Panta Rhei More Quality of Traffic by centralized or decentralized - - PowerPoint PPT Presentation

▶
panta rhei
SMART_READER_LITE
LIVE PREVIEW

Panta Rhei More Quality of Traffic by centralized or decentralized - - PowerPoint PPT Presentation

Mar 09, 2024 377 likes •807 views

Panta Rhei More Quality of Traffic by centralized or decentralized Control ?! Heraclitus All things are in (540 480 B.C.) constant flux Prof. Dr.-Ing. Dr. h. c. E. Schnieder Overview Traffic System view Control arrangements and

slide-1
SLIDE 1

Panta Rhei

More Quality of Traffic by centralized or decentralized Control ?!

  • Prof. Dr.-Ing. Dr. h. c. E. Schnieder

Overview

Traffic – System view Control arrangements and objectives Control problems, approaches and solutions

Braunschweig, 17.08.2009

Heraclitus (540 – 480 B.C.) All things are in constant flux

slide-2
SLIDE 2

Panta rhei

Traffic and its Contextual Environment

politics / law human / society ecology science economy technology traffic

Braunschweig, 17.08.2009

slide-3
SLIDE 3

Panta rhei

Traffic and its Axiomatic System Properties

state traffic quality structure traffic elements behavior traffic processes function traffic organization system traffic global | local

Braunschweig, 17.08.2009

slide-4
SLIDE 4

Panta rhei

Hierarchical Structure of Traffic Control

Braunschweig, 17.08.2009

slide-5
SLIDE 5

Panta rhei

control ressources control function

local global distributed vehicle control vehicle navigation level crossing control

?

centralized intersection control switch control train control truck dispatching fleet dispatching flow control traffic guidance

Structuring of Traffic Control into Functions and Ressources

Braunschweig, 17.08.2009

slide-6
SLIDE 6

Panta rhei - Fundamental decomposition of traffic control into function and ressources StationaryTraffic Behavior at the Fundamental Diagram

  • traffic flow increases to its global maximum,

any further increase in density leads to instability and decrease of flow

  • behavior expressed by the fundamental diagram

may be denoted as suboptimal

  • collective expression of drivers’ behavior

Braunschweig, 17.08.2009

slide-7
SLIDE 7

Panta rhei - Fundamental decomposition of traffic control into function and ressources Objectives and Means of Traffic Control

Braunschweig, 17.08.2009

quality global increase safety increase capacity and flow decrease travel time and confidence decrease fuel consumption decrease emission means local homogenization of driving behavior increase stability indrease driver reliability increase immediate intervention measurement, control, actuating

slide-8
SLIDE 8

Panta rhei - Fundamental decomposition of traffic control into function and ressources Traffic Flow Interpretation - Microscopic and Macroscopic Traffic Flow Models

macroscopic level traffic flow Q traffic density K average velocity vm microscopic level single vehicle ireal position si_real velocity vi_real … single vehicle ireal position si_real velocity vi_real … single vehicle i position si velocity vi …

Source: tfhrc.gov, stadtentwicklung-berlin.com

Braunschweig, 17.08.2009

slide-9
SLIDE 9

Traffic Flow Interpretation - Microscopic and Macroscopic Traffic Flow Models Panta rhei - Fundamental decomposition of traffic control into function and ressources

highway

va vb xa xb Δs x1 Q(x0,t) v(x0,t) Q(x1,t) v(x1,t) x0 Δx

microscopic macroscopic

Braunschweig, 17.08.2009

slide-10
SLIDE 10

Panta rhei - Fundamental decomposition of traffic control into function and ressources Traffic Flow Interpretation - Microscopic and Macroscopic Traffic Flow Models highway

va vb xa xb Δs x1 Q(x0,t) v(x0,t) Q(x1,t) v(x1,t) x0 Δx

ADAS uses microscopic control input parameters ADAS influences macroscopic traffic characteristic

Braunschweig, 17.08.2009

slide-11
SLIDE 11

Panta rhei - Fundamental decomposition of traffic control into function and ressources Traffic Flow Interpretation - Microscopic and Macroscopic Traffic Meassures

va vb xa xb Δs x1

Q(x0,t) v(x0,t) Q(x1,t) v(x1,t)

x0 Δx

400 800 1200 1600 10 20 30

Verkehrsstärke [Fzg/h] Verkehrsdichte [Fzg/km]

Fundamentaldiagramm, 07.05.2008 Autobahn A2 Richtung Berlin, AS Lauenau

2.2 2.24 2.28 2.32 2.36 2.4 x 10

6

28 30 32 34 36 38 Messzeit [ms] Geschwindigkeit (m/s) vleader Messdaten vego simuliert vego Messdaten

