ABOUT THE SPEAKER
Chris Gerdes - Mechanical engineer
An autonomous car may seem like a thing of the distant future, but mechanical engineer Chris Gerdes is racing to make it a reality today.

Why you should listen

Imagine a car that can drive itself -- that with the push of a button can get you home safely when you’re too tired to drive or have had a night of one too many drinks. Dr. Chris Gerdes , the Director of the Center for Automotive Research at Stanford (conveniently acronymed CARS), and his team are developing a robotic race car, capable of driving at outrageous speeds while avoiding every possible accident. Gerdes’ research focuses on the development of driver assistance systems for collision avoidance, as well as on new combustion processes for engines.

Prior to teaching at Stanford, Gerdes was the project leader for vehicle dynamics at the Vehicle Systems Technology Center of Daimler-Benz Research and Technology North America. His work at Daimler focused on safety analysis.

More profile about the speaker
Chris Gerdes | Speaker | TED.com
TEDxStanford

Chris Gerdes: The future race car -- 150mph, and no driver

Filmed:
806,444 views

Autonomous cars are coming -- and they're going to drive better than you. Chris Gerdes reveals how he and his team are developing robotic race cars that can drive at 150 mph while avoiding every possible accident. And yet, in studying the brainwaves of professional racing drivers, Gerdes says he has gained a new appreciation for the instincts of professional drivers
- Mechanical engineer
An autonomous car may seem like a thing of the distant future, but mechanical engineer Chris Gerdes is racing to make it a reality today. Full bio

Double-click the English transcript below to play the video.

