Prospects and Challenges for Automated and Autonomous Vehicles

by Tracy Zafian, Research Fellow

avcrash
May 2018: A car driven with Tesla’s Autopilot driver assist system crashed into a parked police car. The Autopilot system is designed to be used on limited access highways, not on roadways such as this one. (Source: Uncredited/AP/Rex/Shutterstock)

In March 2018, an Uber Volvo operating in automated driving mode, with a driver at the wheel, hit and killed a pedestrian in Arizona. It was the first autonomous vehicle (AV)-related pedestrian fatality. There have been other crashes involving autonomous test vehicles and additional fatalities involving lower-level automated vehicles since.  Also in March, a driver in California was killed when his Tesla, in autopilot mode, crashed into a concrete highway lane divider and caught fire. Last month, a Tesla, in autopilot mode, crashed into a parked firetruck in Utah and into a parked police car in California.  Injuries were minor in these instances.

All of these crashes have occurred while there has been a driver at the wheel making decisions about when to use and disengage the automated driving assist system. As reported in a USA Today article, following the Utah firetruck crash, Tesla issued a statement saying, “When using Autopilot, drivers are continuously reminded of their responsibility to keep their hands on the wheel and maintain control of the vehicle at all time,” and, “Autopilot is designed for use on highways that have a center divider and clear lane markings.” These conditions are not always met when crashes occur; for example, the driver in the Utah crash admitted to being distracted by their phone before the crash. From the National Transportation Safety Board’s preliminary findings from investigating the Uber pedestrian crash, there was no warning given to the safety driver before the crash. Current AV technologies, such as Tesla’s Autopilot, are referred to by car manufacturers as “driver assistance systems” but it is not clear that all drivers understand their limitations, including the need for drivers to monitor the driving environment and stay involved in the driving process.

The Society of Automotive Engineers has developed a classification system for autonomous vehicles. The classification includes six levels (Level 0-5); with Level 5 being fully autonomous and Level 1, containing some automated features such as adaptive cruise control and parking assist. Most current automated driver assistance systems are Level 1 or 2, meaning that drivers still need to be actively involved.

Driver assistance systems and autonomous vehicles hold great promise for improving safety and mobility, but AV technologies are still relatively new, and numerous challenges remain. A number of universities and researchers in Massachusetts are exploring this topic. In April, a commentary by MIT AgeLab researchers, “People must retain control of autonomous vehicles,” was published in Nature magazine (link is for the article). In their remarks, Dr. Ashley Nune, Dr. Bryan Reimer, and Dr. Joseph Coughlin, Age Lab Director and UMass Transportation Center Research (UMTC) Affiliate, focused on two areas – safety and liability – that need urgent attention as policies and regulations are developed for autonomous and semi-autonomous vehicles. They write that, in their view, “some form of human intervention will always be required. Driverless cars should be treated much like aircraft, in which the involvement of people is required despite such systems being highly automated. Current testing of autonomous vehicles abides by this principle. Safety drivers are present, even though developers and regulators talk of full automation.” The researchers’ piece ends with key points for policymakers preparing AV legislation to consider:

  • Driverless does not, and should not, mean without a human operator
  • More information should be shared with operators/drivers about how well different autonomous and driver assist systems are working, including their reliability and limitations
  • Operators should need to demonstrate that they understand the autonomous and driver assist systems in their vehicles and should be tested on their understanding and competence at periodic intervals
  • Remote monitoring networks should be established and shift time guidelines considered for workers monitoring AVs.

In May, a forum held at Harvard University’s T.H. Chan School of Public Health on “Self-Driving Cars: Pros and Cons for the Public’s Health.” (A recording of this session and a transcript are available at this link.) Dr. Jay Winsten, Associate Dean for Health Communication at Harvard, said there is hope right now around the potential for autonomous and highly automated vehicles to reduce traffic deaths.  He also addressed the hype around this: “I think both the media and some of the manufacturers and developers have been going a little too far in setting public expectations for what to expect, especially in the short-term and in the medium-term.” The panelists discussed that initially most vehicles will be highly automated (SAE Level 2), not autonomous (Level 3-5), and the deployment of the autonomous vehicles is likely to occur first for long-distance, highway-based commercial transport and in urban areas for shuttles and other short-distance trips. There are some challenges including the current reliability and drivers’ understanding of AV technologies, including the need for drivers to stay alert while behind the wheel and the safety of vulnerable road users. There are also concerns regarding regulation. The federal government through the National Highway Safety Administration has developed some guidelines regarding autonomous vehicles and automated driving systems. However, there are currently no federal regulations in place regarding autonomous vehicles. Therefore, currently regulations are primarily set at the state level.

