Periodic Reporting for period 1 - Shape-IT (Supporting the interaction of Humans and Automated vehicles: Preparing for the EnvIronment of Tomorrow)
Reporting period: 2019-10-01 to 2021-09-30
The shift to automated road transport has occurred much faster than was expected just a few years ago. Therefore, there is currently a significant skills shortage; major industrial employers and academic departments alike are having considerable difficulty recruiting qualified personnel. The shortage of qualified researchers with multidisciplinary skills spanning the human, computer, and engineering sciences is particularly problematic.
The foremost aim of the SHAPE-IT project is to address this shortage by training early-stage researchers (ESRs) who can design user-centered and safe vehicle automation. Their work will facilitate the transition to safe automated vehicle mobility in the cities of the future. Developers of automated vehicles (AVs) must ensure that humans both inside and outside an AV can understand its capabilities and intent so they can predict its response to their actions (as in AV interactions with non-automated road-users). It is important that everyone (users of AVs as well as other road users) trust the AVs “just enough”: overreliance on the automation is a safety concern, and unwarranted distrust may result in hesitancy using and interacting with AVs.
In addition, developing safe and trusted automation requires tools that can assess the benefits of automation technologies before they reach the road (through road-user models and virtual simulations). From a societal perspective, automated vehicles must be safe, so developers of vehicle automation must have a “safety first” mindset – which is communicated to ESRs within SHAPE-IT.
The SHAPE-IT project is leading the way in addressing AV designs that consider the safety of humans outside the vehicle, by investigating factors affecting vulnerable road users (VRUs), particularly pedestrians and cyclists. Currently, this group makes up more than half of the traffic fatalities worldwide. The safety of this road-user group is consequently a critical societal challenge; the AVs in the cities of the future may be part of the solution. To make this transition safely, the interactions between AVs and VRUs must be considered in the design of the automated systems and the human-machine interfaces.
The SHAPE-IT overall objectives aim to ensure a successful transition to automated traffic in the cities of the future. By doing so, society will reap the full benefits of AVs in our cities.
The foremost aim of the SHAPE-IT project is to address this shortage by training early-stage researchers (ESRs) who can design user-centered and safe vehicle automation. Their work will facilitate the transition to safe automated vehicle mobility in the cities of the future. Developers of automated vehicles (AVs) must ensure that humans both inside and outside an AV can understand its capabilities and intent so they can predict its response to their actions (as in AV interactions with non-automated road-users). It is important that everyone (users of AVs as well as other road users) trust the AVs “just enough”: overreliance on the automation is a safety concern, and unwarranted distrust may result in hesitancy using and interacting with AVs.
In addition, developing safe and trusted automation requires tools that can assess the benefits of automation technologies before they reach the road (through road-user models and virtual simulations). From a societal perspective, automated vehicles must be safe, so developers of vehicle automation must have a “safety first” mindset – which is communicated to ESRs within SHAPE-IT.
The SHAPE-IT project is leading the way in addressing AV designs that consider the safety of humans outside the vehicle, by investigating factors affecting vulnerable road users (VRUs), particularly pedestrians and cyclists. Currently, this group makes up more than half of the traffic fatalities worldwide. The safety of this road-user group is consequently a critical societal challenge; the AVs in the cities of the future may be part of the solution. To make this transition safely, the interactions between AVs and VRUs must be considered in the design of the automated systems and the human-machine interfaces.
The SHAPE-IT overall objectives aim to ensure a successful transition to automated traffic in the cities of the future. By doing so, society will reap the full benefits of AVs in our cities.
Results from some of the 15 unique sub-projects are described here, starting with studies performed on-line (e.g. surveys and interview studies). We now understand that bicyclists are hesitant to be equipped with devices that communicate with AVs, for instance. We have collected some experts’ thoughts about 1) the future of automated driving and 2) the use of augmented reality for pedestrians interacting with AVs. We have also learned that human factors experts should be part of the agile AV development teams, and iterative testing that includes human factors issues should be part of the development cycle.
Switching to in-lab experiments, one ESR has completed a study of a method for assessing how transparent the user perceives a specific vehicle automation to be. Another found that experienced drivers are less sensitive to risk and trust the AV more in dynamic driving, and that nearby road user’s behaviors (relative motion) influence occupants’ perceived risk and trust. Two ESRs worked jointly on a simulator study to investigate how different-colored lighting in the periphery can be used to improve AV user’s trust in automation, by having the lights convey the vehicle’s level of trust to its user. One of the ESRs is focusing on how physiological measures (e.g. EEG) can improve our understanding of users’ trust in automation and the transparency of the vehicle systems.
Other ESRs have re-used data from previous projects: one ESR analyzed existing data to determine how AV users perceive the driving styles of different automated vehicles. Another re-used naturalistic driving data to understand how humans negotiate who goes first in situations where the road is narrow. Their results will help AV engineers develop systems that “behave” in human-like ways.