00:16
So, how many of you have ever
0
506
1407
00:17
gotten behind the wheel of a car
1
1913
1655
00:19
when you really shouldn't have been driving?
2
3568
5687
00:25
Maybe you're out on the road for a long day,
3
9255
1905
00:27
and you just wanted to get home.
4
11160
1490
00:28
You were tired, but you felt you could drive a few more miles.
5
12650
2647
00:31
Maybe you thought,
6
15297
1199
00:32
I've had less to drink than everybody else,
7
16496
2017
00:34
I should be the one to go home.
8
18513
1736
00:36
Or maybe your mind was just entirely elsewhere.
9
20249
4591
00:40
Does this sound familiar to you?
10
24840
1454
00:42
Now, in those situations, wouldn't it be great
11
26294
2898
00:45
if there was a button on your dashboard
12
29192
1593
00:46
that you could push, and the car would get you home safely?
13
30785
6343
00:53
Now, that's been the promise of the self-driving car,
14
37128
2293
00:55
the autonomous vehicle, and it's been the dream
15
39421
2627
00:57
since at least 1939, when General Motors showcased
16
42048
3249
01:01
this idea at their Futurama booth at the World's Fair.
17
45297
3302
01:04
Now, it's been one of those dreams
18
48599
1943
01:06
that's always seemed about 20 years in the future.
19
50542
4214
01:10
Now, two weeks ago, that dream took a step forward,
20
54756
2683
01:13
when the state of Nevada granted Google's self-driving car
21
57439
3265
01:16
the very first license for an autonomous vehicle,
22
60704
3600
01:20
clearly establishing that it's legal for them
23
64304
2245
01:22
to test it on the roads in Nevada.
24
66549
1810
01:24
Now, California's considering similar legislation,
25
68359
3727
01:27
and this would make sure that the autonomous car
26
72086
2408
01:30
is not one of those things that has to stay in Vegas.
27
74494
2977
01:33
(Laughter)
28
77471
2096
01:35
Now, in my lab at Stanford, we've been working on
29
79567
3784
01:39
autonomous cars too, but with a slightly different spin
30
83351
3487
01:42
on things. You see, we've been developing robotic race cars,
31
86838
4248
01:46
cars that can actually push themselves to the very limits
32
91086
4120
01:51
of physical performance.
33
95206
2240
01:53
Now, why would we want to do such a thing?
34
97446
2613
01:55
Well, there's two really good reasons for this.
35
100059
2100
01:58
First, we believe that before people turn over control
36
102159
3959
02:02
to an autonomous car, that autonomous car should be
37
106118
2834
02:04
at least as good as the very best human drivers.
38
108952
3254
02:08
Now, if you're like me, and the other 70 percent of the population
39
112206
3305
02:11
who know that we are above-average drivers,
40
115511
2193
02:13
you understand that's a very high bar.
41
117704
3175
02:16
There's another reason as well.
42
120879
2392
02:19
Just like race car drivers can use all of the friction
43
123271
3576
02:22
between the tire and the road,
44
126847
1280
02:24
all of the car's capabilities to go as fast as possible,
45
128127
3177
02:27
we want to use all of those capabilities to avoid
46
131304
3345
02:30
any accident we can.
47
134649
1588
02:32
Now, you may push the car to the limits
48
136237
2050
02:34
not because you're driving too fast,
49
138287
1967
02:36
but because you've hit an icy patch of road,
50
140254
2160
02:38
conditions have changed.
51
142414
1704
02:40
In those situations, we want a car
52
144118
2761
02:42
that is capable enough to avoid any accident
53
146879
3720
02:46
that can physically be avoided.
54
150599
2678
02:49
I must confess, there's kind of a third motivation as well.
55
153277
4267
02:53
You see, I have a passion for racing.
56
157544
2256
02:55
In the past, I've been a race car owner,
57
159800
2764
02:58
a crew chief and a driving coach,
58
162564
2555
03:01
although maybe not at the level that you're currently expecting.
59
165119
3855
03:04
One of the things that we've developed in the lab --
60
168974
2704
03:07
we've developed several vehicles --
61
171678
1704
03:09
is what we believe is the world's first
62
173382
2235
03:11
autonomously drifting car.
63
175617
2365
03:13
It's another one of those categories
64
177982
2513
03:16
where maybe there's not a lot of competition.
65
180495
2423
03:18
(Laughter)
66
182918
1408
03:20
But this is P1. It's an entirely student-built electric vehicle,
67
184326
3822
03:24
which through using its rear-wheel drive
68
188148
2078
03:26
and front-wheel steer-by-wire
69
190226
1565
03:27
can drift around corners.
70
191791
2067
03:29
It can get sideways like a rally car driver,
71
193858
2200
03:31
always able to take the tightest curve,
72
196058
1715
03:33
even on slippery, changing surfaces,
73
197773
3304
03:36
never spinning out.
74
201077
1616
03:38
We've also worked with Volkswagen Oracle,
75
202693
2368
03:40
on Shelley, an autonomous race car that has raced
76
205061
3424
03:44
at 150 miles an hour through the Bonneville Salt Flats,
77
208485
3070
03:47
gone around Thunderhill Raceway Park in the sun,
78
211555
4471
03:51
the wind and the rain,
79
216026
2639
03:54
and navigated the 153 turns and 12.4 miles
80
218665
5018
03:59
of the Pikes Peak Hill Climb route
81
223683
1562
04:01
in Colorado with nobody at the wheel.
82
225245