As described in an earlier Innovative Outlook article, Governor Baker and Massachusetts state officials have largely taken the approach that it is better not to regulate AVs through legislation, as the technologies are still evolving and legislation can be difficult to modify once passed. Panelist Deborah Hersman of the National Safety Council, shared those concerns, saying “We’ve got to find out how to do this differently” so that any regulations keep up with changing technology. Herman also urged there be more transparency and data sharing regarding specific AV technologies and how well they perform, saying that NTSB investigations after a crash can be challenged by lack of access to such data.

In June 2018, MassDOT entered into a Memorandum of Understanding with several municipalities to help facilitate and expand autonomous vehicle testing on roadways in Massachusetts. As described in a MassDOT blog article, “Following the signing of this MOU, MassDOT and the participating communities will finalize a universal application for companies to use when seeking to test autonomous vehicles and the participating municipalities will identify locations and roadways suitable for autonomous vehicle testing. ‘This agreement will allow companies to responsibly develop and test autonomous vehicle technology in Massachusetts, while ensuring there are uniform safety guidelines in place,’ said Governor Baker [at the MOU signing] . ‘The MOU builds on the existing autonomous vehicle testing framework while simplifying the process for municipalities to work with innovative companies that are seeking to advance transportation, create jobs in our nation leading innovation economy, and improve our quality of life in the Commonwealth.’  …Said Lieutenant Governor Karyn Polito, ‘By creating a standardized process and working collectively with local officials, we can generate economic growth and support our communities as they play a role in the future of innovation and motor vehicle automation.’ Fourteen communities signed the MOU initially, including Boston, Worcester, Arlington, Boston, Braintree, Brookline, Cambridge, Chelse, Medford, Melrose, Newton, Revere, Somerville, Weymouth, Winthrop, and Worcester.  In addition, the Massachusetts Department of Conservation and Recreation also joined the MOU, allowing Commonwealth-owned parkways to be available for autonomous vehicle testing.

 

 

Portland Maine – Thinking Ahead for Autonomous Shuttles

by Melissa Paciulli, Manager of Research

 

shuttle

 

The Town Planner and Legislators are working together to be prepared for a potential autonomous shuttle that connects the Portland Transportation Center with the waterfront and downtown area to assist tourists. New legislation was drafted by Representative Heather Sanborn, which would allow cities and towns to start pilot programs in partnerships with state agencies with autonomous vehicles, as reported in the Portland Press.

The proposed legislation could set up Maine to be a leader in pilot programing for autonomous transit.  Companies such as the global data company Inrix, have been in touch with Town officials about collecting data on the city streets, necessary for autonomous navigation. There are no current companies lobbying for the first pilot in the area, however the legislation, which is slated for a January 2018 review, is a first step in the process of making this a reality.

New Year, New UMass Human Performance Lab Web Site

By Tracy Zafian, UMTC Research Fellow

HPL
Photo Source: Shannon Roberts, Human Performance Lab, UMass Amherst

The Human Performance Lab (HPL) based at the University of Massachusetts-Amherst has a brand new look! The HPL was originally created by Professor Donald Fisher in the 1990s and is world-renowned for its work on teen driver training. In 2016, the University of Massachusetts Transportation Center (UMTC) Affiliate Researcher Professor Shannon Roberts joined the HPL and now serves as the HPL co-director, overseeing research activities of the lab and day-to-day operations. Professor Roberts’ research is focused on driver feedback, in-vehicle interface design, automated vehicles, and teen/novice drivers. Her research group’s web site has information on her team and other interests.

With Dr. Robert’s arrival, the lab underwent significant changes. Upgrades include a new vehicle (2015 Ford Fusion), five new projectors with an expanded field of view of 330 degrees, new channels for displaying the side and rear-view mirrors and a new in-vehicle display.  The HPL has also obtained other new equipment including Virtual Reality (VR) headsets for delivering training and using with simulations, and a new heart-rate monitor to use with participants in lab studies.  Coming soon will be an instrumented vehicle for use with on-road studies.  The upgraded equipment will significantly expand the lab’s research capabilities.  One recent new area of research for the lab involves autonomous vehicles.  This is the transfer of driving control from driver to vehicle, and drivers’ awareness of their surroundings and ability to respond to potential roadway hazards as they switch from autonomous modes that require more attention and input from the driver, to those that require less attention.

