assimilating self-driving cars into society: travel impacts & transport policy choices

ASSIMILATING SELF‐DRIVING CARSINTO SOCIETY:

Travel Impacts & Transport Policy Choices

Panelists Kara Kockelman & Lisa Loftus‐Otway, UT Austin

Self-driving cars are game changers.

Revolution or Evolution....Where do you think we are?

A lot can happen in 13 years…

1900: 5th Ave NYC Easter Parade

Spot the car!

1913: 5th Ave NYC Easter Parade

Source: George Grantham Bain Collection

Spot the horse!

Introduction

“You can count on one hand the number of years it will take before ordinary people can experience (AVs).”–Sergey Brin, at the 2012 signing of California’s SB 1298.

Opportunities for CAVs U.S. Safety In 2014, 6.0 million crashes in the U.S. resulting in 32,675 deaths & 

>$500+ billion in comprehensive costs. Driver error is primary cause of > 90% of U.S. crashes. 40% of fatal crashes involve alcohol, drugs, fatigue &/or distraction.  AVs can dramatically impact safety by reducing human errors.

U.S. Congestion 7 billion hours of delay & $160 billion  losses in 2014. Reductions possible via traffic smoothing, tighter headways, 

cooperative adaptive cruise control (CACC) & fewer crashes.

Opportunities (2)

Traveler Behaviors Car‐sharing & ride‐sharing Increased mobility for 

elderly, disabled, & children? Parking benefits Latent & induced VMT

Freight Movement Reduced labor & thus shipping costs Improved fuel economies from tight‐headway drafting

0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

Midnight

3:00

 AM

6:00

 AM

9:00

 AM

Noo

n

3:00

 PM

6:00

 PM

9:00

 PM

Midnight

Vehs ≤ 10 yrs

Vehs ≤ 15 yrs

All vehs

Part 1

What Are AVs Worth?- to Society & to Individual Owners

Annual Per-AV Economic ImpactsAssumed Market Share

10% 50% 90%CrashesLives Saved 1,100 9,600 21,700Economic Savings per AV per year $430 $770 $960Comprehen. Savings per AV per yr $1,390 $2,480 $3,100

CongestionTravel Time Savings (M Hours) 756 1,680 2,770Fuel Savings (M Gallons) 102 224 724Savings per AV per year $1,320  $590  $550 

Other ImpactsParking Savings per AV per year $250  $250  $250 VMT Increase 2.0% 7.5% 9.0%Change in Total # Vehicles ‐4.7% ‐23.7% ‐42.6%

(U.S. Context, $2012)

Totaled Social Benefits

Assumed Market Share

10% 50% 90%Annual U.S. Savings: Economic Benefits Only $26 B $102 B $201 BAnnual U.S. Savings: Comprehensive Benefits $38 B $211 B $447 BSavings Per AV per year: Econ. Benefits Only $2,000  $1,610  $1,760 Savings Per AV per year: Comprehen. Benefits $2,960  $3,320  $3,900 Net Present Value (NPV) of AV Benefits minus Purchase Price (Econ. Benefits Only) $5,200 $7,250 $10,400Net Present Value of AV Benefits minus Purchase Price (Comprehensive Benefits) $12,500

$20,300 per AV sold $26,700

Added Purchase Price $10,000 $5,000 $3,000

U.S. Industry Impacts, at 100% AdoptionIndustry Industry Size

($B/yr)Industry Impact

($B/yr)% Change in

Industry$ per Capita per Year

Insurance $180B/yr $108B/yr 60% $339/person/yrFreight Transportation $604 $100 17% $313

Land Development $931 $45 5% $142Automotive $570 $42 7% $132

Personal Transport $86 $27 31% $83Electronics & Software

Technology$203 $26 13% $83

Auto Repair $58 $21 36% $66

Digital Media $42 $14 33% $44

Medical $2,700 $12 0% $36

Oil and Gas $284 $10 4% $31

Construction/Infrastructure $169 $8 4% $24

Traffic Police $10 $5 50% $16

Law $277 $3 1% $10

Industry-based Totals $6,113 $420B/year 7% $1,318/person/year

Adding in Additional Effects:Travel Time “Productivity” Rises

+ Pain & Suffering from Crashes Fall

Economy-Wide (non-Industry-based) Effects

Economic Impact ($B/yr) $ per Capita per Year

Productivity enRoute

$645 Billion/year $2,022/person-year

Pain & Suffering + other Crash Costs

$488 B/year $1,530

Additional Effects $1,133 B/year $3,552/person-year

Overall Totals (industry + other)

