Human Computer Interaction in Autonomous Vehicles
Particularly in the public press and in non-scientific debate, automated vehicles are occasionally perceived as completely replacing a human driver. Although automated vehicles have seen some progress, no systems are yet capable of doing this. Instead, we contend that interactions between humans and automated systems in the context of automobiles should be seen as a partnership between the two. In other words, instead of choice between the human and the system, the human actively participates in specific driving-related duties when appropriate.
For many years, research has been heavily influenced by human-Computer interaction. After all, effective interactions with the car and any equipment that are integrated into or transported in the car are necessary for safe manual control of a vehicle. Surprisingly, despite the rise of autonomous driving, human-computer interface more generally, the place of humans in cars continue to be crucial considerations in vehicle design.
The study of HCI in autonomous vehicles has been growing recently. In the following situations, automation is used more and more:
• Situations requiring speed and safety, like those in hospitals, power plants, and transportation
• Application by non-professional users, such as in robot vacuums, personal medications, and intelligent chatbots
• In physical systems that dynamically interact with their environment, such as drones, robot vacuums, and aircraft.
Dynamic Humans and Dynamic Systems: Mode Confusion
1) In many ways, people are dynamic. The first is that users may not use systems as intended. For instance, reducing the possibility of human error is one reason for automated driving. Systems attempt to accomplish this by minimising the amount of fundamental vehicle tasks that the driver must complete, allowing them to concentrate more on other crucial driving-related tasks like monitoring traffic and foreseeing hazardous situations. Unfortunately, studies have found that people perform tasks unrelated to driving more frequently when there is a greater transfer of responsibility from the human to the machine. As a result, situational awareness is diminished (as opposed to improved), and reactions to safety-critical events take longer (as opposed to shorter). This is what is known as “irony of automation,” where the introduction of automation changes human behaviour and creates new, unrelated problems rather than resolving an issue (such as reducing human errors). It can also be viewed in the broader context of people repurposing technology that they are already familiar with from other types of HCI technology use. The person is still dynamic.
2) Humans learn and unlearn skills, habits, and knowledge, which is another way they are dynamic. These factors work together to shape assumptions about a system and particular interaction patterns. The proper use of a system, however, may be constrained by incomplete knowledge (of its operation) and habits that make it difficult to pick up new skills. Although it might not always be available right away, active training could aid in overcoming these biases (for example: might not be there immediately after a software update).
3) Automated systems and the environments in which they operate are dynamic, just like humans. First, automated systems frequently learn from their environment as an instance of artificial intelligence, which can shape and modify their responses over time. Second, automated systems are frequently created for use in particular scenarios. The Society of Automotive Engineers (SAE) has identified specific levels of automation for a car’s functionality in the case of automated vehicles, which can be applied in particular “operational design domains” or contexts. However, the environment in which a car is driving can change over time and space, which can affect how well the system works and how reliable it is. For instance, on common highways with normal traffic and weather, adaptive cruise control might keep a vehicle’s speed and a safe distance from vehicles in front of it. However, the context may alter what is safe: The system might not respond appropriately if there is an unexpectedly heavy snowfall, and the car’s obligations may not be fulfilled.
Negotiated Attention Interleaving
1) It is crucial to create precise frameworks, theories, and systems to negotiate and communicate transitions and to direct attention in human-computer interaction settings because both humans and machines are dynamic and system states can change.
2) Interruption studies have influenced both research and human-computer interaction. How interruptions are handled in professional settings is the subject of extensive observational studies, critical experimental tests, and detailed theoretical models and frameworks. However, the typical interruption study makes the assumption that people are engaged in a task that is urgent and occasionally life-threatening (like manually operating a vehicle) (such as a phone call).
3) When automated driving is enabled, this changes. If automated vehicles live up to their promises, then the need for the human to contribute to the driving task will decrease as they become more capable of driving over lengthy stretches of road with little to no human assistance. As a result, people may begin working on non-driving activities such as checking email, getting ready for a meeting, or participating in a video chat with a distant conversant. The human might then regard these tasks as being of utmost importance. The human will occasionally need to operate a vehicle, though. However, the human might perceive the request from the car to begin driving as a disruption of the original non-driving task that they were working on (this is known as a “transition of control”). Therefore, in automated vehicles, driving is interrupted by other tasks rather than other tasks being interrupted by driving (such as a phone call).
Conclusion:
Human-automation interaction and human interaction with automated vehicles are exciting fields that are currently in development. Although there have been many technological advancements, the human experience is also being taken into account more and more. Additionally, humans continue to play a significant role in many automation scenarios, including automated driving. Given that both people and automated systems are dynamic, the full range of HCI methods and theories have the potential to shed light on a wide range of users and scenarios. Human behaviour can be better understood by using HCI and its parent, partner, and related disciplines. Engineering advancements, for instance, can influence the design of automated systems, and interruption handling theory can influence interventions for human-automation interaction.
References:
https://www.frontiersin.org/articles/10.3389/fpsyg.2019.02757
https://interactions.acm.org/archive/view/march-april-2020/automated-driving