Braunschweig, 17.08.2009

slide-12
SLIDE 12

Panta rhei

control ressources control function

local global distributed vehicle control vehicle navigation level crossing control

?

centralized intersection control switch control train control truck dispatching fleet dispatching flow control traffic guidance e.g. light-signal system or traffic message signs with induction loops

Fundamental Decomposition of Traffic Control into Function and Ressources

Braunschweig, 17.08.2009

slide-13
SLIDE 13

Panta rhei- Centralized traffic control ressources with global control functions

Central Problem of Traffic Flow Measurement

Source: Schick, Peter: Einfluss von Streckenbeeinflussungsanlagen auf die Kapazität von Autobahnabschnitten, Inst. f. Straßen- und Verkehrswesen 2003

500 1000 1500 10 20 30 40 50 60 70 80 90 100

time[s] position vehicle i [m] vehicle positions

lead vehicle

  • veh. 1
  • veh. 2
  • veh. 3
  • veh. 4
  • veh. 5
  • veh. 6

Road Side Unit Road Side Unit

Braunschweig, 17.08.2009

slide-14
SLIDE 14

Panta rhei- Centralized traffic control ressources with global control functions

Central Problem of Traffic Flow Measurement – Sampling Theorem

Source: Schick, Peter: Einfluss von Streckenbeeinflussungsanlagen auf die Kapazität von Autobahnabschnitten, Inst. f. Straßen- und Verkehrswesen 2003

Braunschweig, 17.08.2009

500 1000 1500 10 20 30 40 50 60 70 80 90 100

time[s]

lead vehicle

  • veh. 1
  • veh. 2
  • veh. 3
  • veh. 4
  • veh. 5
  • veh. 6

Road Side Unit Road Side Unit Road Side Unit

slide-15
SLIDE 15

Panta rhei- Centralized traffic control ressources with global control functions

performance criteria ego vehicle collective travel time arrival time traffic availability traffic prediction riding comfort homogenity fuel consumption fuel consumtion of the fleet distance

  • verspeed

traffic safety

Global and Local Control Functions / Performance Criteria

∫

→ =

end

s s

dt T min

∫

→

end

s s

dt a min

2

∫

→

end

s s

Pdt min

Braunschweig, 17.08.2009

min →

T

σ

) ( ) ( → > → = ∆

a

v v p s p

slide-16
SLIDE 16

Panta rhei - Centralized traffic control ressources with global control functions

  • bjective

considered traffic level methodologies distance velocity

  • perational

ego vehicle local vehicle control

  • classical control e.g.

− state space (Roppenecker et al.) −frequency domain

  • robust control e.g.

− Sliding Mode Control [Utkin, Slotine] − Quantitative Feedback Theory [Ackermann et al.] − H∞-Theory [Isidori et. al] flow safety density quality congestion tactical collective of vehicles / fleet coordination of vehicles by cooperative control

  • potential fields [Ögren et al.]
  • Sliding Mode Control [Gazi et al.]
  • Ljapunov-based [Jadbabaie et al.]
  • Receding-Horizon (MPC) [Jadbabaie]
  • graph theory [Baillieul, Fax]
  • Distributed Consensus [Olfati-Saber,

Fax, Murray]

  • Petri-Net decision making

Traffic Control Objectives and Methodologies

Braunschweig, 17.08.2009

slide-17
SLIDE 17

Panta rhei

control ressources control function

local global distributed vehicle control e.g. by means of robust control

?

centralized intersection control switch control train control truck dispatching fleet dispatching flow control traffic guidance

Fundamental Decomposition of Traffic Control into Function and Ressources

Braunschweig, 17.08.2009

slide-18
SLIDE 18
  • for n-th order SISO system
  • definition of sliding surface S(t)
  • tracking means:
  • control law:

equivalent control via:

  • parameter k by stability criterion:

nominal (lin. single track) model model uncertainty

  • design parameters
  • lin. optimization of

Panta rhei - Distributed control ressources with local control functions

Local Control Principles - Sliding Mode Control

d

x x x − = ~ x t t x s

n

~ d d ) , (

1 −

      + = λ ) , ( = t x s ) , ( = t x s 

S(t)