As it has since its beginning, the lab, based in the Department of Mechanical and Industrial Engineering, continues to collaborate with other departments at UMass-Amherst including Civil Engineering, the UMTC, Computer Science, Electric and Computer Engineering, and Psychology.

Do You Trust Vehicles To Do The Driving?

by Courtney Murtagh and Melissa Paciulli, Manager of Research

While Boston is positioned to become a hub for autonomous vehicles, buy-in from the public, remains a potential hurdle.   New research based out of the Massachusetts Institute of Technology (MIT), indicated that people aren’t so sure about self-driving cars.  The study surveyed 3,000 Bostonians of different ages. Half of those interviewed, said they would never buy a self-driving car due to safety, and citing that they did not trust technology.

Researchers at UMass Amherst are studying driver trust to determine how technology impacts human behavior.  Foroogh Hajiseyedjavadi, a PhD Candidate in the Transportation Engineering program at the University of Massachusetts Amherst, was interviewed this past week about her research.

“Inappropriate level of trust in the technology, whether it is over trust or under-trust, would negatively affect the benefits of that technology,” said Hajiseyedjavadi.

Hajiseyedjavadi is evaluating drivers’ trust in automated vehicles and believes that understanding peoples trust in automated vehicles will help enhance human automation interaction models.

“Automated driving is expected to enhance traffic safety and flow,” said Hajiseyedjavadi in her executive summary. “The system will not be as effective if users do not accept it or do not utilize it appropriately.”

Hajiseyedjavadi’s research is in two phases; the first phase being a 68 question survey distributed to participants online with questions that include general demographics, their driving history, and questions about their psychological and personal traits. This phase also addressed a persons’ previous experiences with automated systems, like ATM’s and vending machines, and their experience with computers.“The hope is that the results of this part will give some basic understanding of the level of trust consumers have even before experiencing the technology,” said Hajiseyedjavadi.

The second phase of the research was conducted in the Arbella Human Performance Lab at UMass Amherst, on a driving simulator which included simulation of autonomous controls within the vehicle. Hajiseyedjavadi and her team recruited 80 people to participate in the study and programmed different driving scenarios.

During the simulated drives, there were different levels of automation and functionality that the driver would experience. Either the driver would get an autonomous driving system that worked fine, or the driver would get an autonomous vehicle with a 12 percent or even a 25 percent failure rate.

“We have never scripted a crash. The 12 and 25 percent failure comes when the autonomous vehicle sensors fail. This occurs at either a pedestrian crossing or an intersection. For these failures, it is up to the driver to correct the mistake,” said Hajiseyedjavadi.

Sensors are set up during the simulation that measure the vehicles speed, lane keeping, acceleration, and deceleration. Another set of data is the physiological data of the subject. The driver wears a heart rate and variability sensor. There are also two sets of video cameras showing the hand movements on the steering wheel and the foot petals. This is to see when drivers are engaging in the system.

After the simulations the team would ask the participants to complete a second questionnaire about their level of trust while driving and their mental workload.

“All this data combined is hopefully going to give us a better understanding of the level or trust and how people are interacting with the vehicle,” said Hajiseyedjavadi.

“Trust is one of those things that control people’s use of the system,” continued Hajiseyedjavadi citing why this study is so important.

She said in order to improve drivers’ trust in automated vehicles, scientist need to improve the level of reliability of the system.

Communicating and teaching drivers how to use the technology is also essential, according to Hajiseyedjavadi.

“We are definitely going to see automated driving cars in the future,” said Hajiseyedjavadi. “I don’t know exactly how many years but we are definitely going to have them soon.”

