Large-scale simulations for automated mobility

Results for Zurich and Paris

Sebastian Hörl

Paris School of Economics

27 January 2020

The street in 1900

http://www.loc.gov/pictures/item/2016800172/

The street today

https://commons.wikimedia.org/wiki/File:Atlanta_75.85.jpg

The street of tomorrow?

  • Autonomous Mobility

  • Mobility as a Service

  • Mobility on Demand

  • Electrification

  • Aerial Mobility

Julius Bär / Farner

I. MATSim

Agent-based models

Senozon VIA

https://pixabay.com/en/traffic-jam-stop-and-go-rush-hour-143391/

MATSim

Home

Work

Shop

Home

until 8am

9am to 6pm

6:15m to 6:30pm

from 6:45pm

walk

public

transport

walk

MATSim

Network simulation

Scoring of the plans

Selection and modification

Scenario

MATSim with Discrete Choice Model

Network simulation

Scoring

Replanning

Network simulation

Choice Model

Hörl, S., M. Balac and K.W. Axhausen (2019) Pairing discrete mode choice models and agent-based transport simulation with MATSim, presented at the Annual meeting of the Transport Research Board 2019, January 2019, Washington D.C.

https://pixabay.com/en/zurich-historic-center-churches-933732/

II. AMoD in Zurich

Fleet control

Load-balancing heuristic

Simple heuristic, fast runtime

 

Gloal Euclidean Bipartite Matching

Standard algorithm in Operations Research

?

?

Assignment

Car by Adrien Coquet from the Noun Project

Hail by Bradley Avison from the Noun Project

Example algorithms

Load-balancing heuristic

Undersupply

  • More vehicles than requests
  • When request pops up
    find closest vehicle

Oversupply

  • More requests than vehicles
  • When vehicle gets available
    find closest request

able to serve remote demand

able to serve remote demand

able to serve peek demand

Hörl, S., C. Ruch, F. Becker, E. Frazzoli and K.W. Axhausen (2019) Fleet operational policies for automated mobility: a simulation assessment for Zurich, Transportation Research: Part C, 102, 20-32.

Example algorithms

Global Bipartite Matching

  • Find matching of vehicles and requests that minimizes total Euclidean (or network-based) empty distance
     
  • Nice benchmark because sole objective is to minimize empty distance

Hörl, S., C. Ruch, F. Becker, E. Frazzoli and K.W. Axhausen (2019) Fleet operational policies for automated mobility: a simulation assessment for Zurich, Transportation Research: Part C, 102, 20-32.

Fleet control

Feedforward Fluidic Optimal Rebalancing Policy

Linear program matching a priori known trip rates

 

Adaptive Uniform Rebalancing Policy

Linear program distributing vehicles uniformly

Car by Adrien Coquet from the Noun Project

Hail by Bradley Avison from the Noun Project

?

?

?

?

Redistribution

Example algorithms

Feed-forward fluidic rebalancing strategy

Travel time

Rebalancing flows

Arrival rate

Transition probability

Hörl, S., C. Ruch, F. Becker, E. Frazzoli and K.W. Axhausen (2019) Fleet operational policies for automated mobility: a simulation assessment for Zurich, Transportation Research: Part C, 102, 20-32.

Example algorithms

Feed-forward fluidic rebalancing strategy

Travel time

Rebalancing flows

Arrival rate

Transition probability

  • Informed algorithm from historical data
     
  • Minimization of rebalancing time
     
  • Minimization of waiting time?

Hörl, S., C. Ruch, F. Becker, E. Frazzoli and K.W. Axhausen (2019) Fleet operational policies for automated mobility: a simulation assessment for Zurich, Transportation Research: Part C, 102, 20-32.

Example algorithms

Adaptive Uniform Rebalancing Policy

Vehicles per zone

Current requests

  • Minimization of rebalancing time
     
  • Working on current information
     
  • Uniform distribution of vehicles

Hörl, S., C. Ruch, F. Becker, E. Frazzoli and K.W. Axhausen (2019) Fleet operational policies for automated mobility: a simulation assessment for Zurich, Transportation Research: Part C, 102, 20-32.

Fleet control

Hörl, S., C. Ruch, F. Becker, E. Frazzoli and K.W. Axhausen (2019) Fleet operational policies for automated mobility: a simulation assessment for Zurich, Transportation Research: Part C, 102, 20-32.

Fleet control

Hörl, S., C. Ruch, F. Becker, E. Frazzoli and K.W. Axhausen (2019) Fleet operational policies for automated mobility: a simulation assessment for Zurich, Transportation Research: Part C, 102, 20-32.

Fleet sizing with dynamic demand

Fleet sizing with dynamic demand

Fleet sizing with dynamic demand

What do we know about automated vehicles?

Cost structures?

User preferences?

System impact?

Cost Calculator for automated mobility

Stated preference survey

MATSim simulation

1

2

3

AMoD Cost Calculator

Bösch, P.M., F. Becker, H. Becker and K.W. Axhausen (2018) Cost-based analysis of autonomous mobility services, Transport Policy, 64, 76-91

AMoD Survey

Felix Becker, Institute for Transport Planning and Systems, ETH Zurich.

AMoD Survey

13 CHF/h


 

AMoD

Taxi

19 CHF/h

 

 

Conventional

Car

12 CHF/h
 

 

Public

Transport

VTTS

Car by Adrien Coquet from the Noun Project

Bus by Simon Farkas from the Noun Project

Wait by ibrandify from the Noun Project

AMoD

AMoD Survey

Car by Adrien Coquet from the Noun Project

Bus by Simon Farkas from the Noun Project

Wait by ibrandify from the Noun Project

13 CHF/h
 

 

AMoD

Taxi

19 CHF/h

 

 

Conventional

Car

12 CHF/h
 

 

Public

Transport

VTTS

 
21 CHF/h

 


32 CHF/h

AMoD

Model structure

Cost calculator

Plan modification

Discrete Mode Choice Extension

Mobility simulation

Prediction

Price

Trips

- Utilization

- Empty distance, ...

