Tasks

The fundamental objective of this workpackage is to illustrate the significant contributions to the improvement of the design flow for embedded systems in the automotive industry that can be brought by the introduction of hybrid system modeling, analysis, synthesis and verification techniques. The overall work is divided in four tasks.

Task 4c.1: Hybrid models for automotive control
Due to the rich discrete-continuous interactions in automotive mechanical subsystems, hybrid modelling plays an important role in WP4c, as witnessed by the significant results obtained during the first year. Initially devoted mainly to HCCI and GDI engines modelling, this task has been extended to include spark ignition Multipoint Injection Engines, fuel consumption and energy management, exhaust after-treatment systems, powertrains, braking system and vehicle dynamics. In particular, the partners will focus the activities within this task on: 

Case study 1: Hybrid modeling of engines. 
The Homogeneous Charge Compression Ignition (HCCI) combustion engine principle lacks direct ignition timing control. Since auto-ignition of a homogeneous mixture is very sensitive to operating conditions, combustion stabilization is critical and fast cycle -based combustion phasing control is necessary for reliable operation. The HCCI combustion engine constitutes a very interesting case study for hybrid system analysis and control, the combustion exhibiting continuous-time dynamics and the engine operation being periodic and asynchronous. Effective control achieving robust combustion stabilization can be obtain only on the basis of a detailed hybrid model of the HCCI engine that describes the interactions between the discrete and continuous dynamics in the system.

The experience gathered developing this case study will allow the partners to analyze advantages and difficulties encountered in hybrid modelling for automotive applications, provide motivations for new directions in hybrid system modelling techniques and clear indications for their fruitful introduction in industry.

Milestones

M4c.1.1	Analysis of hybrid interactions in engines, powertrains and vehicle dynamics	Month 4
M4c.1.2	Analysis of injection, ignition and combustion processes in GDI and HCCI engines	Month 4
M4c.1.4	Advantages achieved using hybrid modelling with respect to mean-value modelling	Month 18

Deliverables (downloadable from HYCON deliverables repository)

D4c.1.1	Hybrid models of automotive powertrains and braking systems	Month 12
D4c.1.2	Hybrid models of GDI SI engines suitable for control design	Month 12
D4c.1.3	Hybrid models of HCCI engine suitable for control design	Month 12
D4c.1.4	Preliminary report on Case study 1: Hybrid modeling of engines	Month 28
D4c.1.5	Applicability of hybrid analysis and verification methodologies to automotive models	Month 30
D4c.1.6	Mid-term report on Case study 1: Hybrid modeling of engines	Month 39
D4c.1.7	Functionalities and limitations of tools for simulations and analysis of hybrid models in automotive appl.	Month 42
D4c.1.8	Final report on Case study 1: Hybrid modeling of engines	Month 48

 

Task 4c.2: Hybrid control in automotive applications
In the first year, the automotive control task covered a wide range of applications ranging from engine and vehicle dynamics control to efficient tail-pipe emission control. The application of hybrid design methodologies to automotive control problems will be further investigated. Optimal, switching, discrete-event and multi-rate approaches to controller design will be considered. In particular, research activities will be directed to 

Case study 2: Optimal and MPC in vehicle dynamics control. 
Hybrid MPC is among the most successful approaches to hybrid controller design, as demonstrated by the many applications reported including several in automotive. Optimal control problems for discrete-time hybrid systems can be solved by modelling the hybrid system as a mixed logical dynamical (MLD) system and solving the MPC control problem by mixed-integer programming (MIP). For efficient on-line implementation, the MPC control law can be expressed explicitly as a collection of affine state feedbacks and of corresponding polyhedral cells in the state space. The application of hybrid MPC in vehicle dynamics control is investigated considering both longitudinal dynamics (e.g. adaptive cruise control) and vertical and lateral dynamics (e.g. suspension control and yaw control).

The resulting hybrid control algorithms will be compared with the standard ones employed in the industry in terms of performances, implementation requirements and calibration complexity. The quality of the proposed hybrid control strategies will be assessed by experimental results, either obtained using partners’ test bed facilities or carried out by our industrial partners.