3473
04:04
(Laughter)
83
228718
1448
04:06
(Applause)
84
230166
5566
04:11
I guess it goes without saying that we've had a lot of fun
85
235732
3279
04:14
doing this.
86
239011
1304
04:16
But in fact, there's something else that we've developed
87
240315
3657
04:19
in the process of developing these autonomous cars.
88
243972
3055
04:22
We have developed a tremendous appreciation
89
247027
3871
04:26
for the capabilities of human race car drivers.
90
250898
3817
04:30
As we've looked at the question of how well do these cars perform,
91
254715
4345
04:34
we wanted to compare them to our human counterparts.
92
259060
3279
04:38
And we discovered their human counterparts are amazing.
93
262339
5680
04:43
Now, we can take a map of a race track,
94
268019
4023
04:47
we can take a mathematical model of a car,
95
272042
2370
04:50
and with some iteration, we can actually find
96
274412
2903
04:53
the fastest way around that track.
97
277315
1625
04:54
We line that up with data that we record
98
278940
2533
04:57
from a professional driver,
99
281473
1433
04:58
and the resemblance is absolutely remarkable.
100
282906
4107
05:02
Yes, there are subtle differences here,
101
287013
3212
05:06
but the human race car driver is able to go out
102
290225
3127
05:09
and drive an amazingly fast line,
103
293352
2335
05:11
without the benefit of an algorithm that compares
104
295687
2330
05:13
the trade-off between going as fast as possible
105
298017
2608
05:16
in this corner, and shaving a little bit of time
106
300625
2037
05:18
off of the straight over here.
107
302662
1902
05:20
Not only that, they're able to do it lap
108
304564
3457
05:23
after lap after lap.
109
308021
2375
05:26
They're able to go out and consistently do this,
110
310396
2912
05:29
pushing the car to the limits every single time.
111
313308
4128
05:33
It's extraordinary to watch.
112
317436
3169
05:36
You put them in a new car,
113
320605
2066
05:38
and after a few laps, they've found the fastest line in that car,
114
322671
3902
05:42
and they're off to the races.
115
326573
3877
05:46
It really makes you think,
116
330450
1146
05:47
we'd love to know what's going on inside their brain.
117
331596
4871
05:52
So as researchers, that's what we decided to find out.
118
336467
4541
05:56
We decided to instrument not only the car,
119
341008
1812
05:58
but also the race car driver,
120
342820
2495
06:01
to try to get a glimpse into what was going on
121
345315
2769
06:03
in their head as they were doing this.
122
348084
2186
06:06
Now, this is Dr. Lene Harbott applying electrodes
123
350270
3950
06:10
to the head of John Morton.
124
354220
1232
06:11
John Morton is a former Can-Am and IMSA driver,
125
355452
2989
06:14
who's also a class champion at Le Mans.
126
358441
1800
06:16
Fantastic driver, and very willing to put up with graduate students
127
360241
3496
06:19
and this sort of research.
128
363737
1855
06:21
She's putting electrodes on his head
129
365592
2672
06:24
so that we can monitor the electrical activity
130
368264
2112
06:26
in John's brain as he races around the track.
131
370376
2832
06:29
Now, clearly we're not going to put a couple of electrodes on his head
132
373208
3195
06:32
and understand exactly what all of his thoughts are on the track.
133
376403
3270
06:35
However, neuroscientists have identified certain patterns
134
379673
3407
06:38
that let us tease out some very important aspects of this.
135
383080
3761
06:42
For instance, the resting brain
136
386841
1847
06:44
tends to generate a lot of alpha waves.
137
388688
2155
06:46
In contrast, theta waves are associated with
138
390843
3752
06:50
a lot of cognitive activity, like visual processing,
139
394595
3184
06:53
things where the driver is thinking quite a bit.
140
397779
3048
06:56
Now, we can measure this,
141
400827
1663
06:58
and we can look at the relative power
142
402490
1985
07:00
between the theta waves and the alpha waves.
143
404475
2200
07:02
This gives us a measure of mental workload,
144
406675
2442
07:05
how much the driver is actually challenged cognitively
145
409117
3567
07:08
at any point along the track.
146
412684
1786
07:10
Now, we wanted to see if we could actually record this
147
414470
2942
07:13
on the track, so we headed down south to Laguna Seca.
148
417412
3038
07:16
Laguna Seca is a legendary raceway
149
420450
2016
07:18
about halfway between Salinas and Monterey.
150
422466
2301
07:20
It has a curve there called the Corkscrew.
151
424767
2087
07:22
Now, the Corkscrew is a chicane, followed by a quick
152
426854
2717
07:25
right-handed turn as the road drops three stories.
153
429571
2746
07:28
Now, the strategy for driving this as explained to me was,
154
432317
3766
07:31
you aim for the bush in the distance,
155
436083
2022
07:34
and as the road falls away, you realize it was actually the top of a tree.
156
438105
3025
07:37
All right, so thanks to the Revs Program at Stanford,
157
441130
3304
07:40
we were able to take John there
158
444434
1473
07:41
and put him behind the wheel
159
445907
964
07:42
of a 1960 Porsche Abarth Carrera.
160
446871
2439
07:45
Life is way too short for boring cars.
161
449310
3698
07:48
So, here you see John on the track,
162
453008
1759
07:50
