ICVES2017-Presentation Presentation June 2017 DOI: - - PDF document

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/318131535

ICVES2017-Presentation

Presentation · June 2017

CITATIONS READS

68

3 authors, including: Some of the authors of this publication are also working on these related projects: International research Chair Drive for All View project Astrobots Coordinations for Dark Matter Studies View project Ezequiel González Debada École Polytechnique Fédérale de Lausanne

8 PUBLICATIONS 27 CITATIONS

SEE PROFILE

Laleh Makarem École Polytechnique Fédérale de Lausanne

23 PUBLICATIONS 334 CITATIONS

SEE PROFILE

All content following this page was uploaded by Ezequiel González Debada on 04 July 2017.

The user has requested enhancement of the downloaded file.

A Virtual Vehicle Based Coordination Framework For Autonomous Vehicles in Heterogeneous Scenarios

Ezequiel Debada, Laleh Makarem and Denis Gillet ICVES 2017 - Vienna

Heterogeneity raising awareness Coordination at intersection is not only a collision avoidance problem ´Likely´ reality Vs centralization and homogeneity? Towards better intersection usage

20 sec motivation

2

Heterogeneity raising awareness Coordination at intersection is not only a collision avoidance problem ´Likely´ reality Vs centralization and homogeneity? Towards better intersection usage

What if

heterogeneity

was part of the design process?

20 sec motivation

3

Preliminary control and case study

Interaction mechanism

Targeted scenario

Why another coordination framework? Objective

Designing an heterogeneity-friendly framework that

We will

present

Maximizes the potential impact Simplifies the execution of cooperative maneuvers Allows complex interaction

4

Scenario

Cooperativeness

Connected Distributed Heterogeneous Cooperative uV cAV ACC

5

uV (unconnected Vehicle) ACC (Advanced Cruise Control) cAV (connected AV)

Scenario

Cooperativeness

Connected Distributed Heterogeneous Cooperative uV cAV ACC

What to broadcast > How to maximize cooperation impact >

?

6

uV (unconnected Vehicle) ACC (Advanced Cruise Control) cAV (connected AV)

Inspired Improves

cooperative effort interaction among heterogeneous agents regulation capabilities of CAVs

Requires

positions and speed

by

lane change indicators increasing maneuver headway requiring only one smart vehicle increasing maneuver options

to model and execute cooperative maneuvers

Virtual Vehicles

also referred to as phantom vehicles, projections

Framework building block

7

Inspired Improves

cooperative effort interaction among heterogeneous agents regulation capabilities of CAVs

Requires

positions and speed

by

lane change indicators increasing maneuver headway requiring only one smart vehicle increasing maneuver options

to model and execute cooperative maneuvers

Virtual Vehicles

also referred to as phantom vehicles, projections

Framework building block

8

Inspired Improves

cooperative effort interaction among heterogeneous agents regulation capabilities of CAVs

Requires

positions and speed

by

lane change indicators increasing maneuver headway requiring only one smart vehicle increasing maneuver options

to model and execute cooperative maneuvers

Virtual Vehicles

also referred to as phantom vehicles, projections

Framework building block

9

Inspired Improves

cooperative effort interaction among heterogeneous agents regulation capabilities of CAVs

Requires

positions and speed

by

lane change indicators increasing maneuver headway requiring only one smart vehicle increasing maneuver options

to model and execute cooperative maneuvers

Virtual Vehicles

also referred to as phantom vehicles, projections

Framework building block

10

Inspired Improves

cooperative effort interaction among heterogeneous agents regulation capabilities of CAVs

Requires

positions and speed

by

lane change indicators increasing maneuver headway requiring only one smart vehicle increasing maneuver options

to model and execute cooperative maneuvers

Virtual Vehicles

also referred to as phantom vehicles, projections

Framework building block

11

uV cACC cAVs uV en en en cACC en en en cAV en ex ex

Exogenous, and Endogenous projections

Every vehicle, no matter its kind, is assigned one projection while maneuvering

(broadcast) (guessed)

uV (unconnected Vehicle) cACC (connected autonomous cruise control) cAV (connected autonomous vehicle)

EGO Nearby car

Virtual Vehicles

also referred to as phantom vehicles, projections

Framework building block

to model and execute cooperative maneuvers

Control architecture

Projection management Control references (execution)

13

Control architecture

Where and when to project? How to control (and coordinate) virtual and real car?

Projection management Control references (execution)

14

Control architecture

Where and when to project? How to control (and coordinate) virtual and real car?

Projection management Control references (execution)

Depends on the maneuver

15

Case study Single lane roundabouts

Perfect test-bench for collision avoidance - cooperation - and exploiting knowledge of traffic dynamic

16

Projection placement Execution

control of projected and real car Real car coupling law Projection

= car-following model closest gap reachable can be communicated does not disturb* circulating traffic

Case study Single lane roundabouts

Perfect test-bench for collision avoidance - cooperation - and exploiting knowledge of traffic dynamic

17

Models and simulator Longitudinal control

Gipps car-following model

Merging criteria for uVs and ACC-cars

uVs ACC gap-acceptance cooperative effort

Dynamics

Longitudinal Dynamics, second order model

18

Simulation example The video is available at https://goo.gl/1hV413

Experiments and preliminary results

Case 1 2 3 4 5 6 uV 200 100 100 100 cACCs 200 100 50 cAVs 200 100 50

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 1500 2000 2500 3000 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 1400 1600 1800 2000 2200

Throughput Scenario ID Experiments

6 cases (heterogeneity) and 40 scenarios (traffic level)

Cases Results

20

Experiments and preliminary results

Case 1 2 3 4 5 6 uV 200 100 100 100 cACCs 200 100 50 cAVs 200 100 50

Experiments

6 cases (heterogeneity) and 40 scenarios (traffic level)

Cases Results

5 10 15 20 2 4 6 8 10 12 14 16

Waiting time Circulating speed

5 10 15 20 2 4 6 8 10 12 14 16

Waiting time Circulating speed

21

Wrapping up

Performance is reflected on the intersection state Exploiting traffic dynamic for smarter CAVs Heterogeneity requires an extensive evaluation Friendly integration of heterogeneous parties

22

Thank you

Any question?