Milestones

M4c.2.1	Analysis of event-based and time-triggered actions in multivariable engine control	Month 6
M4c.2.2	Identification of the hybrid behaviors in traction and braking control	Month 6
M4c.2.3	Performance improvements using hybrid control design methodologies	Month 18

Deliverables (downloadable from HYCON deliverables repository)

D4c.2.1	Formulation of hybrid control problems for vehicle dynamics and engine control	Month 12
D4c.2.2	Preliminary report on Case study 2: Optimal and MPC in vehicle dynamics control	Month 24
D4c.2.3	Industrial feasibility of hybrid control algorithms in automotive applications	Month 30
D4c.2.4	Mid-term report on Case study 2: Optimal and MPC in vehicle dynamics control	Month 36
D4c.2.5	Functionalities and limitations of tools for hybrid controller design in automotive applications	Month 42
D4c.2.6	Final report on Case study 2: Optimal and MPC in automotive control	Month 48

 

Task 4c.3: Design methodologies for embedded automotive control systems
Hybrid system analysis techniques allow designers to evaluate both the performances of the control algorithms at the functional level and the degradation due to their implementation on hw/sw platforms. Implementation platform evaluation will be obtained by introducing in the closed-loop model an abstract representation of the main effects due to the implementation and by analysing the behaviour for the resulting system. Exploration of the implementation parameters space will be performed to evaluate the trade-off between cost of the implementation platform and closed-loop performances. Combinations of randomized algorithms, formal analysis, and simulations will be used to sketch the critical scenarios at which the limits of performance of fault-tolerant controllers are reached. The proposed techniques will be evaluated in 

Case study 3: Platform-based design of an ECU.
The platform-based design methodology provides concepts and techniques to achieve an efficient design, aimed at maximizing reuse at each design step and achieving early verification with abstracted information from possible implementation platforms. In this context, a platform is a layer of abstraction that hides the unnecessary details of the underlying implementation and yet carries enough information about the layers below to prevent design iterations. Platform-based design has been successfully applied to different domains including automotive. Abstraction and verification are essential tools for an effective application of platform-based design. Hybrid system modelling, simulation and verification techniques play clearly a fundamental role in this respect, since they are instrumental to a successful application of platform-based design in the development flow for embedded automotive control systems.

Milestones

M4c.3.1	Analysis of the major effects of the implementation on closed-loop performances	Month 8
M4c.3.2	Criteria for the evaluation of different implementation platforms	Month 18
M4c.3.3	Analysis of requirements for hybrid models in automotive software product lines	Month 7
M4c.3.4	Standardisation and research policies: links to AUTOSAR and ARTEMIS	Month 33

Deliverables (downloadable from HYCON deliverables repository)

D4c.3.1	Hybrid models representing implementation details of control algorithms	Month 12
D4c.3.2	Preliminary report on Case study 3: Platform-based design of ECUs	Month 24
D4c.3.3	Randomized algorithms versus sthocastic methods in the platform based design	Month 30
D4c.3.4	Mid-term report on Case study 3: Platform-based design of ECUs	Month 36
D4c.3.5	Potential impact of hybrid techniques in automotive networked control design	Month 42
D4c.3.6	Final report on Case study 3: Platform-based design of ECUs	Month 48

 

Task 4c.4: Dissemination activities
Significant efforts will be devoted to dissemination of the results achieved in this workpackage among the industrial partners. Effective dissemination will be achieved by involving HYCON full and premium member automotive companies in the research activities. The second HYCON workshop on “Automotive applications of hybrid systems”, scheduled in May 2006, will represent a major occasion for sharing the results and strengthening the collaborations with industry.

Milestones

M4c.4.1	In depth look at the dissemination activities, recommendations for the future	Month 12
M4c.4.2	Feedbacks from WP4c-IAB at the HYCON & CEmACS Workshop on Automotive Systems and Control	Month 21
M4c.4.3	In depth look at the dissemination activities, recommendations for the future	Month 18

Deliverables (downloadable from HYCON deliverables repository)

D4c.4.1	Proc. of the 1st HYCON workshop “Automotive applications of hybrid systems”	Month 12
D4c.4.2	Proc. of the HYCON & CEmACS Workshop on Automotive Systems and Control	Month 21
D4c.4.3	Projects on automotive applications of hybrid systems and dissemination activities	Month 24
D4c.4.4	Projects on automotive applications of hybrid systems and dissemination activities	Month 39