$1.4 Trillion per year! $4,419 per person-year

Part 2

Forecasting Americans’ Long-Term Adoption of Connected & Autonomous

Vehicle (C/AV) Technologies

Willingness to Pay (WTP)Average WTP

Average WTP(if WTP > 0)

% of Respondents with $0 WTP

Electronic Stability Control $52 $79 33.4%Lane Centering $205 $352 41.7%Left Turn Assist $119 $221 46.1%Cross Traffic Sensor $169 $252 32.8%Adaptive Headlight  $203 $345 41.1%Pedestrian Detect  $145 $232 37.5%Adaptive Cruise Control $126 $202 37.7%Blind Spot Monitoring $160 $210 23.7%Traffic Sign Recognition $93 $204 54.4%Emergency Automatic Braking  $183 $257 28.7%Level 3 Automation $2,438 $5,470 55.4%Self‐parking Valet System $436 $902 51.7%Level 4 Automation $5,857 $14,196 58.7%Connectivity (DSRC) $67 $111 39.1%

Simulating Fleet EvolutionVehicle inventoryDemographicsTravel Patterns

Technology evolution

Transactiondecision model

(multinomial logit)

Add technologiesto old vehicles

Sell a vehicleand buy vehicles

Buy vehicles

Sell a vehicle

Add connectivityif WTP ≥ Price

Buy new or used? (Logit)

LV4 WTP ≥ Price

LV3 WTP ≥ Price

Dispose ofthe oldest vehicle

Add connectivityif WTP ≥ Price

Vehicleis already

LV3 or LV4End: Do nothing

End: Dispose ofthe oldest vehicle

New

Used

End:Add LV4End: Add LV1,

LV2, or self-parking valetif WTP ≥ Price

No

End:Add LV3

No

No

Yes

Yes

Yes

Same process for each household, every year.

Predicted Shares of US Light-duty Vehicles

59.5%

83.5%100.0% 100.0%100.0% 100.0%100.0%

0.0%

100.0%

0%

25%

50%

75%

100%

2015 2020 2025 2030 2035 2040 2045

DSRC‐based Connectivity

43.0%

43.8%

24.8%

43.4%43.2%

70.7%59.7%

0.0%

87.2%

0%

25%

50%

75%

100%

2015 2020 2025 2030 2035 2040 2045

Level 4 Automation

Part 3Agent-Based Models for Shared AVs

+ = &

Less than 20% of newer (& 15% of all) personal vehicles are in‐use at peak times, even with 5‐minute pickup & drop‐off buffers.

Car‐sharing programs like ZipCar &  Car2go have expanded quickly, with the number of U.S. users doubling every  year or two, over the past decade.

Shared Autonomous Vehicles (SAVs) can help overcome car‐sharing barriers, like return‐trip certainty &  vehicle access distances.

Agent-Based Model Framework Grid‐based 10 mi x 10 mi urban area with 0.25‐sq. mile zones. Trip generation:

Poisson‐based PK & OP counts for trip generation, every 5 minutes. Higher trip production & attraction rates closer to city center. Mostly round‐trip travel, with 78% travelers returning via SAVs. Random departure times & trip distances (2009 NHTS).

SAVs travel at fixed speeds, with 5 min. intervals.

0.0%0.5%1.0%1.5%2.0%2.5%3.0%3.5%4.0%4.5%5.0%

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

00.10.20.30.40.50.60.70.80.91

0 1 2 3 4 5 6 7 8 9 101112131415

Midnight ‐ 3 AM

3 AM ‐ 6 AM

6 AM ‐ 9 AM

9 AM ‐ Noon

Noon ‐ 3 PM

3 PM ‐ 6 PM

6 PM ‐ 9 PM

9 PM ‐ Midnight

Trip Generation Trip Distances (mi.)Dwell Times (hrs.)

Example: One SAV’s 24-hour Journey

Higher AM Trip Attraction

Higher PM Trip Attraction

• Red ArrowsSAV Relocation

• Blue ArrowsServing Riders

5 mi

10 mi

SE

Case Study Results 100 days were simulated to assess SAV travel implications.

Parameter ValueService area 10 mi. x 10 mi.Outer trip generation rate 9 trips/cell/dayCBD edge trip generation rate 27 trips/cell/dayCBD core trip generation rate 30 trips/cell/dayOff‐peak speed 33 mphPeak speed 21 mphAM peak 7 AM ‐ 8 AMPM peak 4 PM ‐ 6:30 PMTrip share returning by SAV 78%

Scenario Results Each SAV replaced 9 to 13 

conventional vehicles. Avg. wait time ≈ 2.8 min. 11% new/induced (empty‐

vehicle) travel. Yet 5% to 50% (GHG vs. VOCs) 

life‐cycle emissions reductions, thanks to smaller vehicles,fewer cold starts, & less parking infrastructure!