Φ , ,η λ = J u x b x f x

n

⋅ + = ) ( ) ( ˆ

) (

f f F ˆ − =

Braunschweig, 17.08.2009

slide-19
SLIDE 19

Panta rhei - Distributed control ressources with local control functions

  • Application e.g. for lateral vehicle-trailer control
  • max. lateral deviation:
  • 15 cm – worst case scenario
  • 11 cm – wind gust at straight ahead driving
  • stabilizes pendular oscillations due e.g. to cross winds
  • may be used as a driver assistance system to provide driving safety

Example Simulation:

  • velocity: 190 km/h
  • straight ahead driving
  • additionally:

disturbance at steering angle as step of 15 deg. at t=3 s

Local Control Principles - Sliding Mode Control

Braunschweig, 17.08.2009

slide-20
SLIDE 20

Panta rhei - Distributed control ressources with local control functions

Local Control Principles - Quantitative Feedback Theory

Specification of performance requirements by means of boundary function

e.g. control variable

Boundary functions Boundary curves

Discretization of frequency

ω1 ω2 ω3

Prefilter Controller Plant Sensor Braunschweig, 17.08.2009

slide-21
SLIDE 21

Panta rhei - Distributed control ressources with local control functions

Local Control Principles - Quantitative Feedback Theory - Robustness Criteria

  • multiplicative uncertainty
  • direct specification as function ∆M
  • in form of virtual templates directly integrated

into calculation of boundary curves Family of plants

structured unstructured

( ) ( ) ( ) [ ]

s s G s G

M

∆ + = 1

ω1 ω2 ω3 ω4

] 30 .. 20 [ und ] 5 .. 1 [ ], 10 .. 1 [ with = = = b a k

amplitude (db)

phase angle (°)

Braunschweig, 17.08.2009

slide-22
SLIDE 22

Panta rhei - Distributed control ressources with local control functions

Local Control Principles - Nonlinear H∞-Control e.g. for Longitudinal Control

Block diagram of the controlled system physical structure of system

Motor Antriebs- strang

αgas nM

Rad Fahrzeug

nRad vFzg MM MD FFzg vFzg

Motor Antriebs- strang

αgas nM

Rad Fahrzeug

nRad vFzg MM MD FFzg vFzg

power train engine wheel vehicle

engine gear- box clutch clutch

vehicle

Braunschweig, 17.08.2009

slide-23
SLIDE 23

Panta rhei

control ressources control function

local global distributed vehicle control e.g. by means of robust control

?

centralized intersection control switch control train control truck dispatching fleet dispatching flow control traffic guidance

Fundamental Decomposition of Traffic Control into Function and Ressources

Braunschweig, 17.08.2009

slide-24
SLIDE 24

Panta rhei

control ressources control function

local global distributed vehicle control e.g. by means of robust control

?

centralized intersection control switch control train control truck dispatching fleet dispatching flow control traffic guidance

Fundamental Decomposition of Traffic Control into Function and Ressources

Braunschweig, 17.08.2009

slide-25
SLIDE 25

Panta rhei

control ressources control function

local global distributed vehicle control vehicle navigation level crossing control multi agent system distributed consensus combined with Petri nets centralized intersection control switch control train control truck dispatching fleet dispatching flow control traffic guidance

Fundamental Decomposition of Traffic Control into Function and Ressources

Braunschweig, 17.08.2009

slide-26
SLIDE 26

Panta rhei - Distributed control ressources with global control functions

Agent Oriented Paradigm for De-centralized Organization and Operation

Multi Agent System Agent − autonomous entity − goal-oriented and rule based − pro-active and reactive − cooperative and interactive − mobile − system of interacting agents − complex system behavior based upon simple rules − robust and scalable

Braunschweig, 17.08.2009

slide-27
SLIDE 27

Panta rhei - Distributed control ressources with global control functions

− decision maker „on board“ − flexible usage

  • f road capacity

− predetermined operation − stringent allocation

  • f track capacity (slots)
  • rder management/ acquisition
  • transp. exec.

trip planning fleet dispatching

  • rganization
  • peration

planning horizon

minutes days weeks/months

planning horizon

minutes days weeks/months

  • rder management/

acquisition fleet dispatching/

  • rder allocation

transport execution trip planning

  • rganization
  • peration

Realization for Road and Rail

Braunschweig, 17.08.2009

slide-28
SLIDE 28

Panta rhei - Distributed control ressources with global control functions

Realization for Rail

slide-29
SLIDE 29

Panta rhei - Distributed control ressources with global control functions

  • requirements to cooperative control methodology
  • accomplish objectives for optimization of traffic behavior
  • give a mathematical framework which
  • enables the engineer to model perception and communication between

entities (vehicles)

