Au­ton­o­mous ve­hi­cle to be tested on area roads part of a global rev­o­lu­tion

The Woolwich Observer - - COMMENT -

ONE DAY, TIME SPENT in the car will be spent catch­ing up on our favourite TV shows, read­ing a book or even nap­ping. That day isn’t to­day, but we’re get­ting there. When it ar­rives, it will bring with it ma­jor so­ci­etal changes.

Tech­no­log­i­cal de­vel­op­ment is the key, with a global re­search ef­fort un­der­way. An­other step on the road to that fu­ture – good, bad or in­dif­fer­ent, de­pend­ing on your take – will be vis­i­ble close at hand, as the Water­loo Cen­tre for Au­to­mo­tive Re­search (WatCAR) at the Univer­sity of Water­loo rolls out on-road test­ing of its au­ton­o­mous ve­hi­cle, a Lin­coln MKZ hy­brid sedan nick­named Au­tonomoose.

The univer­sity got the go-ahead from the prov­ince last fall to test the technology on pub­lic roads. Now, re­searchers are ready to do just that.

It’s an ex­cit­ing time for Krzysztof Czar­necki, a lead re­searcher and pro­fes­sor in Water­loo’s Depart­ment of Elec­tri­cal and Com­puter En­gi­neer­ing, and his team at the Water­loo In­tel­li­gent Sys­tems En­gi­neer­ing Lab (WISE Lab).

“It’s going to be a huge mile­stone for our project,” he says of the on-road tests set for mid-Oc­to­ber.

His own area of re­search cen­ters on de­ci­sion mak­ing sys­tems – for in­stance, the abil­ity of a self-driv­ing car to iden­tify and pre­dict pedes­trian be­hav­iour. There’s the hard­ware in­volved – the cam­eras and sen­sors – and then the soft­ware needed to in­ter­pret the data and act on it. That kind of ar­ti­fi­cial in­tel­li­gence is the most dif­fi­cult part of de­vel­op­ing self-driv­ing cars that can in­ter­pret an en­tire scene out on the road and re­act ac­cord­ingly.

Where hu­man driv­ers can in­ter­act with pedes­tri­ans by eye con­tact and body lan­guage, an au­ton­o­mous ve­hi­cle will have to learn to deal with un­pre­dictable be­hav­iour, he notes.

Safety is a key fac­tor, call­ing for ro­bust and re­dun­dant sys­tems with built-in fail­safes.

Ad­vances will be in­cre­men­tal, just as they’ve been with the in­tro­duc­tion of technology into our cars thus far.

En­gi­neers rate au­ton­o­mous ve­hi­cles on a scale of 0-5, 0 be­ing the tra­di­tional cars that re­quired the driver for all in­puts. We’ve seen lev­els 1 and 2, which range from power-as­sisted func­tions through to cruise con­trol and lane-cen­ter­ing technology. The au­ton­o­mous la­bel doesn’t re­ally ap­ply un­til level 3, which is the next step now.

At level 3, the ve­hi­cle be­comes ca­pa­ble of crit­i­cal func­tions such as nav­i­ga­tion, ac­cel­er­at­ing and brak­ing, but all with close mon­i­tor­ing by the driver – no nap­ping just yet.

Czar­necki notes there are no level 3 cars in pro­duc­tion, but mak­ers such as Tesla and Audi say they’ll be of­fer­ing up such ve­hi­cles in the next year or so.

“That’s where you can take your hands off the steer­ing wheel, but you have to pay at­ten­tion,” he says. “You have to be ready to take over at any time.

“We might be able to see level 3 per­haps as soon as next year.”

Level 4 sees cars be­come gen­er­ally au­ton­o­mous – read­ing a book or even a nap are pos­si­ble, but un­der lim­ited cir­cum­stance, such as typ­i­cal com­mutes.

For all kinds of roads – say the dirt and gravel roads of the ru­ral ar­eas – and con­di­tions, we’ll have to wait for Level 5.

Such ad­vances are five to 10 years out, Czar­necki pre­dicts.