Part 4

What if SAVs Serve Central Austin, & Offer Dynamic Ride-Sharing (DRS)?

One SAV’s24-hour day

Dropoff

Pickup

Travel

Case Study Results

1:10 & 1:8 veh. replacement rates (with & w/o DRS)

System pays for itself with just $1/mile fares!

Electric vehicles (Leaf & Model S) also tested (with inductive charging), using 100 mi x 100 mi region.

DRS saves more emissions ‐& VMT even falls (vs. BAU).

Measure With DRS Without DRSSAV fleet size 1,855 2,181Veh. replacement rate 9.95 8.47Average wait time 57 sec 47 sec% Waiting > 10 min. 0.60% 0.33%5‐6 PM avg. wait 3.0 min 2.4 minAvg. total trip time 14.4 min 13.8 minNew VMT introduced 4.90% 7.92%# rides shared 5,754 0% VMT shared 4.50% 0%

24‐hour days simulated with 56,300 to 270,000 trips served. Excellent Level of Service (typ. wait time < 3 min.)

Part 5

What if SAVs Serve the Entire Region? And Are SAElectricVs?

Shared Autonomous Electric VehiclesSAEVs

or

Station Generation via 30-day Initial Run

Check for unmet requests

Find closest SAEV

SAEV has range to meet trip request?

Unmet request Send SAEV to 

serve tripYes

t=t+1sec

Next timestep: t=t+1sec

No new requests

None available

Is a charging station also in 

range?

No

Create new station at vehicle’s location

No

SAEV begins charging

SAEV heads to closest station

SAEV removed from 

consideration

Try again, in next timestep Yes

Charging Station Locations

Charging stations generated based on demand.  Number of charging stations formed is dependent only on vehicle range. 

Stations formed for 200‐mile range (left) & 60‐mile range (right) 

Central Austin Station Locations

Assuming 60 mile range + 4 hour charge time + 5:1 travelers per SAEV  (= 28 stations over 6 x 10 mi area)

Austin SAEV ResultsScenario Gas 

SAV

Short‐Range SAEV

Long‐Range Fast Charge

Long‐Range SAEV

Short‐Range but Fast Charge

Long‐Range,Fast Charge, Smaller Fleet

Range (mi) 525 60 200 200 60 200Recharge/Refuel Time (min) 2 240  30 240 30 30# of Charging/Gas Stations 19 155 155 155 155 155

Fleet Size (# vehicles) 5,893 5,893 5,893 5,893 5,893 4,124Avg. Daily miles per Vehicle 452 201 354 441 355 501

% of Unserved Trips 1.62 60.6 19.4 2.67 16.2 15.2Avg. Daily Trips per Vehicle 28.5 11.4 23.4 28.2 24.3 35.1

Avg.  Wait Time Per Trip (min) 4.45 9.82 8.76 5.49 6.16 9.55% Unoccupied Travel 6.05 13.1 7.88 6.86 14.2 8.62% Travel for Charging 0.65 5.59 1.26 1.05 5.34 1.27

• Fleet size is key to lower response times. Tripling fleet size (from 9:1 to 3:1 travelers per SAEV) lowers average response times by >75%.

• Longer charge times increase response times (& unserved trips rise 19% to 61%)• Longer ranges lower empty VMT, but fast‐charging improves response times.• Trips in Austin’s urban core are served best (e.g., never exceed 30‐min wait times).

SAEV Cost Assumptions

• Conventional BEV Costs: $25,000 (short range) to $35,000 (long‐range)• Self‐driving Technology Cost: $5,000 to $25,000 per vehicle• Battery Replacement: $100 ‐ $190 per kWh (once per vehicle life)

• Vehicle Maintenance: 5.4¢ to ‐6.6¢ per mile

• Insurance & Registration: $550 ‐ $2,200 per vehicle‐year• Electricity: 8¢ to 20¢ per kWh

• Level II Chargers: $8,000 ‐ $18,000 each• Level II Charger Maintenance: $25 ‐ $50 per year, per charger

• Fast (Level III) Charger: $10,000 ‐ $100,000 per charger• Fast Charger Maintenance: $1,000 ‐ $2,000 per year, per charger

• Station Properties: $1,980 to $6,900 per vehicle space (based on location)