Other ESRs have focused on determining what information is needed to evaluate the safety of AVs. Specifically, two worked on developing mathematical models of the interaction between AVs, pedestrians, and bicyclists. A third has developed AI (Machine Learning) methods that better predict what pedestrians might do in traffic. A fourth ESR researched the development and application of statistical methods to improve the speed and validity of virtual safety assessments. A fifth ESR has performed a study on methods measuring AV users’ perceptions of safety, showing promising results.
In summary, the research in the SHAPE-IT project is quite diverse, addressing many different facets of the design and development of AVs. Nonetheless, the common goal of all the sub-projects is to facilitate safe, transparent, and trusted AVs in the cities of the future for both AV users and VRUs.
Switching to in-lab experiments, one ESR has completed a study of a method for assessing how transparent the user perceives a specific vehicle automation to be. Another found that experienced drivers are less sensitive to risk and trust the AV more in dynamic driving, and that nearby road user’s behaviors (relative motion) influence occupants’ perceived risk and trust. Two ESRs worked jointly on a simulator study to investigate how different-colored lighting in the periphery can be used to improve AV user’s trust in automation, by having the lights convey the vehicle’s level of trust to its user. One of the ESRs is focusing on how physiological measures (e.g. EEG) can improve our understanding of users’ trust in automation and the transparency of the vehicle systems.
Other ESRs have re-used data from previous projects: one ESR analyzed existing data to determine how AV users perceive the driving styles of different automated vehicles. Another re-used naturalistic driving data to understand how humans negotiate who goes first in situations where the road is narrow. Their results will help AV engineers develop systems that “behave” in human-like ways.
Other ESRs have focused on determining what information is needed to evaluate the safety of AVs. Specifically, two worked on developing mathematical models of the interaction between AVs, pedestrians, and bicyclists. A third has developed AI (Machine Learning) methods that better predict what pedestrians might do in traffic. A fourth ESR researched the development and application of statistical methods to improve the speed and validity of virtual safety assessments. A fifth ESR has performed a study on methods measuring AV users’ perceptions of safety, showing promising results.
In summary, the research in the SHAPE-IT project is quite diverse, addressing many different facets of the design and development of AVs. Nonetheless, the common goal of all the sub-projects is to facilitate safe, transparent, and trusted AVs in the cities of the future for both AV users and VRUs.
At present, extensive resources are being devoted to developing AVs—with the primary focus on reaching the technological maturity needed to have safe, self-driving cars on the road. SHAPE-IT addresses a no-less-critical focus, the human factor. The expected outcomes of SHAPE-IT include new and improved methods for assessing how transparent an AV design is (how effectively the AV “communicates” with people both inside and outside). We also expect to improve several components of the virtual safety assessment and AV algorithm designs, such as modeling the way road users interact (including both cognition-based and more data-driven AI methods). Results from SHAPE-IT are expected to apply to interactions inside the AV (e.g. design strategies improving a variety of human/AV interactions) as well as outside (e.g. communication between AVs and surrounding road users, perhaps through on-vehicle, wearable, or on-bike devices).
The scientific outcomes from SHAPE-IT are likely to help the automotive industry develop safer and more user-centered AVs, and to help policy-makers and road authorities make more informed decisions about AVs. Further, the dissemination of SHAPE-IT results to the broader public may improve people’s understanding of AVs, facilitating a reasonable, calibrated level of trust in AV performance. These actions will in turn likely speed up the use of AVs, with resulting improvements in safety and mobility.
In addition to moving the research on the topic substantially forward, at the end of this project the 15 ESRs in SHAPE-IT will have acquired the key research skills required to study and model human behavior in traffic. Taking a user-centered approach, they will be able to improve the interaction experiences of automated vehicles, while making the design safe. They will have unique competences gained from studying, designing, and evaluating AVs for urban use. In their future employment they will be able to bridge the gap between the domains of human factors specialists and hard-core AV engineers, having been trained in a variety of complementary methods.
The scientific outcomes from SHAPE-IT are likely to help the automotive industry develop safer and more user-centered AVs, and to help policy-makers and road authorities make more informed decisions about AVs. Further, the dissemination of SHAPE-IT results to the broader public may improve people’s understanding of AVs, facilitating a reasonable, calibrated level of trust in AV performance. These actions will in turn likely speed up the use of AVs, with resulting improvements in safety and mobility.
In addition to moving the research on the topic substantially forward, at the end of this project the 15 ESRs in SHAPE-IT will have acquired the key research skills required to study and model human behavior in traffic. Taking a user-centered approach, they will be able to improve the interaction experiences of automated vehicles, while making the design safe. They will have unique competences gained from studying, designing, and evaluating AVs for urban use. In their future employment they will be able to bridge the gap between the domains of human factors specialists and hard-core AV engineers, having been trained in a variety of complementary methods.