For now, re­searchers are fo­cus­ing on get­ting the technology to the point where it can deal with the mul­ti­ple chal­lenges of chang­ing con­di­tions. In rolling out the Au­tonomoose next month, UW teams will start on quiet roads be­fore ramp­ing up to busier streets and win­ter con­di­tions, mak­ing it more dif­fi­cult for the car.

“The main chal­lenges are re­ally there: win­ter driv­ing, night-time driv­ing – mak­ing sure the car does the right thing in all con­di­tions.”

The hard­ware and soft­ware chal­lenges, par­tic­u­larly the ar­ti­fi­cial in­tel­li­gence, are so com­plex that small steps, tweaks along the way, are the norm.

Czar­necki points to Google’s tests of au­ton­o­mous ve­hi­cles, where dis­en­gage­ment episodes – when the hu­man driver is asked to take over – oc­cur about once ev­ery 5,000 miles, up from about ev­ery 2,500 miles a year ago.

Un­like hu­man driv­ers, who all have to learn the skill in­di­vid­u­ally when they come of age, au­ton­o­mous ve­hi­cles can ac­tu­ally “learn” from the vast amounts of data gath­ered by re­searchers around the world. That al­lows for ad­vances on many fronts.

While the safety of the technology is a fac­tor in both its clear­ance for gen­eral road use and its ac­cep­tance by the pub­lic, driver safety is also a ma­jor, well, driver of au­ton­o­mous ve­hi­cles: de­vel­oped as ex­pected, such cars would lead to fewer ac­ci­dents.

“In my mind, there’s no doubt that a ro­botic ve­hi­cle will be safer,” says Czar­necki, not­ing com­put­ers don’t get tired or an­gry or inat­ten­tive.

That’s a big deal, as some 95 per cent of ac­ci­dents are due to hu­man er­ror – “Many of those can be pre­vented.”

Get­ting to that point, how­ever, will come with im­pacts well be­yond safer, nap-filled com­mutes.

Au­ton­o­mous ve­hi­cles have the po­ten­tial to elim­i­nate all driv­ing jobs, from cab­bies to tran­sit work­ers. Let’s look at the truck­ing in­dus­try, one of the largest em­ploy­ers and a place where there are jobs for those with­out higher ed­u­ca­tion. Ac­cord­ing to the Amer­i­can Trucker As­so­ci­a­tion, there are 3.5 mil­lion pro­fes­sional truck driv­ers in the U.S. – it’s the most com­mon job in more than half the states. Add to that an­other 5.2 mil­lion in jobs within the truck­ing in­dus­try. On top of that, there are mil­lions more in re­lated jobs, from lo­gis­tics right on through to the din­ers and mo­tels that cater to driv­ers.

Now, imag­ine all of the jobs dis­ap­pear­ing as a re­sult of driver­less trucks.

The technology al­ready ex­ists to­day, just as it does

with driver­less cars, an­other use of technology des­tined to dis­place jobs such as cab­bies and couri­ers. Driver­less buses and trains will elim­i­nate the need for tran­sit work­ers, many of them an in­creas­ing bur­den on gov­ern­ments and tax­pay­ers.

Many in­dus­tries are likely to un­dergo mas­sive changes, as the technology is likely to be elec­tric, elim­i­nat­ing much of the petroleum-based busi­ness for ve­hi­cles. More­over, we may not even own in­di­vid­ual cars, in­stead adopt­ing a sys­tem based on a smart­phone app sum­mon- ing a ve­hi­cle as needed. (Our cars cur­rently spend about 90 per cent of their time parked, un­used). That would dra­mat­i­cally al­ter the auto in­dus­try, along with the en­tire ser­vice busi­ness and the likes of in­sur­ance com­pa­nies.

Fewer cars trav­el­ling more ef­fi­ciently would change the face of our cities, re­duc­ing the de­mand for park­ing spa­ces, for in­stance, and spark­ing a re­use of land now ded­i­cated to cars.

In short, the world would look much dif­fer­ent.

Whether the pros, of which there are many, out­weigh the (tem­po­rary) cons re­mains to be seen.

Newspapers in English

Newspapers from Canada

© PressReader. All rights reserved.