CIO Insider

CIOInsider India Magazine

Separator

We May Be Closer to Flying Cars Than We Think, Says This New Study

Separator

It is 2023, yet neither flying, nor fully autonomous vehicles are on the road. Renowned automobile manufacturers, from Mercedes Benz to upstarts like Tesla, appear to be having difficulty launching genuinely secure autonomous driving systems. But, a group of academics from The École polytechnique fédérale de Lausanne (EPFL—the Swiss Federal Institute of Technology Lausanne), a public research university in Lausanne, Switzerland, and JTEKT, a Japanese steering systems supplier, have put together a viable substitute for completely autonomous driving technologies: a clear human-machine partnership.

According to the reports, too much automation in a vehicle's operation can be harmful rather than helpful. This is because fully automated autonomous driving technologies encourage human drivers to become careless and unduly reliant on them. But the researchers have developed and road-tested a haptic-based automated driving system that works with the driver to produce a system that is a good compromise between a manual vehicle and a completely autonomous vehicle.

According to the researchers, deploying completely autonomous driving vehicles on a large scale is exceedingly difficult, and it is unlikely to happen in the next five years. You cannot just exclude the driver from the equation. The out-of-the-loop problem or driver abuse when using ADAS (advanced driver assistance systems) features indicates the still immature state of the current technology, according to the experience obtained in the market with Level 2 vehicles. It is crucial for safety and social acceptance to comprehend driver engagement.


Together with data from cameras, radar, and LIDAR, the researchers' collaborative steering system also incorporates information from the car's steering column. Unlike the most current autonomous driving systems, which are either on or off with nothing in between, the system actively promotes interaction between the driver and the system.

Robert Fuchs, a member of the research team, claims that when people are passively watching an autonomous system rather than actively participating in it, they lose the ability to respond. The researchers sought to boost driver engagement by utilizing automation for this reason.

The system's interaction, arbitration, and inclusion features can be widely categorized. The system starts by defining cooperation, co activity, collaboration, and competitiveness as four distinct types of human-robot interactions.

Identifying the Interaction
Collaboration is when a human and an automation system work together to achieve a common objective, as is the case with modern electric power steering systems that lessen the load on the driver's hands by increasing their input during manual operation. With existing lane aid systems that keep drivers centered in the lane they are currently driving in, co-activity is defined as a situation where the person and the system have different goals, but their activities have an impact on one another.

Collaboration occurs when a human and an automated system work together to accomplish a common goal. It takes a more driver-centered approach, where, for instance, if the system gives the driver control of the steering, it nonetheless stands by to provide backup when the driver is not using it.

Competition describes the extreme situations in which the system and the driver behave in opposition to one another. Although it is an extreme scenario, it can be put into practice in the form of automated emergency steering when the driver cannot avoid a collision. The driver shouldn't be viewed as being completely rejected by the competition.

Arbitrating between Interaction Modes
The system arbitrates or shifts between the several modes after determining the appropriate interaction mode for the given situation. When driving on a highway, for instance, the system might be in co-activity mode when it assists the driver in staying in the center of the lane until it suddenly senses an oncoming collision.

The new technology was connected to a typical sedan for field testing, carried out with five drivers' assistance on a JTEKT test course in Japan. The researchers tested the driver's perceptions of steering smoothness and simplicity of lane-changing during the tests with a particular focus

The system will have to work against the driver at this point because it's possible that the driver won't act quickly enough to prevent the crash. The system will then transition from co-activity to competition mode to counteract the driver's activities and prevent collisions.

Inclusion of the Driver and their Inputs
The system considers the driver's inputs and every junction, which is the last and possibly most important component. The collaborative driving system will recalculate the vehicle's trajectory and integrate it into its measurements, unlike other autonomous driving systems that treat the user's intervention by moving the steering wheel as an "override."

Testing the New Collaborative Steering System
Once it was developed, the system needed to be tested and validated to determine how it affected the driver's safety and comfort. To test it, the researchers ran numerous trials with a human driver operating a detachable power steering system and a virtual driver simulating a virtual driver. The researchers conducted field experiments with a customized test vehicle in addition to testing on a driving simulator.

The new technology was connected to a typical sedan for field testing, carried out with five drivers' assistance on a JTEKT test course in Japan. The researchers tested the driver's perceptions of steering smoothness and simplicity of lane-changing during the tests with a particular focus. The testing verified that the system has great potential to improve comfort while lowering driver effort.

Current Issue
ARETE: Pioneering Cyber Risk Solutions & Transforming The Future Of Cybersecurity