Automated vehicles have received much attention recently, particularly the Defense Advanced Research Projects Agency Urban Challenge vehicles, Google's self-driving cars, and various others from auto ...manufacturers. These vehicles have the potential to reduce crashes and improve roadway efficiency significantly by automating the responsibilities of the driver. Still, automated vehicles are expected to crash occasionally, even when all sensors, vehicle control components, and algorithms function perfectly. If a human driver is unable to take control in time, a computer will be responsible for precrash behavior. Unlike other automated vehicles, such as aircraft, in which every collision is catastrophic, and unlike guided track systems, which can avoid collisions only in one dimension, automated roadway vehicles can predict various crash trajectory alternatives and select a path with the lowest damage or likelihood of collision. In some situations, the preferred path may be ambiguous. The study reported here investigated automated vehicle crashing and concluded the following: (a) automated vehicles would almost certainly crash, (b) an automated vehicle's decisions that preceded certain crashes had a moral component, and (c) there was no obvious way to encode complex human morals effectively in software. The paper presents a three-phase approach to develop ethical crashing algorithms; the approach consists of a rational approach, an artificial intelligence approach, and a natural language requirement. The phases are theoretical and should be implemented as the technology becomes available.
As automated vehicles receive more attention from the media, there has been an equivalent increase in the coverage of the ethical choices a vehicle may be forced to make in certain crash situations ...with no clear safe outcome. Much of this coverage has focused on a philosophical thought experiment known as the "trolley problem," and substituting an automated vehicle for the trolley and the car's software for the bystander. While this is a stark and straightforward example of ethical decision making for an automated vehicle, it risks marginalizing the entire field if it is to become the only ethical problem in the public's mind. In this chapter, I discuss the shortcomings of the trolley problem, and introduce more nuanced examples that involve crash risk and uncertainty. Risk management is introduced as an alternative approach, and its ethical dimensions are discussed.
•AVs are struck from behind 4.8 times more per mile than conventional vehicles.•Most of the crash rate difference was found in urban driving.•Different definitions of “urban” across data sets limit ...the significance.
Automated vehicle developers in California are required to submit records of crashes and distances traveled in autonomous mode for all vehicles in their fleets. Several studies have investigated this database to compare automated vehicle crash rates with national rates. Although automated vehicles are struck from behind in 73 % of their autonomous mode crashes, this is the first study to compare automated vehicle struck-from-behind crash rates to national rates using equivalent crash definitions. Rear-end collisions have substantial public health and economic impacts, representing a third of all collisions and $3.9 B in annual economic costs. In this study, automated vehicles in autonomous mode were found to be struck from behind at 4.8 times the rate of human-driven vehicles in a naturalistic driving study. When controlling for driving environment, the rates for AVs were 5.0 times higher for urban driving and not significant for business/industrial driving, although these results are for different manufacturers, complicating the results. Automated vehicles were more likely to be struck when stopped than when moving compared to human-driven vehicles, suggesting that automated vehicles’ decisions about where and when to stop or remain stopped at intersections are more plausible contributing factors than unexpected rates of deceleration.
The article by Fleetwood1 in this issue of AJPH provides an overview of the public health implications of highly automated vehicles, with a focus on the ethics of a vehicle's behavior when a crash is ...unavoidable, that is, its "ethical crashing algorithms." ASSIGNING VALUES As the authors point out, decisions about how to respond in crash situations require that the vehicle, and by proxy, its designers and software developers, assign values to various objects; otherwise, the car would treat a traffic cone and a pedestrian identically. The trolley problem is useful in discussions because it is fairly well known, represents clear choice with only two distinct alternatives, and assumes completely certain outcomes with obvious moral consequences.3 These attributes strike vehicle developers as unrealistic and naive4; real driving dilemmas have many subtle choices, uncertain outcomes, and often an obviously superior course of action, for example, apply the brakes. The distribution of...
Traffic Signal Control with Connected Vehicles Goodall, Noah J.; Smith, Brian L.; Park, Byungkyu (Brian)
Transportation research record,
01/2013, Letnik:
2381, Številka:
1
Journal Article
Recenzirano
Odprti dostop
The operation of traffic signals is currently limited by the data available from traditional point sensors. Point detectors can provide only limited vehicle information at a fixed location. The most ...advanced adaptive control strategies are often not implemented in the field because of their operational complexity and high-resolution detection requirements. However, a new initiative known as connected vehicles allows the wireless transmission of the positions, headings, and speeds of vehicles for use by the traffic controller. A new traffic control algorithm, the predictive microscopic simulation algorithm, which uses these new, more robust data, was developed. The decentralized, fully adaptive traffic control algorithm uses a rolling-horizon strategy in which the phasing is chosen to optimize an objective function over a 15-s period in the future. The objective function uses either delay only or a combination of delay, stops, and decelerations. To measure the objective function, the algorithm uses a microscopic simulation driven by present vehicle positions, headings, and speeds. The algorithm is relatively simple, does not require point detectors or signal-to-signal communication, and is completely responsive to immediate vehicle demands. To ensure drivers' privacy, the algorithm does not store individual or aggregate vehicle locations. Results from a simulation showed that the algorithm maintained or improved performance compared with that of a state-of-the-practice coordinated actuated timing plan optimized by Synchro at low and midlevel volumes, but that performance worsened under saturated and oversaturated conditions. Testing also showed that the algorithm had improved performance during periods of unexpected high demand and the ability to respond automatically to year-to-year growth without retiming.
