When crossing the street one of the firstthings most pedestrians do when they see an oncoming vehicle is make eye contact with the driver. This is one way to ensure that the driver has seen you. Being seen by the driver is important to ensuring that you can cross the streetsafely.Now imagine doing the same scenario as a pedestrian, only when you attempt to make eye contact with the driver you discover that the vehicle has no [31](1.passenger 2. driver 3. observer). Do you cross the street? That situation is likely to become an everyday occurrence with the widespread [32](1. adaptation 2. regulation 3. hearsay) of automated vehicles (AVs). The answer to that question cannot be to wait and let the AV pass and then cross the street.To be fully integrated into our society,AVs need to be navigated in much the same way as other vehicles.
Research on pedestrians interactions with manually driven vehicles has highlighted the important role of communication between pedestrians and vehicle drivers in ensuring safe interactions. This communication is often done through [33](1.timed 2. digital 3. verbal) exchanges, hand gestures, or eye contact between pedestrians and vehicle drivers. The removal of the driver presents new challenges to facilitating the communication needed to ensure pedestrian safety.The research on pedestrian-AV communication can be divided into those examining AV-to-pedestrian communication and those examining pedestrian-to-AV communication.
First,research on AV communication with pedestrians focuses on leveraging the use of devices on the AV to promote communication with the pedestrians. The most commonly studied devices are light-emitting diode (LED) message boards. These LED message boards are located on various parts of the AV (e.g.,side panels, windshields, and overhead). Research is being conducted to determine the best locations for placing the LED boards on AVs. There is also ongoing research on what information these message boards should display.For example, should they display what the AV is currently doing(i.e., stopping)or what the pedestrian should be doing(i.e., cross now). One of the biggest limitations to the use of LED messages is related to scalability. An LED board might display a message intended for one pedestrian but be read by another pedestrian. For example, an AV’s LED board might display a message that it is safe for pedestrian “A” to cross but also have the message read by pedestrian “B” whom the AV was [34](1.cautious 2. unaware 3. independent) of and to whom the AV did not intend to communicate that it was safe to cross. This could result in at least one pedestrian misreading the AV’s intention. Another example of the scalability problem is the increase in the cognitive load：[35](1.balanced 2. i1nposed 3. transfeITed) on a pedestrian as the number of AVs with LED boards increases. As the number of AVs that the pedestrian [36](1.encounters 2. utilizes 3. maneuvers) increases, so does the number of potential LED messages to read. A pedestrian reading one message from one AV is certainly manageable but messages from two, three, or four become somewhat more difficult.This is especially t1ue when you factor in the habits and behaviors associated with many pedestrians such as text messaging and e-mail reading.In addition, when you [37](1.couple 2. double 3. compare) the first scalability problem with the second scalability problem it bec01nes easy to see how issues related to scalability can magnify. Scalability problems are not [38](1. irreplaceable 2. insurmountable 3. discouraging), but they do present ongoing challenges with the use of LED boards as a standalone solution.
Second, the removal of the driver presents another problem – the ability of the pedestrian to communicate with the AV. Common ground or a shared understanding helps to promote communication. One important source of common ground between pedestrians and drivers is based on their shared experiences. In many cases drivers have been pedestrians and pedestrians have at least ridden in a vehicle, [39](1.and yet 2. if not 3. let alone) driven a vehicle. This creates com1non ground between the driver and the pedestrian, which facilitates communication. However, AVs have not been pedestrians and AVs do not always [40](1. assist 2. challenge 3. mimic) human drivers in their behavior or decision- making. Both make it difficult for the AV and the pedestrian to establish common ground. Researchers are conducting studies to determine how pedestrians communicate their intention implicitly through their body language and behavior.Models employing machine learning are being developed to teach AVs how to interpret implicit communication from the pedestrians so that they can react to them correctly. However, the dynamic and emergent nature of these interactions makes modeling these interactions particularly challenging.
コメントを残す