Customizing Your Self-Driving Car

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.

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.

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%.

Figure 3(a). Collision probabilities when channel access probabilities are equal for all vehicles
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.

Written by Hossein Pishro-Nik, UMTC Research Affiliate and 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


GM Rolling Out AV Fleet

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.”

Written by Tracy Zafian, UMTC Research Fellow. 

Safety First! Are You a Distracted Driver or a Distracted Pedestrian?

The annual number of pedestrians hit and killed by vehicles in the United States is now at its highest level in more than 20 years. In March 2017, the Governors Highway Safety Association (GHSA) released a report showing an 11 percent increase rise in the number of pedestrian deaths between 2015 and 2016, and a 25 percent increase in these deaths over the past five years. The report estimates here were almost 6,000 pedestrian fatalities in 2016 and pedestrias now account for 15 percent of all traffic deaths. The rise in pedestrian fatalities from 2015 to 2016 was the highest annual increase in both the total number and percentage growth in the 40 years that these national data have been recorded.

The GHSA figurpedses are calculated based on pedestrian fatalities for January to June 2016 and then extrapolated for the rest of the year. For this six-month period, 2,660 pedestrians died in traffic crashes nationwide. Four states accounted for 43 percent of these fatalities: California (405 pedestrian deaths); Florida (277); Texas (242); and NewYork (137). Massachusetts had 38 pedestrian deaths in this time frame( 1.4 percent of the total).

Source: Seattle Times

The GHSA identified several factors that could be contributing to the rise in pedestrian deaths, including the following.

  • More driving. People are driving more now, with the economy improving and gas prices down from their historic high levels ($4+/gallon) earlier this decade. Federal Highway Administration data released in February 2017 show that in 2016, people in cars, minivans, SUVs, and trucks drove a record 3.22 trillion miles on the nation’s roads and highways. This is an increase of 3 percent over 2015, and the fifth straight year of increased total mileage.
  • Alcohol. According to the GHSA report, 15 percent of pedestrian taffic deaths involve a drunk driver, and 34 percent of the pedestrians killed in traffic accidents themselves have blood alcohol levels above the legal limit of 0.08.
  • Lack of pedestrian visibility. Many of the pedestrian fatalities occurred in conditions where the pedestrians may not be very visible to drivers. The GHSA found that 74 percent of pedestrian deaths occurred at night, and 72 percent of those killed were not at a roadway intersection.
  • In recent years, as cell phones and other portable communication and entertainment devices have become more ubiquitous, there has been an increase in crashes and injuries attributed to distraction. Drive distraction is considered one of the top three causes of traffic fatalities in general—the other top causes are alcohol and vehicle speed—and one of three main causes for pedestrian fatalities. The National Highway Transportation Safety Administration (NHTSA) found that driver distraction contributed to 3,477 traffic crash-related deaths and 391,000 injuries in 2015. As discussed in a recent National Public Radio piece, there are also concerns about the impact of pedestrians’ own distractions on pedestrian safety

A comprehensive research literature review on the impact of electronic device use on pedestrian safety was conducted by Robert Scopatz and Yuying Zhou (2016). The literature review was part of a larger research project examining whether electronic device use by drivers and pedestrians significantly affects pedestrian safety. The literature review included sections on distracted pedestrians, distracted drivers, and pedestrian-driver interactions, and examined real-world studies, simulator studies, and other collected data in these three areas. There have been no studies thus far showing a direct cause-and-effect link between distraction and pedestrian crash risk. Nonetheless, there is clear evidence that distracted drivers face increased crash risks and that distraction impacts how pedestrians walk, react, and behave, including safety-related behaviors

Scopatz and Zhou found only one study (Brumfield and Pulugurtha, 2011) to date that examined pedestrian-vehicle conflicts and the role of distraction due to handheld electronic device use. That study’s researchers observed 325 pedestrian-vehicle interactions at seven midblock crosswalks on a university campus in Charlotte, North Carolina. They found that 29 percent of pedestrians and 18 percent of drivers were noticeably distracted (talking on a cell phone or texting) at the time the pedestrian and vehicle were nearing the crosswalk. Further, the researchers calculated that distracted drivers were more than three times more likely to be involved in a conflict at the midblock crosswalks than distracted pedestrians. Government legislators in Montreal, Quebec, and New Jersey have proposed banning cell phone texting for pedestrians while they are crossing the street. These proposals have not received much support thus far.

Research is needed to dig deeper into the issues around pedestrian fatalities with specific focus on distraction.