 

Snow, Sensors and AVs – A Wintery Mix

by Tracy Zafian, Research Fellow

The headline of a Canadian news article last year, warned that “We may never see self-driving cars anywhere it snows.”  That article discussed the experiences of Sam Abuelsamid, a former automotive engineer turned tech writer and his experience with cars’ automated assist features in snowy conditions. One snowy day, as Abuelsamid drove around, his car’s radar sensor became covered with slush. As a result the car’s adaptive cruise control (ACC) disengaged, and an alert was given that the radar sensor needed to be cleaned off before the ACC could be used again. Abuelsamid surmised that to use the ACC in these weather conditions, he would need to pull over and clean off the radar sensor repeatedly as he kept driving. He chose to keep the adaptive cruise control disengaged instead. In another example, Abuelsamid was in a parking lot and there were large fluffy snowflakes falling. The car’s parking assist sensors detected the snowflakes and thought they were potential obstacles, thereby triggering repeated audible alerts to warn of their presence.

Can autonomous vehicles be taught to be smarter than this, and will they ever be able to perform well in the snow? Some car manufacturers including Ford and General Motors have been testing self-driving cars in snowy conditions and the results are promising.

One issue in the winter, as mentioned above, is that the sensors and cameras for vehicle safety features can be covered with snow or ice. This can happen even with new non-autonomous vehicles and as described in this winter’s safety guide, it’s important to keep them clean. Some car manufacturers are using small wipers or defrost technology to keep the sensors and cameras clear.

Ford, similar to other companies, has been developing high fidelity 3D maps of the roads its self-driving cars will travel. The maps include road geometries and line markings, road signs, and other nearby features. Ford’s self-driving cars have numerous sensors and cameras, including a LIDAR (LIght Detection And Ranging) sensor on the top of the car which provides a 360-degree view, radar, and cameras on the front, back and sides of the vehicle. With these detailed maps, even if some data aren’t available at a particular time, a Ford self-driving car could still have enough information to know its location, the location of other road users, and potential hazards. This demo video shows a self-driving car in snowy conditions at Ford’s full-scale outdoor test facility in Michigan.

Researchers at the Massachusetts Institute of Technology (MIT) Lincoln Laboratory have been creating maps of roadway sub-surfaces, for use by self-driving cars. This mapping involves the use of Localizing Ground-Penetrating Radar (LGPR). To generate the maps, the LGPR equipment is mounted on the undersize of a vehicle to collect data during an initial drive. On future drives, the LGPR data is compared to the base map generated early to determine an autonomous vehicle’s location. One advantage of the LGPR approach to localization in that it doesn’t rely at all on optical images of the roadway or surrounding environment which could be obscured in certain weather conditions. This video describes MIT’s LGPR technique in more detail.

winteravIn testing with autonomous vehicles in snowy conditions, both Ford’s and MIT’s approaches have been shown to allow autonomous cars to achieve locational accuracy within a few centimeters. One possible limitation of these methods, at least at the writing of this article, , is that they both depend on detailed mapping of the roadway environment and for MIT, the subterranean environment- in clear weather conditions, in advance. This means that autonomous vehicles using such mapping can only “drive” on roadways which have already been mapped to the detailed level needed.

Bridgewater State University becomes a “LivingLab”

By Uma Shama and Lawrence Harman, Bridgewater State University

This summer, Bridgewater State University (BSU) has become a “LivingLab” with the BSU GeoGraphics Lab using campus parking lots to research Small Unmanned Aircraft System Remote Sensing Platforms for transportation.

Lab co-directors and UMTC Research Affiliates Dr. Uma Shama and Mr. Lawrence Harman are using aircraft systems produced by the company DJI to develop image archives of near-empty parking lots on campus with the research goal of using those archives as a baseline for analysis of BSU’s smart parking initiatives.  The aircraft systems (sUAS and software) are low-cost, but they are still able to create high-resolution 2D and 3D web mapping data that emphasizes transportation flows within the LivingLab area.

The research is also working to create Imagery Collection Techniques that comply with the flight rules established in 2016 by the Federal Aviation Administration.  These rules include restrictions on operating unmanned aircraft over people or private property.

Potential applications of the research outside of Bridgewater State, include real-time monitoring and time-series analysis for evaluating mobility investments made by MassDOT, municipalities, and the Massachusetts Bay Transportation Authority.

Partnerships have been a key part of the project’s success.  Collaboration between the BSU Facilities Management and Planning Department, the BSU Police Department, the BSU Aviation Science Department, and the Veterans and Military Affairs Office have been established.  Partnerships have also extended to regional planning and regional transit agencies.

This “LivingLab” initiative builds on a decade of remote sensing research that seeks to apply spatial information technologies to mobility management issues and will help to improve public services locally, regionally, and globally.

