New decision-making model could minimize injuries in inescapable crashes with autonomous cars

In an effort to minimize potential damage and injuries as a result of inevitable car crashes with automated vehicles, scientists have developed new decision-making and motion-planning system

Tesla model 3. Tesla is a company that is well known for making great strives in the field of autonomous vehicle technology - Image Credit: Vlad Tchompalov via Unsplash - HDR tune by Universal-Sci

Tesla model 3. Tesla is a company that is well known for making great strives in the field of autonomous vehicle technology - Image Credit: Vlad Tchompalov via Unsplash - HDR tune by Universal-Sci

In an effort to minimize potential damage and injuries as a result of inevitable car crashes with automated vehicles, scientists have developed new decision-making and motion-planning system and published the result their efforts in the journal: IEEE Transactions on Intelligent Transportation Systems.

Moments after an autonomous vehicle using this new algorithm realizes that an impending crash is unavoidable it starts calculating al accessible options and ultimately decides what action will generate an outcome with the least amount of harm done.

Amir Khajepour, professor at the University of Waterloo, explains that the focus of the science team is aimed at what minimizing potential negative consequences. When a pending crash is imminent, the first principle that a self-driving car should adhere to is to avoid an impact with a pedestrian or pedestrians as the results of such a crash would have the highest probability of causing the most severe injuries. 

If it is possible to avoid such a collision, the next step is to look at, and weigh other conditions such as relative velocities, angles of potential impact, and variations in weight and vehicle type to decide the best possible move. The vehicle might, for example, decide to steer in a different direction or start braking. Dongpu Cao, another Professor at the University of Waterloo, mentioned that the entire traffic environment sensed by the self-driving car is taken into consideration in the process. 

Khajepour added that it is important to note that the idea that automated vehicles will prevent crashes entirely is a myth. Even though safety will probably improve significantly, there are way too many uncertain variables for autonomous cars to manage them all without any mishaps. In other words, accidents will likely still occur, and thus this newly developed system is a necessity.

In recent years the amount of car manufacturers jumping on the EV / autonomous driving vehicle bandwagon has increased massively. Making the creation of a decent decision-making and motion-planning system for potential crashes more relevant by the d…

In recent years the amount of car manufacturers jumping on the EV / autonomous driving vehicle bandwagon has increased massively. Making the creation of a decent decision-making and motion-planning system for potential crashes more relevant by the day. - Image Credit: Y.Leclercq via Wikimedia Commons

Self-driving cars are able to restrict damage when an accident is utterly inevitable as they are much more aware of their surroundings with the information they receive via various types of sensors. In addition, these vehicles are capable of making 100's of decisions per second regarding this information. 

The main idea behind this new system is to generate an optimal response in an emergency circumstance by the use of pre-defined calculations. 

However, the scientists did not try to factor in complex ethical questions like the safety of the occupants of the vehicle versus that of pedestrians or other motorists that might become a part of a pending crash. 

These ethical questions should best we mapped and decided upon by carmakers and lawmakers themselves. As of now, there is at least a start for a framework to implement future determinations made by lawmakers regarding these difficult ethical quandaries. 

If you are interested in this subject you can read the related journal article here: Crash Mitigation in Motion Planning for Autonomous Vehicles

Sources and further reading: University of Waterloo press release via EurekAlert / Self-driving car

More interesting articles on the subject of self-driving cars:


If you enjoy our selection of content please consider following Universal-Sci on social media: