DACCORD
www.hcg.nl/daccord

Development and Application of
Co-ordinated Control of Corridors

Telematics Applications Programme
TRANSPORT

 Table of Contents

 TECHNICAL ABSTRACT 4

2.1 Relation with Work Package WP04

2.2 Overview of Control Strategies in DACCORD

2.2.1 Point, Link, and Network Control
2.2.2 Tactical and Strategic Approaches
2.2.3 The DACCORD Sites

3.1 Motorway-To-Motorway Control (MTMC)

3.1.1 ALINEA: Ramp-Metering Strategy
3.1.2 Barrier Devices Modelling
3.1.3 ALINEA Strategy

3.2 Speed Recommendation and Lane Use

3.3 Co-ordinated Ramp Metering

3.3.1 Amsterdam A10-West Co-ordinated Ramp Metering

3.3.1.1 Background
3.3.1.2 Basic Principles of Co-ordinated Ramp Metering

3.3.2 A Non-minimal State Variable Feedback Approach to Co-ordinated Ramp Metering

3.4 Network-level Co-ordinated Ramp Metering

3.4.1 Introduction
3.4.2 Mathematical Problem Formulation
3.4.3 Linear Programming Approach (LPA)
3.4.4 Non linear Programming Approach (NLPA)

3.5 Traffic Information Display

3.5.1 Introduction
3.5.2 Strategy for Travel Time Display on Variable Message Signs

3.6 Advanced Routing Control Strategies

3.7 Integrated Motorway Control Strategies

Technical Abstract

	Motorway-to-motorway control.
	Speed recommendation and lane use.
	Co-ordinated ramp metering.
	Network level co-ordinated ramp metering.
	Traffic information display.
	Advanced routing control.
	Integrated motorway network control.

For each one of these control measures, control strategies are discussed and further presented in the corresponding annexes.

Objectives of Work Package WP06

The objectives of WP06, according to Technical Annex 1, is to develop generic, co-ordinated control tools for periurban traffic networks including a number of different control measures (ramp metering, motorway-to-motorway control, VMS, VDS, route guidance), and to apply them to the DACCORD sites.

The work in Work Package WP06 will be included in three deliverables:

Deliverable D06.1 : Co-ordinated control strategies.
Deliverable D06.2 : Co-ordinated control software.
Deliverable D06.3 : Off-line simulation results.

The objective of this deliverable D06.1 is to report on the control strategies developed and control strategies that are implemented (or will be implemented) in the DACCORD sites. These control strategies include:

	Motorway-to-motorway control.
	Speed recommendation and lane use.
	Co-ordinated ramp metering.
	Network level co-ordinated ramp metering.
	Traffic information display.
	Advanced routing control.
	Integrated motorway network control.

The above mentioned strategies will be presented in a general way in the main part of the deliverable. Their more detailed description will be part of the corresponding annexes attached at the end of this document. The contents of the annexes are as follows.

In Annex A the METANET traffic flow model is described in some detail. The detailed modelling of links and nodes and also some additional features of METANET, developed within DACCORD, are presented. This annex serves as a reference for the Annexes B and C of this deliverable.

In Annex B the modelling of the Amsterdam motorway network is presented. The Amsterdam network was modelled according to the METANET specifications and the resulting network is described. The major part of Annex B is concerned with the quantitative validation of the Amsterdam motorway network. The goal of the quantitative validation is to estimate METANET’s model parameters and to validate the modelling accuracy. The tool META was used for this purpose. The methodology used and the results obtained are discussed in detail.

In Annex C a control strategy based on non-linear optimisation is presented. The methodology used is general enough to lead to both co-ordinated and integrated control strategies may be obtained from its use. The traffic flow model that is used is the METANET model. The optimisation algorithm for obtaining optimal control trajectories and the traffic flow model are parts of a generic control software tool called AMOC (Advanced Motorway Optimal Control). AMOC can calculate optimal control trajectories for a combination of control measures (VMS and/or ramp metering) for motorway networks of arbitrary topology. AMOC will be the subject of Deliverable D06.2.

In Annex D, the routing control problem is discussed from several angles. The DACCORD user requirements with regard to routing are summarised. The existing routing control experiments are reviewed. A summary is provided of the existing theory of routing control. The interaction of information and control in routing is explained. The advantages and disadvantages of currently used routing control measures are discussed. At the theoretical level, the routing control problem is formulated as a dynamic game problem. Attention is restricted to a Nash equilibrium and the control problem of determining a Nash equilibrium is discussed. For simple networks it is argued that a Nash equilibrium exists. A control algorithm is specified that produces an input trajectory that is an approximation to the input trajectory produced by a Nash equilibrium control law.

In Annex E the problem of integrated control is discussed. Control measures have been implemented to provide information and directions to drivers. Such measures so far have been installed in an incremental way. There is a need for integrated control of motorway networks in which actions of the control measures all strive for the same objective. In this annex a conceptual model for integrated traffic control is proposed. The control objectives, control functions, and control tasks at the network, link, and point level of a hierarchical model are defined and related.

Annex F is concerned with the development of a new co-ordinated ramp metering control strategy based on a non-minimal state variable feedback approach. The research involves data-based modelling and linearisation of the motorway traffic system, coupled with a non-minimal state variable feedback approach to control system design. This approach yields Proportional-Integral-Plus Linear Quadratic (PIP-LQ) optimal controllers. By co-ordinating the metering at several on-ramps, traffic can be more equally distributed over the entrances to the motorway, resulting in fewer disturbances and smaller queues. The proposed PIP-LQ approach derives from earlier research into Non-Minimal State Space (NMSS) control system designs and can be considered as one particular implementation of Generalised PIP (GPIP) control

Annex G is concerned with the development and the investigation of the traffic impact of the Motorway-To-Motorway Control (MTMC). In order to evaluate the impact of such control action on the traffic conditions of the Paris test site, variants of the well known traffic responsive ramp metering control strategy ALINEA will be implemented on A6a entrance on the BPI using two possible control devices: signal lights and barriers. Annex G is focused on the description of the strategies using the signal lights or the geometric capacity restriction (barriers) devices.

In Annex H, an integrated control approach for traffic corridors is developed. The control measures which are produced in a co-ordinated way to improve traffic performance in traffic corridors include signal control, ramp metering and route guidance, and Motorway-To-Motorway control. The common traffic control objective, incorporated in this framework, is the minimisation of all delays in the corridor. The numerical solution of a formulated large-scale optimal control problem will be investigated by application of the linear and non linear programming methodology. To evaluate the integrated control approach, the macroscopic modelling tool METACOR, including traffic assignment modelling, is used.

In Annex I the assessment of centralised co-ordinated ramp metering control strategies for the A10-West is reported. There are two levels of co-ordination which vary in the level of enforcement that can be imposed from the CTMS on the local ramp meters. The first level of co-ordination is the decision which local systems should meter and which should not. The second level of co-ordination is the decision which cycletime a local system should use. The control strategies implemented and assessed for the A10-West are based on both levels of co-ordination. Hence, based on the above distinction the control strategies for co-ordinated ramp metering that are dealt with are the RWSCEN and ALICEN for the first level, and RWSCOR and METALINE for the second level of co-ordination. Moreover, a ‘ZERO-situation’ is also considered as a ‘DO-NOTHING’-control strategy.

In Annex J the text strategy for the display of traffic information on Variable Message Sign A13 The Hague-Rotterdam is described. The objective of the text strategy is to translate a computed travel time over a part of a road network into a message to be displayed on the variable message sign. When the computed travel times are assumed not to be reliable, then the text strategy allows for the presentation of a computed queue length rather than a computed travel time. The formulated text strategy addresses three main issues: choice of stretch of road; travel times, and queue lengths. In Annex K a generally applicable text strategy which differs slightly from the text strategy for the travel time display on the A13 DRIP is presented. This strategy was developed after the text strategy of the A13 DRIP and is more general in nature.

Finally, in Annex L, the control strategies of speed recommendation and lane use implemented in the Italian test site are presented. The strategy of speed recommendation is implemented both for the optimal exploitation of the available capacity of the motorway (it takes into account also the atmospheric conditions like fog, rain, road surface status, etc.) and for slowing down the traffic when some anomalous event (for example an accident, a queue, etc.) happens on a section of the motorway. In case of accidents, roadworks or whatever event that implies the closure of some lane, another control strategy acts on the VMS-control system together with the recommended speed: the lane use strategy. When the system detects the presence of a lane closure, the VMS stations before the correspondent section display for each lane a symbol (generally a colored vertical or oblique arrow) which indicates if the driver has to keep the lane or if he has to change it and move to the parallel one. In normal conditions a vertical green arrow is displayed.