Self-driv­ing cars vul­ner­a­ble to cy­ber­at­tack, ex­perts warn


Hack­ers pose a real dan­ger to self-driv­ing ve­hi­cles, U.S. ex­perts are warn­ing, and car­mak­ers and in­sur­ers are start­ing to fac­tor in the risk.

Ex­pected on the road by 2020 or even sooner, driver­less cars should have a wide range of cut­ting-edge tech­nolo­gies such as elec­tronic sen­sors — a group of cam­eras, radar, sonar and Li­DAR ( light de­tec­tion and rang­ing) — com­manded re­motely us­ing soft­ware that senses road widths, iden­ti­fies signs and even road­blocks.

But like con­nected ve­hi­cles and their on­board mul­ti­me­dia sys­tems, th­ese new self-driv­ing tech­nol­ogy el­e­ments — which were meant to make the cars safe and re­li­able, could end up leav­ing them vul­ner­a­ble to hacker strikes, ac­cord­ing to U.S. se­cu­rity firms Mission Se­cure Inc (MSi) and Per­rone Ro­bot­ics Inc.

A hacker re­cently boasted of hav­ing en­tered the elec­tronic sys­tems of the U.S. jet he was trav­el­ing on, and of hav­ing changed its tra­jec­tory. He claimed he did so us­ing the in-flight Wi-Fi sys­tem.

The two se­cu­rity com­pa­nies, work­ing with the Uni­ver­sity of Vir­ginia and the Pen­tagon, have run tests that have shown they be­lieve it is pos­si­ble to hack into and dis­rupt the multi-sen­sor sys­tem.

One trial was to change how the car re­sponded when it en­coun­tered an ob­sta­cle.

“One attack sce­nario forces the car to ac­cel­er­ate, rather than brake, even though the ob­sta­cle avoid­ance sys­tem (us­ing Li­DAR) de­tects an ob­ject in front of the car. Rather than slow­ing down, the car hits the ob­ject ... at high speed, caus­ing dam­age to the car and po­ten­tial threat to the life and safety of the pas­sen­gers in the car un­der attack and in the car be­ing struck,” ac­cord­ing to the re­port avail­able on MSi’s web­site.

“If an attack were car­ried out suc­cess­fully, au­to­mo­bile man­u­fac­tur­ers have no means of quickly gath­er­ing in­for­ma­tion for foren­sic anal­y­sis or to rapidly deploy ad­di­tional pro­tec­tions to cars in re­sponse to new and evolv­ing at­tacks,” the re­port warns.

Ac­cord­ing to th­ese ex­perts, hack­ers pen­e­trate the sys­tem through wire­less con­nec­tions.

MSi and Per­rone Ro­bot­ics, which are work­ing on a sys­tem to counter cy­ber at­tacks, be­lieve the sit­u­a­tion poses “sig­nif­i­cant chal­lenges and risks for the au­to­mo­tive in­dus­try, as well as to

public safety.”

In­sur­ance pre­mium re­views?

Most of the car­mak­ers gear­ing up their own au­ton­o­mous car projects did not re­ply to up­date re­quests from AFP.

But sources close to the in­dus­try say the chance of the sys­tem be­ing hacked has been fac­tored in through­out the man­u­fac­tur­ing process.

In­ter­net gi­ant Google, for ex­am­ple, is be­lieved to have a team of top pro­gram­mers tasked with try­ing to hack into the soft­ware in their own self-driv­ing pro­to­type car, which is ex­pected to get on­road testing within the next few months. Google de­clined to com­ment for this story.

U. S. in­sur­ers are con­cerned about safety, and whether the new tech­nolo­gies can cut the risk of ac­ci­dents hap­pen­ing.

This could force them to re­think their con­tracts and to re­cal­cu­late pre­mi­ums.

At first, pre­mi­ums could rise be­cause the price of self-driv­ing cars will be high due to the cost of em­bed­ded tech­nolo­gies and re­pairs, in­surer Na­tion­wide told AFP.

But this could be par­tially off­set by the wider use of ve­hi­cles decked out with ac­ci­dent-pre­vent­ing tech­nolo­gies.

For State Farm, an­other U.S. in­surer, the big pic­ture is what counts.

“As con­nected and au­to­mated ve­hi­cle tech­nol­ogy re­duces or elim­i­nates some risks that driv­ers face to­day, new risks are likely to emerge. We are fo­cused on the big pic­ture — how can we adapt to th­ese changes and con­tinue to de­liver value to our cus­tomers,” the com­pany said in an email to AFP.

Newspapers in English

Newspapers from Taiwan

© PressReader. All rights reserved.