How to Think About Driverless Vehicles Goodall, Noah J.
American Journal of Public Health,
09/2018, Letnik:
108, Številka:
9
Journal Article, Book Review
Recenzirano
Odprti dostop
The automation of driving may also put five million American professional drivers out of work, enable a foreign government to halt traffic via cyberattack, allow police to effortlessly lock down and ...cordon off a city, or paradoxically increase congestion as individual car travel becomes so cheap and comfortable as to discourage public transit to the point of elimination. Machine vision- that is, the ability for a computer to tell the difference between a car and a pedestrian from video stills-is of obvious importance for a self-driving vehicle navigating a busy road. The Tesla that fatally struck the side of a white tractor trailer in Florida had both a detection and classification error, as the car's radar was looking too low for bumpers and missed the elevated truck trailer, while the camera saw the truck but interpreted the white truck as part of the sky.3 A minor crash between a Waymo automated vehicle and a bus in 2016 was an error ofprediction, as the Waymo vehicle incorrectly assumed that the bus would wait for Waymo to merge into traffic.4 These errors can be drastically reduced if vehicles can simply communicate their location (I'm here!), status (I'm a car!), and intentions (I'm about to merge!) directly with nearby vehicles.
AbstractComputer-driven vehicles will behave differently from human-driven vehicles due to changes in perception abilities, precision control, and reaction times. These changes are expected to have ...profound impacts on capacity, yet few models of automated driving are based on empirical measurements of computer-driven vehicles in real traffic. To this end, this paper investigates characteristics of an early form of longitudinal control automation, a commercially available adaptive cruise control (ACC) system driven in real traffic. Two car-following models were calibrated to a vehicle with ACC. First, the Intelligent Driver Model was reformulated to comply with ACC design standards then calibrated to match speed and range data from the test vehicle. The vehicle with ACC was found to decelerate less severely than predicted by the model when tested in severe braking and unimpeded acceleration scenarios. Second, the Wiedemann 99 model was calibrated because it is the default car-following model in the traffic microsimulation software program Vissim and can therefore be implemented cheaply and quickly in sophisticated models of roadways worldwide. Four parameters of the Wiedemann 99 model were measured directly from field observations of the test vehicle: standstill distance, start-up time, unimpeded acceleration profile, and maximum desired deceleration. Simulation results in Vissim were found to match the adaptive cruise control in unimpeded acceleration tests. These findings will benefit researchers and modelers seeking more accurate models of car-following behavior with adaptive cruise control and automated longitudinal control.
Many transportation agencies use re-identification technologies to identify vehicles at multiple points along the roadway as a way to measure travel times and congestion. Examples of these ...technologies include license plate readers, toll tag transponders, and media access control (MAC) address scanners for Bluetooth devices. Recent advancements have allowed for the detection of unique MAC addresses from Wi‑Fi and wireless local area network enabled devices. This study represents one of the first attempts to measure the fundamental characteristics of Wi‑Fi re-identification technology as it applies to transportation data collection. Wi‑Fi sampling rates, re-identification rates, range, transmission success rates, and probability of discovery of sensors and mobile devices were measured, and a model of probability of detection is presented. Field tests found that mobile phones routinely experienced significant time gaps between Wi‑Fi transmissions. The study recommends that Wi‑Fi sensors be deployed at low-volume, low-speed roadways, with sensors positioned near intersections where vehicles are likely to slow or stop. Due to Wi‑Fi's relatively low probability of discovery, the technology may produce poor results in applications that require re-identifying vehicles over multiple consecutive sensors.
Most automobile manufacturers and several technology companies are testing automated vehicles (AVs) on public roads. While automation of the driving task is expected to reduce crashes, there is no ...consensus as to how safe an AV must be before it can be deployed. An AV should be at least as safe as the average driver, but national crash rates include drunk and distracted driving, meaning that an AV that crashes at the average rate is somewhere between drunk and sober. In this paper, safety benchmarks for AVs are explored from three perspectives. First, crash rates from naturalistic driving studies are used to determine the crash risk of the model (i.e., sober, rested, attentive, cautious) driver. Second, stated preference surveys in the literature are reviewed to estimate the AV risk acceptable to the public. Third, crash, injury, and fatality rates from other transportation modes are compared as baseline safety levels. A range of potential safety targets is presented as a guide for policymakers, regulators, and AV developers to assist in evaluating the safety of automated driving technologies for public use.
Vehicle manufacturers are beginning to introduce battery electric pickup trucks, with at least three models in production in the United States and at least six others announced. Unlike hybrid and ...plug-in hybrid vehicles, which use batteries to supplement an internal combustion engine, battery electric vehicles are fully reliant on their batteries and have significantly shorter ranges, fewer refueling/recharging options, and may experience shorter ranges under towing, mountain driving, and temperature extremes. This study investigated real-world pickup truck usage data from a large state department of transportation to determine whether battery electric trucks could effectively replace trucks based on manufacturer-stated range as well as early field tests of actual vehicle range under various speeds, idling scenarios, and weather conditions. The results indicated that 97% of pickup truck day trips could be completed on a standard range battery, and 99% on an extended range battery. Among tracked department of transportation fleet pickup trucks, 31% could be replaced by a standard range electric truck with no change in operation; and up to 64% of the fleet with an extended range could be replaced by an extended range electric truck. Dynamically assigning long trips to dedicated conventional engine pickup trucks could further reduce the number of nonelectric trucks required by half.