Human Factors in Driverless Cars
I’m Ed Smith, a Self-Driving Car Accident Lawyer. Autonomous cars promise to reduce accidents, save time, and dramatically improve human efficiency. Manufacturers of self-driving car technology, such as GM, Daimler-Bosch, Ford, Volkswagen, BMW, Renault, and Tesla, along with ride-sharing serving providers such as Uber are already testing semi-autonomous cars on the roads. These cars currently do have a human backup driver present in the vehicle during a road tests. Analysts estimate that the initial phase of commercial release of these vehicles on the roads will also involve some degree of human interaction.
How Will The Human Factor Impact The Safety and Performance of Driverless Cars in California?
This leads us to a vital question: Are these autonomous vehicles being designed in a way that is in harmony with the capabilities and limitations of an average driver, enabling the driver to make timely and safe interventions when needed?
Level of Human Involvement in Self-Driving Cars
According to National Highway Traffic Safety Administration (NHTSA), automated cars are defined as vehicles that can perform some key driving functions, such as braking, throttle, or steering, without requiring direct input from a human driver. Levels of automation in such cars can vary, and drivers are required to interact with the vehicle in consonance with its specific level of automation.
Level 1 Automation
NHTSA says that Level 1 automation in a car requires a driver to operate the vehicle physically with their hands or feet at all times, but the vehicle may include some automated driver assist features.
Level 2 Automation
At this level, a driver’s presence in the vehicle is required for monitoring the roadway. The driver should be readily available to manually control the vehicle quickly in an emergency.
Level 3 Automation
At this level, no monitoring of the roadway or traffic environment is required on the part of a human driver. However, a driver must still be present in the vehicle to take physical control when needed.
Level 4 Automation
The driver’s role is limited to providing inputs for destination and navigation, but a driver is not required to be available for controlling the vehicle at any point during the trip.
Level 5 Automation
The car performs all driving functions in all conditions, but a driver may have an option to control the car.
Watch YouTube Video: How Tesla’s Self-Driving Autopilot Actually Works
The following video discusses how the technology in self-driving cars, such as Tesla, function on the road.
Challenges for a Human Driver in an Autonomous Vehicle
Human drivers have at least some role to perform at all levels of vehicle automation. Two key tasks for drivers at early levels of automation include:
- Monitoring the road environment and the automated system
- Making an emergency physical intervention when required
Brittleness
Automated features are usually employed only in specific conditions, such as highways or when in clear weather situations. In other words, autonomous vehicles in their current form suffer from the limitation of “brittleness”.
Brittleness in this situation means that the autonomous system will perform well in conditions for which it was designed, but human intervention will be required in situations that the AI program was not designed for.
As a result, a human driver faces a challenge while operating these vehicles because they may not realize when the automated system is not performing correctly, or may not understand why it is not performing correctly.
Slower Reaction Time
Research shows that human beings are slow to recognize that there is a problem with an automated system and slow to understand it after it has been identified. Therefore, in the event of a sudden automation failure, a human driver may not be prepared for a quick takeover of manual controls.
Situation recognition and mental awareness of human beings are adversely impacted by vehicle automation. Therefore, the perception and reaction time is greater at higher levels of automation. Research shows that drivers of autonomous cars at Levels 1 and 2 have a delayed braking reaction time by up to 1.5 seconds compared to an average driver operating a manual vehicle.
Complacency
Complacency tends to set in for a driver as the level of automation increases. This may induce them to get involved in distractions such as reading/interacting with a mobile device or talking on their cell phone. As the level of automation goes up, so does the likelihood that a driver will indulge in tasks unrelated to driving.
These distracted drivers are likely to require more time to re-enter the driving loop, regain their situation awareness, and respond to a request for intervention. Depending on the type of non-driving task they may be engaged in, the takeover time for a driver may get delayed by as much as 25 seconds.
To ensure the safety of semi-autonomous cars, the designers of this technology should ensure that the system accommodates a driver who takes the longest to regain driving control and not the “average” driver.