Autonomous Vehicle Research: MassDOT Leads the Way

by Tracy Zafian, UMTC Research Fellow

More U.S. states are considering legislation and regulations for highly automated vehicles (HAVs) testing. Twenty-four states and the District of Columbia have now enacted legislation regarding the testing of highly autonomous vehicles. Only Michigan currently allows the driverless HAVs on public roads; California is considering the same but has not approved it yet.

The federal policy (Federal Automated Vehicles Policy) provides guidance for those developing, testing, and deploying highly automated vehicles. The policy considers current and potential regulatory tools that could be used with these vehicles. The policy also describes the different responsibilities on the federal and state levels, and creates a model for state policy that recommends policy areas for states to consider for automated vehicles.

Figure 1: States with Enacted Legislation for Autonomous Vehicles

IOpicAs of July 27, 2017. Source: National Conference of State Legislatures. http://www.ncsl.org/research/transportation/autonomous-vehicles-self-driving-vehicles-enacted-legislation.aspx

In October 2016, Massachusetts Gov. Charlie Baker signed Executive Order No. 572, To Promote the Testing and Deployment of Highly Automated Driving Technologies (EO 572). EO 572 created a state government working group on autonomous vehicles (AV Working Group). The group’s charge is to “convene and consult with experts on motor vehicle safety and vehicle automation…and [to] work with the Legislature on any proposed legislation necessary to protect the public welfare.” The AV Working Group is led by Katherine Fichter, Massachusetts Department of Transportation (MassDOT) Assistant Secretary for Policy Coordination and Transportation Secretary Stephanie Pollack’s designee to the group. The AV Working Group also includes other MassDOT staff and representatives from the State Police, the Executive Office of Public Safety and Security, Housing and Economic Development, and the State Legislature.

One Center, at the UMass Transportation Center, has recently contracted with UMTC Research Affiliates, at UMass Lowell, to conduct research on the technological developments, regulatory requirements, funding opportunities, and potential benefits of the emerging AV technology to take appropriate actions for the benefit of the citizens of the Commonwealth. The affiliates associated with this research are Chronis Stamatiadis, Nathan Gartner, Yuanchang Xie, and Danjue Chen. This project will provide baseline information pertaining to strategic planning for connected vehicle (CV) technologies. This information will be used by MassDOT to develop a strategic plan for the development and deployment of connected vehicle technology and infrastructure in Massachusetts.

EO 572 authorized MassDOT, with input from the AV Working Group and other technical experts, to develop and issue guidance for testing highly automated vehicles on public roadways in Massachusetts, and includes a process for companies to obtain approval for such testing.

Highly automated vehicle testing on public roadways is under the authority of MassDOT. Presently in Massachusetts, most testing takes place in spaces and courses outside of MassDOT’s jurisdiction, such as universities, private indoor testing facilities, and the former Fort Devens military base.

As described by Boston National Public Radio station WBUR, nuTonomy, a Massachusetts Institute of Technology (MIT) spinoff company, began the first testing of highly automated cars on Boston roads in January 2017. The initial testing area was limited to a 191-acre industrial park in South Boston, the Raymond L. Flynn Marine Park, which has a simple road layout, no traffic signals, and only 3 miles of roadway. At first, testing was approved only for daylight hours and good weather, but then was expanded to nighttime and inclement weather. The company has now logged over 200 miles of automated vehicle driving in the industrial park, with no crashes or incidences. With these results, in April 2017, nuTonomy was granted approval to expand its HAV testing to the Seaport and Fort Point areas. A Boston Globe article discussed this approval and interviewed City of Boston and nuTonomy staff. The Seaport roadways are considerably more complex than the testing roads so far, including more complicated intersections, traffic signals, roadways with multiple lanes, bridges, and a rotary. As before, nuTonomy’s testing in the expanded area initially was for daylight hours and good weather only.

In June 2017, MassDOT granted permission for a second MIT-spinoff company, Optimus Ride, to test highly automated vehicles on Boston roads. As described in a Boston Globe article, Optimus Ride will initially test its vehicles only in the Raymond L. Flynn Marine Park, as nuTonomy did.

