Past Events

PhD Defense

The integration of heterogeneous systems for vehicular sensing

 

Download as iCal file

Tuesday, March 23, 2021, 10:00am - 12:00pm

 

Speaker: Xiaoyang Xie

Location : Remote via Zoom

Committee

Prof. Desheng Zhang (Advisor)

Prof. Zheng Zhang

Prof. Yongfeng Zhang

Prof. Ruilin Liu (external member)

Event Type: PhD Defense

Abstract: With the emergence of connected vehicles and autonomous vehicles in the future, vehicles are being equipped with rich sensing and communication devices, as well as various traffic sensors are being deployed on the road. Based on real-time sensor data from these sensors, tremendous driving assistant applications have been presented to improve urban efficiency, e.g., navigation, HD-live map, and traffic accident recovery, etc. Thus, it is essential to improve vehicular sensing for urban residents. The vehicular sensing could be achieved by modeling mobility patterns of vehicular data from multiple systems, e.g., (1) mobile systems including devices equipped in taxis, buses, trucks, personal vehicles, etc.; and (2) stationary systems including traffic cameras, electronic toll station systems, roadside units, etc. However, most existing works study vehicular sensing on a single system, which may not comprehensively represent the mobility patterns of residents for the whole city, leading to a biased understanding. In this dissertation, we study the integration of multiple vehicular sensing systems and design a vehicular sensing framework, which enables integrated vehicular sensing from three perspectives, i.e.: (1) Integration of mobile systems: we present a real-time sensing task scheduling system, RISC, which schedules sensing tasks for commercial vehicles with constraint sensing resources based on unified modeling for vehicles' mobility patterns; (2) Integration of stationary systems: we design a type-aware travel speed prediction for road segments method, mDrive, by utilizing sparse spatial-temporal data of multiple types of vehicles from traffic camera systems; (3) Integration of hybrid systems: we provide a privacy risk prediction approach based on transfer learning, TransRisk, to predict the privacy risk for a traffic camera system through its small-scale short-term data, and the knowledge learned from large-scale long-term data from existing sensing systems. We implement these three modules of our vehicular sensing framework by real-world data from two Chinese cities, Shenzhen and Suzhou. Our results provide insights for the integration of multiple sensing systems on the improvement of the performance of vehicular sensing from different aspects.

 

Join Zoom Meeting
https://rutgers.zoom.us/j/3062568848?pwd=Q3BKTzQ0NFdTai8rUmVCTW9HaGxHQT09

Join by SIP
This email address is being protected from spambots. You need JavaScript enabled to view it.

Meeting ID: 306 256 8848
Password: 165212
One tap mobile
+13126266799,,3062568848# US (Chicago)
+16465588656,,3062568848# US (New York)

Join By Phone
+1 312 626 6799 US (Chicago)
+1 646 558 8656 US (New York)
+1 301 715 8592 US (Washington DC)
+1 346 248 7799 US (Houston)
+1 669 900 9128 US (San Jose)
+1 253 215 8782 US (Tacoma)
Meeting ID: 306 256 8848