  • integrates these to cooperative control-engineering
  • enables to analyze robustness and performance esp. with

respect to

– communication delays and data dropouts – formation stability, convergence

not supported but by Distributed Consensus (Graph Theory Based) recent approach: combination of Distributed Consensus and Petri-Net based decision making

Cooperative Control Principles - Generic Requirements

Braunschweig, 17.08.2009

slide-30
SLIDE 30

Panta rhei - Distributed control ressources with global control functions

Modeling the ad-hoc-network of C2C-communication by a graph G(N,E)

N is a set of vehicles E is the ordered subset of the cartesian product NxN

Standard consensus algorithm

denotes an arbitrary information variable that is exchanged between the vehicles by the bidirectional graph the information variables converge at each vehicle to the average

This algorithm can be used for control

Generic Approach - Consensus Algorithm

   ∈ = =

∈

else E j i iff a a A

ij p j i ij

) , ( 1 with ) (

} ,..., 1 { ,

∑

=

− − =

n j j i ij i

x x a x

1

) ( 

i

x

the adjacence matrix

Braunschweig, 17.08.2009

slide-31
SLIDE 31

lane 1 lane 2

Panta rhei - Distributed control ressources with global control functions

Road Discretization - Petrinet based Decision-Making

Braunschweig, 17.08.2009

slide-32
SLIDE 32

lane 1 lane 2

Panta rhei - Distributed control ressources with global control functions

Road Discretization - Petrinet based Decision-Making

Braunschweig, 17.08.2009

slide-33
SLIDE 33

Panta rhei - Distributed control ressources with global control functions

  • bjective

criteria party responsible (interessent) party concerned (demander)

ensuring traffic safety traffic safety (risk) state, society traffic object, users of means of transportation, suppliers of infrastructure

  • ptimal economical

benefit and minimum travel time, rapidness

  • ptimal transport

expenses, parking costs, low amount of planning time, low travel time, dwell time of traffic jam infrastructure, means of transportation producer, logistics fleet manager sales market for end-user products and services , means of transportation user, traffic object cooperation, migration and integration net splitting, reachability

  • f objectives,

Park and Ride means of transportation supplier, infrastructure supplier, logistics, commodity supplier client, means of transportation user, logistic objects, logistician, society net load distribution and environmental protection load balancing, road damages, emissions policy, state, society, means of transportation producer and organiser means of transportation user Objectives, Criteria and their Responsible and Demanding Parties

Braunschweig, 17.08.2009

slide-34
SLIDE 34

Panta rhei - Distributed control ressources with global control functions

A Mirical occurs

„Shouldn‘t we get a little more detailed in the implementation stage?“ MIRACLE

Braunschweig, 17.08.2009

slide-35
SLIDE 35

Panta Rhei

More Quality of Traffic by centralized or decentralized Control ?!

  • Prof. Dr.-Ing. Dr. h. c. E. Schnieder

Summary

Traffic – System view Control arrangements and objectives Control problems, approaches and solutions

Braunschweig, 17.08.2009

Heraclitus (540 – 480 A.C.) All things are in constant flux

More Quality of Traffic by decentralized Control !

slide-36
SLIDE 36
slide-37
SLIDE 37

traffic quality ... accidents traffic ... flow noise congestion traffic safety density distance ... frequency ...

Term Properties Characteristics Quantities Value

. . . . . . . . .

comprehension meaning

slide-38
SLIDE 38

Ontologic Model of Transportation and Traffic Notions

traffic

  • bjects of transport

1 * transport infrastructure environment 1 * traffic organization means of transportation 1 * tracks roads 1 * 1 * persons goods trains vehicles 1 * 1 * 1 * 1 * 1 * 1 *

slide-39
SLIDE 39
slide-40
SLIDE 40

Methodology and tools

Safety models and description means

Models State independent failure sources State dependent failure sources FMECA RBD FTA Memoryless distributions General distributions Markov chains Petri nets C++ (Monte-Carlo-Simulation)

slide-41
SLIDE 41

Methodology and tools

Descriptions and results of the simulation

FMECA RBD FTA Markov chains Petri nets C++ (Monte-Carlo-Simulation) Availability, failure rates accident severity

slide-42
SLIDE 42

Methodology and tools

Descriptions and results of the simulation

FMECA RBD FTA Markov chains Petri nets C++ (Monte-Carlo-Simulation)

  • Isograph
  • Relex
  • Reliasoft
  • Reliass
  • …
  • TimeNet
  • CPN-Tools
  • PiTool
  • …
  • Boost-Library
  • MPI-Standard
  • PVM
  • …