- Travel times

- Wait times, ...

Results

Maximum

38k rides

Results

Maximum

38k rides

Results

Visualisation

Automated taxi

Pickup

Dropoff

III. Paris Scenario

https://pixabay.com/en/paris-eiffel-tower-night-city-view-3296269/

The case of Île-de-France

Census data

Récensement de la population

RP

Icons on this and following slides: https://fontawesome.com // Background: Simunto VIA

The case of Île-de-France

Dispositif sur les revenus localisés sociaux et fiscaux
Income tax data

RP

FiLo
SoFi

 

The case of Île-de-France

RP: Flux de mobilité
Commuting data

RP

RP
Mob

 

FiLoSoFi

The case of Île-de-France

Enquête globale de transport
Household Travel Survey

RP

EGT

FiLoSoFi

RP Mob

Enquête national transports et deplacements
Household Travel Survey

ENTD

The case of Île-de-France

Enquête globale de transport
Household Travel Survey

RP

EGT

FiLoSoFi

RP Mob

Enquête national transports et deplacements
Household Travel Survey

ENTD

The case of Île-de-France

Base permanente des équipements
Enterprise census

RP

BPE

FiLoSoFi

RP Mob

EGT

ENTD

The case of Île-de-France

OpenStreetMap
Road network

RP

OSM

FiLoSoFi

RP Mob

EGT

ENTD

BPE

IDFm GTFS
Public transport schedule

GTFS

The case of Île-de-France

RP

FiLoSoFi

RP Mob

EGT

ENTD

BPE

OSM

GTFS

Discrete choice model
Agent behaviour

DCM

The case of Île-de-France

RP

FiLoSoFi

RP Mob

EGT

ENTD

BPE

OSM

GTFS

DCM

The case of Île-de-France

RP

FiLoSoFi

RP Mob

EGT

ENTD

BPE

OSM

GTFS

DCM

The case of Paris

RP

FiLoSoFi

RP Mob

EGT

ENTD

BPE

OSM

GTFS

DCM

The case of Paris

RP

FiLoSoFi

RP Mob

EGT

ENTD

BPE

OSM

GTFS

DCM

The case of Paris

RP

FiLoSoFi

RP Mob

EGT

ENTD

BPE

OSM

GTFS

DCM

The case of Paris

RP

FiLoSoFi

RP Mob

EGT

ENTD

BPE

OSM

GTFS

DCM

The case of Île-de-France

RP

FiLoSoFi

RP Mob

EGT

ENTD

BPE

OSM

GTFS

DCM

The case of Île-de-France

RP

FiLoSoFi

RP Mob

EGT

ENTD

BPE

OSM

GTFS

DCM

The case of Île-de-France

Open
Data

RP

FiLoSoFi

RP Mob

EGT

ENTD

BPE

OSM

GTFS

Open
Software

+

=

Reproducible research

Verifiable results

Integrated testing

  • Detailed documentation of whole open-source pipeline is under preparation

DCM

The case of Île-de-France

CO2 Emmissions

AMoD in Paris

Hörl, S., M. Balac and K.W. Axhausen (2019) Dynamic demand estimation for an AMoD system in Paris, paper presented at the 30th IEEE Intelligent Vehicles Symposium, June 2019, Paris, France.

Travel Behaviour

Zurich model, calibrated
for Paris population

Cost sturcture

Adapted from Berlin

AMoD Simulation

Load-balancing heuristic

AMoD in Paris

Maximum static demand: 2.3M trips

Hörl, S., M. Balac and K.W. Axhausen (2019) Dynamic demand estimation for an AMoD system in Paris, paper presented at the 30th IEEE Intelligent Vehicles Symposium, June 2019, Paris, France.

AMoD in Paris

Maximum static demand: 2.3M trips

Hörl, S., M. Balac and K.W. Axhausen (2019) Dynamic demand estimation for an AMoD system in Paris, paper presented at the 30th IEEE Intelligent Vehicles Symposium, June 2019, Paris, France.

Do it yourself!

  • Simple toy example for Auckland NZ
     
  • Includes
      MATSim,
      Discrete Mode Choice,
      and AMoDeus
     
  • Requires JAVA & Eclipse

Further work

Routing

Lu, Chengqi (2019) Congestion-Aware Operation of Coordinated Autonomous Mobility-on-Demand Systems, Master thesis, Institute for Dynamic Systems and Control (IDSC), ETH Zurich, Zürich, Switzerland.

Further work

Parking

Ruch, C., S. Hörl, R. Ehrler, M. Balac, E. Frazzoli (2020) How Many Parking Spaces Does a Mobility-on-Demand System Require?, Under review.

Further work

Parking

Ruch, C., S. Hörl, R. Ehrler, M. Balac, E. Frazzoli (2020) How Many Parking Spaces Does a Mobility-on-Demand System Require?, Under review.

Further work

Serving real-world taxi demand

San Francisco

Chicago

Zurich

Ruch, C., S. Hörl, J. Hakenberg, E. Frazzoli (2020) The Impact of Fleet Coordination on Taxi Operations, Under review.

Thanks!

Questions so far?

Contact: sebastian.hoerl@ivt.baug.ethz.ch

Large-scale simulation scenarios of automated mobility

By Sebastian Hörl

Large-scale simulation scenarios of automated mobility

Paris School of Economics, 27 January 2020

  • 916