Factors Affecting Driver Response
A human driver’s ability to respond and avoid a car accident when an autonomous feature fails will depend on factors such as the driver’s mental workload, their level of situation awareness, the condition of the roadway at that moment, and the time available to avert a collision.
Either the driver must be able to quickly detect that the driving situation can no longer be managed by the automated system, or the vehicle user interface must be designed in a way to alert the driver to this fact. Once the situation is identified, the driver should have adequate time available to take over manual controls.
The time required for this takeover will depend on the specific driving situation and how much time is required by a human driver to assess that situation and formulate a response. In addition, the time available in such situations also depends on the capacity of the sensors of the autonomous vehicle and the ability of the sensors to predict an emergency need for human intervention.
Practical Hurdles at Advanced Levels of Vehicle Automation
Will autonomous cars be able to ‘think’ and act like humans in actual road conditions at advanced levels of automation? Consider an imaginary situation where you are a pedestrian walking along a countryside road on a dark night.
In this situation, if you find a car getting closer to you, you are likely to be more watchful of your steps even when you are on the roadside at a safe distance. You will exercise more caution because it is dark and you assume that the driver may or may not have noticed you. On the other hand, if you are a pedestrian in an urban area during peak daylight, you would not worry about crossing the street in front of a car at a stop sign. A lot of interaction between a pedestrian and a car driver on the road happens through eye contact, body language, and instinctive behavior. Based on these interactions, they both make mutually safe decisions on the road.
How will the autonomous vehicles interact with pedestrians in different road conditions, at different times of the day, and with people of various age groups and having their own peculiar natures?
Will driverless vehicles be able to account for human caprices, cultures, social habits, and the countless subtle behavioral nuances into their decision-making process? Will other manual drivers and pedestrians have the confidence to share the road with these self-driving cars?
It is one thing for self-driving cars to be on the highways and wide streets, but it entirely another to negotiate their way in communities, neighborhoods, parking lots, and near schools.
Crash Investigations for Self-Driving Vehicles
As automation elements increase in the modern vehicles, the role of a driver is undergoing a transformation. Therefore, when crashes involving autonomous or semi-autonomous cars occur, the experts should take into account this changing role of a driver.
Clearly, any accidents involving self-driving cars will require an analysis and investigation while taking human factors into consideration.
Investigators will have to determine whether any action or inaction of a driver was a factor in an accident, and more importantly, whether the automated system was designed in a way to integrate itself with the known skills and limitations of drivers who operate these systems.
Who Do We Hold Responsible?
One of the most vital questions that will have to be resolved with regard to self-driving vehicles is: who must take the blame when a collision occurs? Can a human driver really be held responsible for a crash when the technology itself is designed to minimize or eliminate their role and involvement in driving?
Can the manufacturers be held responsible for a self-driving car crash? Will insurance companies be prepared to offer coverage for these vehicles with regular, affordable premiums? Or, will these insurance companies look for ways to avoid paying anything for crashes involving autonomous cars.
Answers to these questions are not easy to find. But it is time to explore all these areas and create viable solutions so that the transition to autonomous driving technologies can occur smoothly while safeguarding the interests of the consumers.
Self-Driving Auto Accident Attorney
I’m Ed Smith, a Self-Driving Car Accident Lawyer. When you or your loved one suffers injuries in an accident caused by an autonomous vehicle, you may still be able to be compensated for your injuries and related damages. Please feel free to give me a call anytime at (916) 921-6400 or toll-free at (800) 404-5400 for free, friendly advice.
To read reviews of my legal practice by my former clients, visit the following sites:
I am a member of the Million Dollar Advocates, a forum which recognizes trial lawyers nationwide who have won a verdict or settlement worth $1 million or more for previous clients.
Additional information on my past cases and their outcomes is available on my Verdicts and Settlements page.
Image Attribution: By “Geralt” via Pixabay under CC0 Creative Commons
Source of: Human Factors in Driverless Cars – AutoAccident.com
:br ds [cs 1700] cv