Jaguar Land Rover Makes Eye Contact with New AV Research

August 29, 2018

Stuart Birch

Eye contact: JLR’s pedestrian-communication research Eye Pod. A production communication system is unlikely to use big blue eyes. (JLR)

Jaguar Land Rover (JLR) engineers developing autonomous-vehicle (AV) driving systems are working with cognitive psychologists to determine how users—and particularly pedestrians—might be encouraged to fully trust new AV technology.

For this research study, JLR engineers are using an AV with large, blue virtual “eyes” that can track a pedestrian—although a production version might be expected to accomplish the goal in a less literal fashion.

The JLR psychological work comes alongside a recent report of some in the tech community suggesting pedestrians should be expected to “reprogram” their own behavior to make operation easier for AVs. The story aroused new skepticism regarding the prospects for near-term AV deployment if some developers believe humans should be “trained” on how best to survive in the automated-driving future.

Two-way street

“We want to know if it is beneficial to provide humans with information about a vehicle’s intentions—or whether simply letting a pedestrian know it has been recognized is enough to improve confidence,” said Pete Bennett, JLR’s Future Mobility Research Manager.

Called “trust trials,” the work is part of the United Kingdom government-backed Autodrive project. There are numerous other studies underway to determine the best methods to establish trust between man and machine, some seeking, like JLR’s work, to determine if it is possible to mimic the eye contact and other non-verbal signals pedestrians and human drivers exchange.

JLR is using dedicated autonomous test vehicles—dubbed Eye Pods—fitted with sophisticated vision systems to gather data to assess how much information future self-driving vehicles should share with other road users (but particularly pedestrians) to generate sufficient mutual trust. At present, states JLR, studies indicate that around two-thirds of pedestrians would be concerned about crossing a road populated by AVs.

Communication between a pedestrian and a conventional vehicle driver is a combination of vision (eye contact with the driver), speed and awareness assessment. There is invariably some mutuality about actions taken, or about to be taken, by both parties.

Eye contact. Literally

Giving pedestrians trust that an autonomous vehicle has identified them and will stop has generated an interesting solution from JLR’s research engineers. Red or green lights signal and reassure a pedestrian of the vehicle’s intentions. (JLR).

The test vehicles’ somewhat unnerving Eye Pods were devised by a team of advanced-technology engineers working in JLR’s Future Mobility division. The autonomous and therefore intelligent Pods run on a fabricated city street scene, constantly analyzing pedestrians’ behavior as they wait to cross the road. The company said the research Pod spots pedestrians—and appearing to “look” directly at them, signals via green or red lights that it has identified the pedestrian and simultaneously indicates if it will stop at a pedestrian crossing or take avoiding action as necessary. But it won’t wink, assures JLR.

Pedestrian “trust levels” are constantly monitored before and after the Eye Pod makes overall visual contact to find out whether the vehicle generated sufficient confidence that it would stop for the pedestrian. “Previous studies suggest as many 63% of pedestrians and cyclists say they would feel less safe sharing the road with a self-driving vehicle. Safety remains the number one priority as JLR invests in self-driving technology,” the company said.

The trial is aligned with the brand’s long-term AV-related strategic goals: to make cars safer, free up people’s valuable time, and improve mobility for everyone.

Bennett adds: “It’s second nature to glance at the driver of the approaching vehicle before stepping into the road. Understanding how this translates in tomorrow’s more-automated world is important.”

The trials are part of a wider study exploring how future connected and autonomous vehicles can replicate human behavior and reactions when driving. As part of the work, more than 500 test subjects have been studied interacting with the self-driving Pods designed by U.K. Autodrive partner Aurrigo.

The Autodrive project is a 3-year program that completes toward the end of 2018 and was formed to integrate autonomous and connected vehicles into real-world urban environments. Its targets include showing how autonomous and connected vehicles could solve everyday challenges such as congestion; demonstrating the commercial operation of electric, self-driving “pods” at a city scale and providing insight for key stakeholders and decision-makers, including legislators, insurers and investors.