Connected and Autonomous Vehicles

CCI faculty have been involved with some of the earliest work on vehicular communications and sensing, as well as distributed control algorithms for intelligent traffic signal control. We also have significant expertise in multi-agent autonomous vehicles.

Vehicular Communications (V2X)

CCI faculty have been pursuing research on topics related to vehicular communications:

  • Radio propagation studies for vehicle to vehicle and vehicle to infrastructure settings
  • Protocols for intermittently connected wireless vehicular networks
  • Hybrid Cellular and DSRC/WAVE-based communications for low cost information dissemination of over-the-air updates
  • V2V approaches for disseminating road condition information
  • Coding for low latency vehicular download
  • Rapid vehicular file transfer protocol

Faculty Involved: Andreas F. Molisch, Bhaskar Krishnamachari, Konstantinos Psounis

Illustrative results from Prof. Andy Molisch’s field measurements of doppler spectrum and angular power spectrum for vehicular wireless communication channels.

Vehicular, Edge and Cloud Computing

As the sensing and communication capabilities for vehicles grow, so do their computational requirements. CCI researchers have been exploring various approaches to distribute computation across heteroegeneous elements: on car processors, mobile devices, and edge and cloud server. Some of this research has been supported by and performed in collaboration with General Motors.

Faculty Involved: Bhaskar Krishnamachari, Ramesh Govindan, Murali Annavaram

Intelligent Traffic Control

CCI Researchers have explored:

  • Application of Queuing theory to design more efficient traffic controllers.
  • Decentralized green-time allocation at signalized traffic intersections.
  • Use of real-time transportation data to improve navigation and route planning.
  • Game theoretic approaches to traffic control.
  • Deep Learning applied to traffic flow and impact analysis

Faculty Involved: Ketan Savla, Cyrus Shahabi, Yan Liu

Machine Learning

CCI faculty have expertise and active research  on various machine learning tools and techniques of relevance to autonomous vehicles:

  • Spatio-temporal learning and prediction
  • Deep learning applied to Image recognition and traffic analysis
  • Transfer learning
  • Online reinforcement learning

Faculty Involved: Yan Liu, Fei Sha, Bhaskar Krishnamachari

Autonomous Vehicles

CCI researchers include experts on autonomous robotics, both single and multi-agent systems. Their research into autonomous vehicles spans unmanned ground, air and water vehicles.

Faculty Involved: Gaurav Sukhatme, Nora Ayanian

Vehicular Sensing

We are interested in sensors deployed/mounted on vehicles, to detect other vehicles and pedestrians as well as air quality and other remote-sensing applications. The figure shows a air quality monitoring pilot deployed on a USC shuttle bus in collaboration with Valarm, an LA-area IoT startup. Research in this area spans hardware development, and algorithms and software for efficient large-scale data collection and processing.

Faculty Involved: Hossein Hashemi, Mahta Moghaddam, Bhaskar Krishnamachari, Antonio Ortega