Financial Results: Costs per Mile

Mid‐RangeExpected Costs

per mile

Gasoline SAV

Short‐Range SAEV

Long‐Range SAEV

Fast‐Charge, Long‐RangeSAEV

Fast‐Charge SAEV

Fast‐Charge, Long‐Range Reduced Fleet

Electricity/Fuel 6.39¢/mi 4.51 4.26 4.21 4.57 4.29Vehicle Maint., Admin +Attendants

18.4¢/mi 19.7 18.6 18.4 19.9 18.7

Charger Costs (Land + Infrastructure) n/a 3.57 1.35 2.15 6.30 0.76

Vehicle Purchase  19.6¢/mi 27.7 29.4 28.3 25.3 28.4Battery Costs n/a 1.58 4.91 4.85 1.60 4.95

Total Costs per Mile 45¢/mi 59¢/mi 59¢/mi 59¢/mi 59¢/mi 59¢/miDaily Vehicle Profit ($1/mile fare)

$234/veh‐day $72 $132 $170 $126 $187

#Trips/vehicle‐day 28 trips/veh‐day 11 23 28 24 35

Response time/trip 4.4min 9.8 8.8 5.5 6.2 9.6

Part 6How Should We Modify our Travel

Demand Models & Plan for the Future?

More Complete Model Assumptions Vehicle ownership changes over time (AVs cost more & SAVs 

allow people to avoid ownership). Travel times are less burdensome (for drivers)

Lower values of travel time thanks to more productive (& restful!) in‐vehicle activities

Affect trip mode choice probabilities Travel costs may fall

AVs can head to lower‐cost parking locations Shared AVs reduce overall vehicle‐use costs Dynamic ride‐sharing reduces per‐trip costs even further

Link capacities rise on roadways V2V communications (e.g., CACC) + smart intersections (long term) AVs may eventually follow at shorter headways & distances Hopefully lower travel times & travel‐time unreliability…

However, we also expect… Longer travel distances (more distant destinations). More trip‐making by those presently unlicensed, with 

disabilities &/or other difficulties driving. Less air travel by passengers & rail travel by freight. Possibly larger, less‐efficient vehicles, for longer‐distance 

trips, & more land use sprawl.This means… Rising congestion & infrastructure damage in many locations. Need for smarter system management, including incentives 

for ride‐sharing & non‐motorized travel, route guidance, credit‐based congestion pricing & micro‐tolling ‐ to internalize externalities & operate more efficiently, equitably, & sustainably!

In Conclusion… CAVs offer tremendous benefits for mobility, safety & parking, 

but will add VMT & congestion. SAVs offer a new & exciting (transit?) mode, with each SAV 

replacing ~8 personal vehicles, for same level of motorized trip‐making.

SAVs add 7‐10% extra VMT (though DRS may reduce VMT). Yet SAVs may bring useful travel‐cost savings, emissions benefits

+ profits for transit providers. Traditional travel models cannot capture the details of SAV 

systems & CAV operations. Microsimulation is needed. Smart system management practices are also needed, to avoid 

gridlock, sprawl, greater energy use, & other downsides.

So, what is going on in ‘the law’ around the world?

United States: Federal• National Highway Traffic Safety Administration (NHTSA) 

preliminary policy on Automated Vehicles in 2013– Outlined definitions for Levels 0 through 4 of automation

• NHTSA in September 2016 issued new Policy on Autonomous Vehicles – Adopted SAE J3016 definitions (L0 through Level 5) as their standard.

• Deliberately issued as policy & not regulations, with goal to set stage for consistent national framework but providing flexibility to states.

NHTSA 2016 PolicyDelineated roles/responsibilities for state(s) policy:‐ States retain their traditional responsibilities for vehicle licensing, registration, traffic laws & enforcement, and motor vehicle insurance & liability.

‐ NHTSA continued preemption for interpretations, exemptions, notice, and rulemaking & enforcement authority.

Manufacturer responsibility to determine their system conforms with SAE J3016.

NHTSA’s Framework forVehicle Performance Guidance

U.S.: Federal (2)• NHTSA October 2016: Policy on Cyber Security in 

Autonomous Vehicles– Covers all vehicles not just HAVs & applies to designers, supplies, manufactures & modifiers

• FAST Act 2015, §24302 limitations on data retrieved from Event Data Recorders (EDRs)– NHTSA required to determine amount of time EDRs should capture & record data for retrieval ... to provide sufficient information to investigate a motor vehicle crash

U.S.: State‐level• Over 80 bills are currently in front of U.S. state legislatures on 

this topic.• Nevada created legislation allowing testing in 2011.• California legislation authorized a pilot program in 2014.• Michigan: 2013 allowed testing of automated vehicles as long 

as human was in car.• Michigan: 2016 allows driverless cars to be driven for any of the 

following purposes, no human required to be in car:– Personal use; road testing; as part of a SAVE program or “on‐

demand automated vehicle network;” & as part of a platoon.