Some key questions remain:

  • How distractions (for drivers and pedestrians) exacerbated by hazards that are already present?
  • With the encouragement of Bicycling and Pedestrian activity for healthy communities, how will this impact the grown problem?
  • What type of solutions are States considering for solutions? One recent report published in March of 2017,  Consensus Recommendations for Pedestrian Injury Surveillance aims to offer guidance in tracking, recording and prevention.

By: Tracy Zafian, UMTC Research Fellow with input from Affiliate Researcher, Karin Goins from UMass Medical


YouTube Research Spotlight: Research to Improve At-Grade Rail Crossing Safety

The UMTC Research Section Launches a Research Spotlight YouTube Channel. We are showcasing research currently being conducted on “At-Grade Rail Crossing Safety” by Radhameris Gomez.  Ms. Gomez is a PhD candidate in the UMass Transportation Engineering Program at the University of Massachusetts, Amherst. View the overview video (3 minutes) or the extended video (10 minutes) to find out how she became interested in studying transportation engineering.

TrailCrashes at roadway-railroad intersections happen far too often. Federal Railroad Administration data show that 2,025 such crashes occurred in the United States in 2016, resulting in 265 fatalities and 798 injuries. There have been a number of roadway-rail intersection crashes recently. For example, in Florida, an Amtrak train collision with a car left one person dead; in Arkansas, one person was killed and another injured when their car crossed into a train’s path; and in North Carolina, a train crashed into a car that stopped on the railroad tracks when the safety arms came down, and the car driver was killed. Earlier in March, a freight train collided with a charter bus in Mississippi that had become stuck on a rail crossing with low clearance on the crest of a slope. Four people were killed and others injured; it was the 161st crash since 1976 at that crossing. After a March snowstorm, a local DPW worker in Longmeadow, Massachusetts, died when his snowplow backed onto railroad tracks when a train was coming. At that intersection, there are no gate arms or traffic signals to help warn drivers when a train would be coming; there had been five other crashes and four other deaths at that location since the 1970s.

Previous studies have examined primary contributing factors for grade-crossing train-car crashes and how these crashes can be prevented. Jeff Caird and colleagues at the University of Calgary analyzed over 300 grade-crossing crashes in Canada (2002). They estimated that adding flashing lights to a rail crossing without them has the potential to reduce crashes by over 60 percent, as compared to crossbucks alone. Michael Lenné and colleagues at Monash University in Australia conducted a driving simulator study (2010) on driving behavior at three different types of at-grade rail crossings: stop-controlled, with flashing lights, and with a traffic signal. The researchers found that participants slowed their vehicles the most when approaching rail crossings with flashing lights.

By: Tracy Zafian, UMTC Research Fellow


Simulator Evaluation of the Effectiveness of an Comprehensive Teen Driver Training Program

Novice teen drivers are over represented in crashes, particularly rear end, intersection and run- off-the-road crashes. Their over involvement in these crashes appears to be due to six poorly developed skills: tactical and strategic hazard anticipation, tactical and strategic hazard mitigation, and tactical and strategic attention maintenance. Previous studies had determined that a single skill could be taught in a 45 minute training session. The question addressed here was whether all six possible skills could be taught in a two hour session without reducing the effectiveness of the training of the individual skills. Specifically, the current study examines the development and evaluation on a driving simulator of a training program, ACCEL (Accelerated Curriculum to Create Effective Learning), that is designed to decrease the time it takes teens to become safer drivers over the first 18 months of independent driving by targeting for training the above six behaviors in the most risky crash scenarios. During the evaluation, eye movements were recorded and vehicle measures were collected for a total of 75 novice drivers (16 to 18 14 years with less than 6 months’ experience), of which fifty were ACCEL-trained and 25 were Placebo-trained, and 25 experienced drivers (28 to 55 with at least 10 years’ experience), all untrained. ACCEL training was found to significantly improve the performance of novice drivers in 5 out of the 6 of the trained skills when compared to Placebo trained teens: tactical and strategic hazard anticipation, tactical hazard mitigation, and tactical and strategic attention maintenance. The results are consistent with the hypothesis that combined skill training can be deliver effectively in a relatively short amount of time.

Innovative Strategies for Safer Cycling

Research in progress at the University of Massachusetts underway to evaluate newer bicycle infrastructure treatments such as bike-boxes, merge lanes, and protected intersections to identify patterns around driver behavior and performance when approaching these new innovative bicycle infrastructure treatments. The information collected can then be used to develop countermeasures such as infrastructure geometry, signage, training campaigns, etc. The goal of this information is to promote cycling by mitigating bicycle safety concerns through improving driver awareness at new and unfamiliar bicycle infrastructure treatments. For more information please click here.

By: Eleni Christofa and Nick Fournier