he's going up the hill -- Oh! Somebody liked that --
163
454767
2184
07:52
and you can see, actually, his mental workload
164
456951
2465
07:55
-- measuring here in the red bar --
165
459416
2153
07:57
you can see his actions as he approaches.
166
461569
2343
07:59
Now watch, he has to downshift.
167
463912
3230
08:03
And then he has to turn left.
168
467142
761
08:03
Look for the tree, and down.
169
467903
3993
08:07
Not surprisingly, you can see this is a pretty challenging task.
170
471896
2838
08:10
You can see his mental workload spike as he goes through this,
171
474734
2976
08:13
as you would expect with something that requires
172
477710
2064
08:15
this level of complexity.
173
479774
2809
08:18
But what's really interesting is to look at areas of the track
174
482583
3416
08:21
where his mental workload doesn't increase.
175
485999
2846
08:24
I'm going to take you around now
176
488845
1495
08:26
to the other side of the track.
177
490340
1089
08:27
Turn three. And John's going to go into that corner
178
491429
2336
08:29
and the rear end of the car is going to begin to slide out.
179
493765
2551
08:32
He's going to have to correct for that with steering.
180
496316
2017
08:34
So watch as John does this here.
181
498333
2231
08:36
Watch the mental workload, and watch the steering.
182
500564
2322
08:38
The car begins to slide out, dramatic maneuver to correct it,
183
502886
3672
08:42
and no change whatsoever in the mental workload.
184
506558
3523
08:45
Not a challenging task.
185
510081
2832
08:48
In fact, entirely reflexive.
186
512913
3200
08:52
Now, our data processing on this is still preliminary,
187
516113
3643
08:55
but it really seems that these phenomenal feats
188
519756
2672
08:58
that the race car drivers are performing
189
522428
1610
08:59
are instinctive.
190
524038
1910
09:01
They are things that they have simply learned to do.
191
525948
3390
09:05
It requires very little mental workload
192
529338
2282
09:07
for them to perform these amazing feats.
193
531620
2824
09:10
And their actions are fantastic.
194
534444
3135
09:13
This is exactly what you want to do on the steering wheel
195
537579
2611
09:16
to catch the car in this situation.
196
540190
3337
09:19
Now, this has given us tremendous insight
197
543527
3445
09:22
and inspiration for our own autonomous vehicles.
198
546972
3122
09:25
We've started to ask the question:
199
550094
1928
09:27
Can we make them a little less algorithmic
200
552022
2253
09:30
and a little more intuitive?
201
554275
2449
09:32
Can we take this reflexive action
202
556724
2281
09:34
that we see from the very best race car drivers,
203
559005
2287
09:37
introduce it to our cars,
204
561292
1649
09:38
and maybe even into a system that could
205
562941
1984
09:40
get onto your car in the future?
206
564925
1968
09:42
That would take us a long step
207
566893
1611
09:44
along the road to autonomous vehicles
208
568504
2509
09:46
that drive as well as the best humans.
209
571013
1912
09:48
But it's made us think a little bit more deeply as well.
210
572925
3440
09:52
Do we want something more from our car
211
576365
2968
09:55
than to simply be a chauffeur?
212
579333
1840
09:57
Do we want our car to perhaps be a partner, a coach,
213
581173
4235
10:01
someone that can use their understanding of the situation
214
585408
3087
10:04
to help us reach our potential?
215
588495
4256
10:08
Can, in fact, the technology not simply replace humans,
216
592751
2273
10:10
but allow us to reach the level of reflex and intuition
217
595024
4575
10:15
that we're all capable of?
218
599599
3425
10:18
So, as we move forward into this technological future,
219
603024
1923
10:20
I want you to just pause and think of that for a moment.
220
604947
2821
10:23
What is the ideal balance of human and machine?
221
607768
3775
10:27
And as we think about that,
222
611543
1709
10:29
let's take inspiration
223
613252
1731
10:30
from the absolutely amazing capabilities
224
614983
3329
10:34
of the human body and the human mind.
225
618312
2816
10:37
Thank you.
226
621128
1736
10:38
(Applause)
227
622864
4604
Translated by Morton Bast
Reviewed by Thu-Huong Ha

▲Back to top

ABOUT THE SPEAKER
Chris Gerdes - Mechanical engineer
An autonomous car may seem like a thing of the distant future, but mechanical engineer Chris Gerdes is racing to make it a reality today.

Why you should listen

Imagine a car that can drive itself -- that with the push of a button can get you home safely when you’re too tired to drive or have had a night of one too many drinks. Dr. Chris Gerdes , the Director of the Center for Automotive Research at Stanford (conveniently acronymed CARS), and his team are developing a robotic race car, capable of driving at outrageous speeds while avoiding every possible accident. Gerdes’ research focuses on the development of driver assistance systems for collision avoidance, as well as on new combustion processes for engines.

Prior to teaching at Stanford, Gerdes was the project leader for vehicle dynamics at the Vehicle Systems Technology Center of Daimler-Benz Research and Technology North America. His work at Daimler focused on safety analysis.

More profile about the speaker
Chris Gerdes | Speaker | TED.com