Supported by NSF CNS #1816938 and #1814923
Description
More than half of fatal vehicle crashes in the US today are "run-off-road" (RoR) crashes and roadside barriers are the last means of mitigating their severity. Yet, vehicles of tomorrow are slated to operate on roadside infrastructure designed decades ago. This project aims to bridge this gap by establishing wireless connectivity between vehicles and roadside barriers by developing a novel vehicle to barrier (V2B) communication and networking paradigm called connected barriers (CBs). CBs will (1) complement on-board sensor technologies and existing physical barriers, (2) avoid RoR crashes, (3) minimize the severity of a crash when it is inevitable, and (4) help develop mutual collaborations between the roadside safety and vehicular communication and networking communities, which will lead to robust technology solutions. Based on Department of Transportation's value of a statistical life estimate in 2014, the societal cost of a crash is $9.4 million per fatality, with approximately 10,000 fatal ROR crashes each year. With widespread implementation, this research project could reduce fatal and non-fatal crash rates by up to 70%, which could save as many as 8,000 lives per year in the U.S. alone and billions of dollars in economic losses. This project supports 2 graduate students and its results will be incorporated into several courses at UNL and OSU. Insights from the project results will help establish collaborative efforts to educate both industry and academia in Nebraska and Ohio.
To realize the CB vision, the project explores three main threads: (1) the foundations of V2B communication will be established through analysis of wireless channel characteristics in field measurements, actual car-crash tests, and channel model development. (2) to enable reliable control reactions, very high-fidelity barrier-assisted vehicle localization and vehicle-barrier synchronization solutions will be developed. (3) novel communication algorithms and protocols will be developed to disseminate periodic information about barrier and road conditions and to facilitate real-time information exchange during upcoming RoR events, with an eye towards coexistence with existing communication standards. The developed solutions will be evaluated through encroachment and vehicle crash experiments in the Midwest Roadside Safety Facility, and system evaluations on low-volume suburban and rural road segments in collaboration with the Village of Eagle, Nebraska, and Nebraska's Cass County. Finally, a multi-time-scale simulation platform will be integrated by including empirical measurements of the wireless channel, vehicle and barrier dynamics, traffic dynamics, and communication and network dynamics to provide effective evaluation tools of the CB paradigm
Goals
The major goals of the project are to:
(1) form the foundations of V2B communication through the analysis of wireless channel characteristics through field measurements, actual car-crash tests, and channel model development.
(2) To enable reliable control reactions, an advanced vehicle to barrier integration plan will be enacted. With this plan, very high fidelity barrier-assisted vehicle localization and vehicle-barrier synchronization will be achieved.
(3) Novel communication algorithms and protocols will be developed to disseminate periodic information about barrier and road conditions, and to facilitate real-time information exchange during upcoming RoR events, with an eye towards coexistence with existing communication standards. The proposed solutions will be evaluated through encroachment and real-life vehicle crash experiments. The fundamental results emerging from this high-risk, high-reward research will bring together researchers from roadside safety, wireless communications, and networking.
Personnel
PI: Mehmet C. Vuran (UNL)
PI: Eylem Ekici (OSU)
Co PI: Ronald Faller (UNL)
Co PI: Cody Stolle(UNL)
Students: Mohammad Mosiur Rahman Lunar (UNL),
Ceyhun D. Ozkaptan (OSU)
Outcomes
2019-2020
Major Activities 2019-2020
Indoor and Outdoor Experiments: We have conducted a set of indoor and outdoor experiments with both single Tx-Rx systems and MIMO Tx-RX systems. In addition, a vehicle-to-infrastructure(V2I) experiment is organized on a road in the City of Lincoln, Nebraska.
Data Analysis of Crash Experiments: We analyze three channel metrics: path loss, RMS delay spread, and RMS Doppler spread for developing a V2B channel model. For this analysis, two types of experiments are conducted: static open-space tests and V2B crash tests. Based on these experiments empirical values of path loss, RMS delay spread, and RMS Doppler spread are obtained. Instantaneous velocity and displacement of vehicle antenna with respect to the barrier antennas are measured with high-precision sensors for ensuring the granularity of those quantities. Then the empirical values are investigated with mathematical models to explain the impact of barrier antenna height deployments, vehicle types, barrier types, etc.
Beamforming System Design: We have finished the beamforming system design for V2B communication. Antennas are placed in half-wavelength separation for the 5.8GHz band. Moreover, proper enclosure and cabling for installation in a crash vehicle are finalized. OFDM-based Digital Beamformer Development: We develop an OFDM-based digital beamformer application that is compatible with both GNU Radio and UHD API. The time overhead of each component of this development is also analyzed.