During their HAV roadway testing, nuTonomy and Optimus Ride both have a human operator sitting in the driver’s seat, ready to take over control of the vehicle if needed. This is currently standard for most on-road testing of HAVs. Some companies use two human workers, one in the driver seat and one in the front passenger seat, to help sustain vigilance and monitoring of the HAV’s driving and the ability to switch to manual driving mode if ever needed. As described in its road test application to MassDOT, after 200 miles of testing, Optimus Ride may request MassDOT permission to test its vehicles with passengers.

In terms of legislation and regulations for automated vehicles (AVs), in her keynote talk at a recent conference on Autonomous and Connected Vehicles held at Worcester Polytechnic Institute, Ms. Fichter indicated that Gov. Baker and MassDOT have taken the position that it is better not to regulate AVs through legislation. AV and HAV technologies are still evolving, and legislation can be difficult to modify once passed. In the Massachusetts Legislature, there are currently eight bills that have been filed related to AVs. On July 13, 2017, the AV Working Group held a legislative meeting to discuss them and hear more about them from their proponents. The MassRobotics Consortium has posted its notes from the meeting. Most of the bills include guidance for AV safety and for liability in the event of a crash involving an AV, with no liability assigned to the original manufacturer of a vehicle that has been later converted to an AV. Joint bills S. 1945/H. 1829 also request that all AVs be zero emission vehicles (ZEVs), encourage AVs to be for public transit only in areas with dense populations, provide guidance for AV data collection, and propose having a vehicle-miles-traveled (VMT) tax on AVs. The idea of a VMT-based tax raised questions and issues at the meeting, related to such issues as geographic equity, fuel consumption and encouraging efficient vehicles, and collection of vehicle owners’ travel data, as well as the need for additional revenues as more vehicles are converted to AVs and electric vehicles.

Among the other proposed AV legislation, H. 2742 requires that AVs used for the interstate transport of goods or for transporting eight or more people be required to have a human operator present who can intervene if needed. Bills S. 1938 and H. 3422 both focus on making AVs that do not require a human operator available to the public. Bills H. 1822 and H. 1897 each request that MassDOT submit a report to the state House and Senate leaders “recommending additional legislative or regulatory action that may be required for the safe testing and operation of motor vehicles equipped with autonomous technology.” H. 1897 requests such a report by June 2017, while H. 1822 requests it by March 2019.

At the end of the July AV Working Group meeting, Ms. Fichter recommended the next meeting would be in September 2017. At this meeting, people from the AV industry will present and provide their perspectives regarding AVs and HAV regulation, and how AV technologies will come to market.

 

 

————-

UMTC Affiliates & MassDOT Assistant Secretary Katherine Fichter Present at WPI Conference on Vehicle Automation

By Tracy Zafian, UMTC Research Fellow

In May 2017, Worcester Polytechnic Institute (WPI) held its second annual Connected and Autonomous Vehicles Summer School speaker series, sponsored by the Institute of Electrical and Electronics Engineers Vehicular Technology Society (IEEE VTS). The event included two days of lectures and discussions.

CAV intersection
Photo source: U.S. Department of Transportation
  • Danjue Chen, Professor at UMass-Lowell and UMTC Affiliate, discussed the impacts of connected and automated vehicles (CAVs) on traffic operations and highway traffic flow, and how CAVs can help optimize roadway capacity and traffic control. Professor Chen is the featured researcher in this month’s Innovative Outlook (IO).
  • Hossein Pishro-Nik, Professor at UMass-Amherst and UMTC Affiliate, spoke about Vehicular Ad Hoc Networks (VANETs) for vehicle-to-vehicle and vehicle-roadway infrastructure communications. His talk discussed the relationship between communications and safety in VANETs, how VANETs can be customized for different traffic conditions and individual drivers, and the issues of privacy in VANETs and Internet-connected devices and applications. Professor Pishro-Nik’s research is described in more detail in another post.
  • Jason Rife, Professor at Tufts University, presented information on different GPS-based technologies and applications that can assist with automated vehicles and navigation, even in dense urban areas with limited sky visibility.
  • Bob Sletten, Engineering Manager at Autoliv, a company that develops automotive safety systems for auto manufacturers, spoke about radar technology in automotive applications.
  • Akshay Rajhans, Senior Research Scientist at MathWorks, spoke about model-based design for connected autonomous vehicles. As described in the WPI conference program, “model-based design makes use of computational models of systems under design that are developed, optimized and checked after correctness specifications throughout the design cycle.”
  • Alexander Wyglinski, WPI Professor and organizer of the conference, provided an overview of vehicular communication systems and the fundamental concepts for understanding, designing, and implementing them.