California’s & Michigan’s approaches differ... California: prescriptive approach Required rulemaking by agencies, pilot tests must be authorized, test vehicles do not require driver behind wheel, but qualified test drivers must have ability to take control, minimum insurance surety bond of $5 million.

Michigan: framework approach Initially, Automakers can test AVs as long as human in car in original legislation.  No agency rulemaking required.

Current legislation, allows driverless cars to be driven for multiple  purposes not just road testing, without a human in the vehicle as the AI is considered the driver. 

Stats from California’s pilot.....• 17 testing permits (mainly Tier 1’s, OEMS & technology 

manufacturers)• 26 traffic accidents involving HAVs (Google 22, Delphi 1, Cruise 

1, GMCruise 1, Nissan 1)• Reported disengagements from automated mode between Dec 

1, 2015 & Nov 30, 2016:– Bosch – 1442 – 982 miles driven(MD)– BMW – 1 – 638 MD– Delphi – 178 – 3125 MD– Ford – 3 – 590 MD – Google – 465 – 1,060,199 MD– GM Cruise – 284  – 9970 MD– Nissan – 28 – 4,099 MD– Mercedes – 336 – 673 MD– Tesla – 182 – 530 MD

https://www.dmv.ca.gov/portal/dmv/detail/vr/autonomous/autonomousveh_ol316https://www.dmv.ca.gov/portal/dmv/detail/vr/autonomous/disengagement_report_2016

European UnionEU has not yet passed any legislation specifically on HAVs. 

EU Directive 2007/46 EEC regulates how vehicles should be designed & operated. 

EU Roadworthiness Directive 2014/45 EU sets out basis for roadworthiness. 

Problems with the Vienna Convention on Road  Traffic Article 8’s language stating ‘every driver shall at all times be able to control his vehicle’ incidentally prevented the development & testing of HAVs.

Amendments to the Convention in 2016 allow drivers to take hands off wheel in self‐driving cars.

ECE Regulation 79 also creates an impediment through requirements for specific steering configurations:

advanced driver steering system is only allowed to control steering as long as the driver remains in primary control of vehicle at all times. 

EU Member States Finland, France, Germany, Netherlands, Sweden & UK implemented legislation in 2015/2016 

Finland, Netherlands & Sweden, all have similar systems to  California’s legislation & regulations for pilot tests, & for driver licensing. 

UK issued code of  practice in July 2015, which must be followed by any groups conducting HAV testing. This includes licensing & training provisions, & a risk management process by the testing group.  

UK issued Vehicle Technology and Aviation Bill 2016‐2017 in February 2017 – outlines liability for insurers of automated vehicles.

Activities in Canada, Japan & Australia

Canada has not yet created federal regulation Province of Ontario in 2016 produced legislation & regulations for a pilot program

Japan has allowed road testing, & is working to develop regulations

Australia has not federally legislated National Transport Commission has set out recommendations & policy positions in 2016

Government of New South Wales introduced legislation in September 2015 for road testing.  NSW released a future transport roadmap in fall 2016 that outlined the ministry’s view on the transition to HAVs

Initial Legal Issues Privacy, Liability, Cybersecurity & Freedom of Information Requests / State Open Records Requests all raised as concerns for automated & connected vehicles. No case law yet on these issues.

NHTSA & FTC have noted they are reviewing hacking & privacy of consumer data in HAVs.   Federal statutes also provide penalties under the Computer Fraud and Abuse Act, Digital Millennium Copyright Act, Wiretap Act, & Patriot Act. 

Initial legal issues (2)

Privacy realm three areas have been identified as needing changes to law:1. Autonomy privacy (i.e. an individual’s privacy 

under 4th amendment to the U.S. Constitution e.g. illegal search & seizure); 

2. Personal information privacy, and 3. Surveillance.

• California passed law regarding consumer privacy in HAVs

Initial Conclusions

Many jurisdictions around the world have begun to draft legislation & regulations with a primary focus on pilot testing. 

California has begun to review how it needs to amend its laws & regulations. 

Legal articles have primarily focused on privacy, liability, cyber security & constitutional protections. 

Thank you!Questions & Suggestions?

kkockelm@mail.utexas.eduloftusotway@mail.utexas.edu