Multi-Range OFDM-based Radar Communication System: We investigated the use of pilot-based OFDM radar approaches for joint communication and multi-range radar systems. The system is capable of switching between short range, long range, and medium range radar options while minimizing the age of information of the covered regions, while sustaining communication sessions.
Joint Waveform Design for Radar Communication System: We investigated the joint waveform
design problem for automotive radars enabling communication. This approach leverages the vast bandwidth
available in the automotive radar band (76-81GHz) for sensing as well as for data transmission. Based on our
previous efforts for OFDM pilot-based radar and communication system, this activity focused on the development
of optimal joint waveform design problem.
Significant Results 2019-2020
- A robust path loss model is designed for the V2B channel. This path loss model can address the impact of directivity angle (for directional antennas) and vehicle body impairment in path loss.
- A lognormal RMS delay spread model and a lognormal RMS Doppler spread model is also derived. These models depict the impact of vehicle shape, barrier type, and antenna deployment height in empirically determined RMS delay spread and RMS Doppler spread. Necessary correction procedures of channel state information (CSI) calculation is also analyzed.
- The RMS error of the path loss model with empirical values varies between 4.82dB to 12.65dB. The standard deviation of RMS delay spread varies between 4.03dB to 7.92dB. For RMS Doppler spread the standard deviation is in the range of 0.93dB to 1.12dB.
- By dynamically adjusting the pilot subcarrier locations in the pilot-based OFDM waveform, we achieved a multi-range radar and communication system. The proposed system is first of its kind both for stand-alone radar and communication systems. We formulated a cost minimization problem to minimize a balanced cost of the communication delay and the age of information of the radar operation modes. In particular, for each radar operation mode, we define the age of information as the time difference between the last sensor data update time of this radar operation mode and the current time. We then pose the multi-range radar and communication scheduling as a cost minimization problem with the cost comprising the age of information of each radar operation mode and the communication traffic load. The performance of the proposed greedy algorithm has been compared against the benchmark periodic policy through simulations. The results have shown that the heuristic algorithm has superior performance in terms of data transmission rates as well as the freshness of the sensing information.
- We studied an adaptive OFDM waveform design problem for joint automotive radar and communication systems with given statistics about the extended target and signal-dependent clutter. First, we investigated the power and subcarriers allocation between data and pilot symbols in the OFDM waveform to minimize the channel estimation error and formulate the effective channel capacity. Then, we presented the design problem to maximize SCNR while maintaining baseline effective communication capacity compared to equal power allocated waveform. We showed that the problem is a non-convex QCQFP and propose relaxation and approximation approaches for the problem. The numerical results show that the proposed CA approach solves the problem with good optimal value and low complexity compared to the well-known SDR approach.
Broader Impact Outcomes 2019-2020
- Our team participated in the Raikes School Research Fair organized by UNL. In this fair, the key equipment of V2B research was showcased with crash-test videos for the audiences. The aim was to motivate participants of that fair who were mainly computer science undergrad students in these types of research.
- A detailed overview and principal operation schemes were presented to a staffer from Senator Fischer’s office. The overall impact of this research in Nebraska and all over the U.S. was also explained to her.
- A brief overview of all our researches including V2B was presented at a research program organized by UNL. Senator Fischer and UNL Chancellor Green visited to see that presentation.
2018-2019
Major Activities 2018-2019
Crash Experiment and Data Analysis: We have instrumented more than 30 full-scale vehicle-to-barrier crash tests conducted at Midwest Roadside Safety Facilities (MwRSF) and multiple component tests to investigate signal power, directionality, and continuity. Until now, we have conducted 24 different real-world crash experiments with the collaboration of MwRSF. In each experiment, a vehicle encroaches towards a barrier and eventually crashes on it. For those experiments, we install different antennas with USRP(s) at both vehicles and barriers in different heights. We analyze the data collected from those experiments and observe performance matrices such as received signal strength (RSS), signal to interference and noise ratio (SINR), error vector magnitude (EVM), phase error (PE), bit error rate (BER), and channel coherence time. Based on these experiments optimum position and types of vehicle and barrier antennas are investigated.
OFDM Pilot-Based Radar and Communication System: We developed a joint communication and radar for automotive systems. The hardware system utilizes the 76-77GHz and 77-81GHz bands reserved for automotive radar to exchange data between vehicles in addition to performing radar functions.
The OFDM-based joint radar and communication system has been configured to meet the classification performance levels for Short Range and Long Range sensing. The system has been shown to deliver between 446-1440Mbps data rate based on different deployment scenarios without sacrificing radar performance. This is the first radar-communication system developed for automotive radar systems that can achieve these data rates.
Testbed Development
Resources
Pictures from outreach events.