

[This slide contains four images: (1) a screenshot of a live demo, (2) a screenshot of photos of the road, (3) a screenshot of photos of the road, and (4) a tracking database and code from the database.]

Slide 12: Integration with Automated Traffic Signal Performance Monitoring System (ATSPM)

Slide 13: A live system at Cooper @UTA Blvd, Arlington, TX

[This slide contains two images: (1) a bird’s eye view photo of where the system films in Arlington, TX, (2) a data graph from the live system at Cooper at UTA Blvd in Arlington, TX.]

Slide 14: Thanks!

Acknowledgement: partially funded by National Institute for Transportation and Communities (NITC), a national UTC center

For more details, please contact: Dr. Taylor Li, email: Taylor.Li@uta.edu. Homepage: https://actionlab.uta.edu

↑ Return to top