The keynote speaker at the gathering was Katherine Fichter, Assistant Secretary for Policy Coordination at MassDOT. Ms. Fichter discussed the potential future impacts of driverless vehicles under different scenarios, including a Driverless Utopia and a Driverless Nightmare that were described in Driving Towards Driverless Cars, a blog by Lauren Isaac. Under these scenarios, autonomous vehicles are expected to improve roadway safety, increase vehicle miles traveled, and reduce greenhouse gas emissions, but there are other potential impacts that are less certain. For example, will more driverless cars reduce urban sprawl or increase it, and how will the mobility of low-income people be impacted? As Ms. Fichter discussed, there are questions as well about how autonomous vehicles will be regulated and insured. One big challenge is that current regulations are all based on the idea that vehicles have human operators; this will need to change.

Customizing Your Self-Driving Car

by Hossein Pishro-Nik, University of Massachusetts-Amherst

In the future, intelligent transportation systems (ITSs) will involve connected vehicles, including driver-assisted vehicles and self-driving cars, as well as on-board mobile devices, sensors, and the software and algorithms that govern the functioning of these devices and their communications. Despite recent improvements, each year tens of thousands of lives are lost and billions of dollars are wasted because of traffic inefficiencies in the United States alone. Improvements in the transportation systems could have an enormous impact on lowering these statistics.

In this research, we aim to establish a new approach in design of safety systems, which is based on the individualization and customization of these systems to specific drivers and their environments. This means that wireless communication protocols, as well as algorithms that communicate to users, can be designed in an intelligent way in order to take advantage of all the statistical data that is available regarding the driver and his/her environment.

To accomplish this objective, we can use the technology to collect driver performance data and subsequently learn driver characteristics and driving strategies. This information, along with data collected from other vehicles and roadside units, can be used to customize the technology to each driver. With this, it is possible to adapt warnings or automatic control strategies to each driver. Meanwhile, vehicle-to-vehicle (V2V) communication can be dynamically tuned to make efficient use of finite bandwidth and guarantee the transmission of information critical to safety.

In this way, we should consider that there is an uncertainty of the message delivery between two specific vehicles, while other vehicles might also transmit simultaneously. Our research shows that by proper adaptation of wireless communication and warning algorithms, we can potentially reduce accident fatalities by a considerable amount.

To understand the benefit of V2V communication, consider a traffic stream where a chain of vehicles moves with same speed. When the first vehicle in the chain brakes, the driver of the following vehicle applies the brake after her perception reaction time (PRT). If no intervehicle communications are employed, vehicle Vi applies the brake after the sum of PRTs up to the driver i. With the communications, this time will change to the communications delay plus PRTs of the driver i. This is shown in Figure 1.

Hossein_Fig1new
Figure 1. Communications delay versus sum of PRTs, illustrating the time before a driver in a chain applies the brake

Some drivers may think that some of the received warning messages are not needed, because the drivers are aware of their own response time empirically and they know that they can react to stimuli fast. These warning messages are false alarms for these drivers. These warning messages may frustrate the drivers with an overly high number of false alarms, causing them to ignore warnings or even disable the system. To address this issue, we propose estimating the PRT of drivers and personalizing warning messages based on individual PRTs. Figure 2 shows that at the same accident probability for each driver, the false alarm rate can be reduced by at least 30% by employing the estimated individual distribution instead of the population distribution. Thus, it is of vital importance to minimize false alarms so that the system sends warnings only when they are needed.

Hossein_Fig2new
Figure 2. False alarm rate versus the probability of accident based on using average response time or individual

Now, we should determine how channel access probabilities of vehicles and vehicular communications can be adapted to drivers’ characteristics. In a network of vehicles, each vehicle transmits with a specific probability in the transmission medium. Large channel access probabilities lead the system to excessive interferences and, consequently, low probability of packets being successfully received (success probability), while very small values reduce the success probabilities since the probability of the favorite transmission is low itself. Therefore, there is an optimal value, given both the physical data obtained by vehicular networks and the communications protocol requirements, which results in lower collision probability of vehicles. We can find the expression of packet success probability in a network of vehicles based on channel access probability of vehicle.

