Consumer demand, government legislation, and the automotive industry’s efforts to increase safety are driving the adoption of advanced automotive safety systems that use radar, LiDAR, camera systems, GPS, and vehicle to vehicle (V2V) communications. For such systems to be feasible, they require security to prevent hacking, and they must be competitively priced, reliable, easily repaired, and have a communication system that can transmit and receive all of these signals. This requires a single or standardized multi-mode, multi-application transceiver that can interface with multiple radar and LiDAR systems that are installed around the vehicle. This project aims to design a single transceiver with a software defined radio (SDR) that handles radar, LiDAR, GPS, and V2V communications. REU students will use two portable computers and two USRP Ettus Research B200mini SDR transceivers that interface with Matlab through the MathWorks’ SDR Hardware Support Package to transmit and receive various automotive communication signals in real time. Currently, digital short-range communications (DSRC), and Cellular V2X (C-V2X), are the two technologies that compete to fulfill the communication needs of connected vehicles. Working with a single programmable transceiver provides the ability to design a communication system that can be used by existing safety systems while allowing for reprogramming to accommodate future standards. Because these V2X communication transceivers need to be flexible and multi-modal, software defined radios are a perfect platform for developing such transceivers.
A Cognitive Security Framework for ICPS in Smart Cities
With the proliferation of ICPS (IoT-based Cyber Physical Systems) in private, commercial and government sectors, many IoT networks of different type, size, and sensitivity level exist. The main objective of this project is to introduce an adaptive framework for secure design, implementation, and evaluation of ICPS networks in order to increase security by closing the gap between the existing security mechanisms (e.g. defense in depth and multi-factor authentication), current decentralized IoT security solutions (e.g. ARM trust shell), and the security needs of the IoT-based CPS. This project will use a proactive method to stay ahead of attacks by constantly monitoring network data patterns and identifying threats that are imminent. Deep Learning (DL) methods will be employed that rely on local proximity of patterns to find and construct prediction models of interest through supervised learning. Using a large training set, models of malicious behavior can be constructed and used to detect and block potential threats. The project tasks will include: (1) designing a modular framework for a secure implementation of ICPS, covering deterrent, preventive, detective, corrective, and recovery controls; (2) developing prediction models and DL techniques to classify malicious behaviors and target agents using ICPS log data as the input; and (3) conducting penetration testing by simulating the proposed framework on GENI and Cloudlab test-beds. Several smart city sensors and nodes will be prototyped and attacks and mitigation techniques on smart nodes will be developed as a testing framework. Smart nodes will be built using Raspberry pi and open source testing packages.
Enabling non-V2X devices for V2X World
Vehicle-to-everything (V2X) is a technology that allows vehicles to communicate with road-side infrastructures and other non-mobile traffic systems. Many of the legacy and current road-side infrastructures and other non-mobile traffic systems do not support the V2X communication and protocols. This project will explore creating a hardware-software wrapper for such devices.
Requirement: Advanced Embedded Systems.