We then use a recursive algorithm to tune the transmission probability of each vehicle based on the individual characteristics of drivers. The PRT of the driver, traffic conditions, and communications delay are three factors that play roles in assigning channel access probabilities to vehicles. In simple terms, we categorize the drivers into safe and unsafe drivers based on perception-reaction time. The unsafe vehicles are the ones whose drivers have long perception-reaction time and low distance to the vehicle in front. In other words, unsafe vehicles have higher collision probability. Then we assign different channel access probabilities to unsafe and safe vehicles respectively.

Figure 3(a) illustrates the collision probabilities when channel access probabilities are assumed to be equal for all vehicles. Figure 3(b) shows the scenario in which different channel access probabilities are assigned to unsafe and safe vehicles. The minimum collision probability in the second scenario improves by 25%.

Hossein_Fig3a
Figure 3(a). Collision probabilities when channel access probabilities are equal for all vehicles
Hossein_Fig3b
Figure 3(b). Different channel access probabilities are assigned to unsafe and safe vehicles.

Our simulation results confirm that unsafe vehicles need to inform other vehicles of their perilous situation more frequently than do safer vehicles. In other words, with higher channel access probability for unsafe vehicles, we can achieve lower collision probabilities.

Hossein Pishro-Nik is a UMTC Research Affiliate and an Associate Professor in the Department of Electrical and Computer Engineering (ECE) at UMass-Amherst. This research was supported by the National Science Foundation under Grant CCF– 0844725 (PI: Hossein Pishro-Nik). It is a joint work with ECE PhD students Mohammad Nekoui, Ali Rakhshan, and Mohammad Kohsravi, and Professor Daiheng Ni from the UMass-Amherst Department of Civil and Environmental Engineering. For more information and access to published papers, please visit http://www.ecs.umass.edu/ece/pishro/publications.html.

GM Rolling Out AV Fleet

By Tracy Zafian, UMTC Research Fellow

General Motors Company (GM) announced in mid-June that it completed production of 130 self-driving Chevrolet Bolt electric vehicles for testing automated vehicle (AV) technologies on-road. These highly automated vehicles (HAVs) join GM’s more than 50 Chevrolet Bolts with AV technologies already operating on public roads in San Francisco, Detroit, and Scottsdale, Arizona. In April 2017, Spectrum, the flagship magazine for the Institute of Electrical and Electronics Engineers (IEEE), reported on GM plans to have as many as 300 more self-driving vehicles on-road, presumably including the recently completed 130 vehicles. According to Spectrum, GM would then have the largest HAV fleet on-road not only in the United States, but worldwide. Google-based Waymo has the second-largest AV fleet in the United States, with an estimated 160 vehicles on-road.

GM CEO & Chairman Mary Barra with a new Chevrolet Bolt AV (Photo by Paul Sancya, Associated Press)

In GM’s announcement regarding the 130 new self-driving Bolts, GM Chairman and CEO Mary Barra is quoted: “This production milestone brings us one step closer to making our vision of personal mobility a reality …. Expansion of our real-world test fleet will help ensure that our self-driving vehicles meet the same strict standards for safety and quality that we build into all of our vehicles.” CEO Barra has also said that “no other company today has the unique and necessary combination of technology, engineering and manufacturing ability to build autonomous vehicles at scale.”

The new self-driving version of the Chevrolet Bolt is the second generation of GM’s AVs and is capable of handling almost any roadway situation without human driver intervention. The new Bolts are equipped with the latest technologies in cameras, radar (LiDAR), sensors, and related hardware. “There are even a couple of cameras that are dedicated just to seeing traffic lights to make sure you don’t run red lights,” said Kyle Vogt, CEO of Cruise Automation, a self-driving software company that GM acquired in 2016. The GM HAVs always have an employee in the driver’s seat for safety reasons, just in case any intervention is needed. Almost all states with HAV regulations also have the requirement that a human operator be present.

In 2016, GM also partnered with and invested $500 million in ride-sharing company Lyft. In a recent Forbes article, Cruise CEO Vogt wouldn’t confirm a Reuters report that “thousands” of self-driving Chevrolet Bolt hatchbacks will go into service for ride-hailing company Lyft in 2018, but said it wouldn’t be surprising. “We’ve had a plan in place for a while and it’s going according to schedule. From what I can tell it’s much faster and going to happen much sooner than most people in the industry think,” Vogt said. “We’re planning to deploy in a rideshare environment, and very quickly.