Research Project SASV

Variable Structure Automatic Systems

Project Name: Variable Structure Automatic Systems Dean of the SASV group: Nacer Kouider M'Sirdi Research Projet 2012-2015

1. Introduction

The VSAS project started at the end of 2008, for fundamental research in automatic control and optimization for the Variable Structure Systems with commutations. The considered class of systems (VSS) is multi components (using multiple connected models with commutations) involving exchanges between several parts of the system and the environment. The figure below shows a simple example of a Variable Structure System with model commutations. The table or a vehicle chassis is composed by various sub-models depending on which end point of the system is in contact with the ground. The posture motion involve models commutations (depending on the system state evolution), energy exchanges with environment occur.

The SASV group knowledge in automatic control, robotics field, computer-integrated manufacturing, signal processing, optimization and mechatronics are exploited in several innovating industrial applications, where optimization, respect of the environment, cost reduction and energy saving are the main targets. These last decades, the applied research in automatic control for vehicle dynamics, robotics, computer-integrated manufacturing or for energy and sustainable development is largely studied, especially for safety aspects, user’s comfort, energy saving, performances and driving assistance. The technologies progress and recent knowledge offer new solutions (automatic or assisted) to improve the design of tools such as the monitoring or control systems. The requirement increase and the performances need advanced nonlinear control techniques. Systems Analysis and robust estimation of parameters or variables, robust observations and control are studied and extended for the needs for innovating applications.

2. Research Topics

Variable Structure Automatic Systems are made of several sub systems (systems of systems) with time or state depending commutations. Modelling, Observation and control are open problems to be studied, for this case according to the applications considered. The variations in the structure and commutations imply exchanges between the various parts of the system and the environment. VSAS is interested in: • Analysis and design of robust methods to estimate states, parameters and unknown inputs, • Control of VSS with commutation which are often met in vehicle dynamics and Renewable Energies, • Intelligent control of robotic systems, • Optimization and management of multi energy systems • Optimization and control of the industrial production of manufactured goods and services. Two main research orientations are considered: 1) State estimation, testing, analysis and diagnosis 2) Control, monitoring and optimization. 1) Sate Analysis, observation and testing (of the system and its environment) to lead to automatic recognition of operating situations (embedded estimators, assistance system, diagnosis, …). - the estimation of the interfaces variables for a vehicle, for car comfort analysis and safety enhancement, - on line optimization of system operation also if commutations or structure changes appear,… - automatic recognition of situation, inboard diagnosis and possibly decision-making aid

 Open  Research  Topics:  State  estimation  in  VSAS,  observation  and  identification  for  Variable  Structure  Systems,   Interactions  with  other  systems  with  unknown  dynamics.     The  main  interest  is  to  buid  a  methodology  for  the  state  estimation,  analysis,  running  situation  recognition  and  optimal   control  for  VSAS  and  production  systems.   2) The system control, assistance, diagnosis and monitoring, of complex systems and their interactions with the environment (safety, optimal operation, performance). In complex systems, several sub-systems interact. The main interest is to control the system in presence of the interactions and exchanges with the environment. In a vehicle or a drone or production systems, for example, several

1

control entities can coexist (in addition to the human pilot). How can be handled tasks and how can priorities be shared? The objective is to show that this can operate without any hierarchy, when integrating decision-making, by means of multicriteria analysis.

3. Research Activities

These last decades, the research and the application of control and industrial engineering in the fields of motor cars, robotics, industrial engineering, isoarchic control systems or for energy and sustainable development have been largely studied, especially for safety aspects, user’s comfort, energy saving, performance and control aid. The development in technology and current knowledge offers new solutions (automatic or assisted) to improve or modify the design of tools such as the monitoring or control systems, use of multicriteria decision aid. The requirement increase and the necessary performances need advanced control techniques (nonlinear, robust, adaptive). The system analysis, parameters and states estimations, observations and robust control are studied and extended to the needs for innovating applications. Our research highlights several prospects for applications to control engineering and diagnosis with robotics, motor cars, mobile machines and autonomous systems. Observers and sliding mode control for electric machines, energy and production systems Vehicle Dynamics and transport systems, Driving simulators and remote control Complex Systems in Robotics and mobile machines, Multi physics and multi energies Systems, with Variable Structure, commutations and in presence of random events… o Monitoring and security systems o Production systems with multi-actors and variable composition network (Flexible Manufacturing Systems, Intelligent Manufacturing Systems, Holonic Manufacturing Systems, and Intelligent Supply Chain). We highlight Several prospects in robotic applications of automatic control and diagnosis like mobile robots and autonomous systems as well as in manufacturing systems. The approach thus aims to integrate intelligent elements for control and decision making to get `plug and run' which allows the adaptation to the variability of the SASV. o o o o

Key words: Models of complex systems, multi model, Variable structure, Optimization, Stability and control of variable structure Systems, Identification of VSAS and Sliding Modes, Methods and tools of observation and VSAS control, Manufacturing System, Control of Isoarchic systems, Holonic systems, Multicriteria Decision.

4. Application domains o

The main applications in SASV are: The control of the vehicles to enhance safety,

o

Optimization and control of systems with multiple energy sources, applications to small buildings, smart buildings,

o

Control and optimization of production systems (goods and services), with intelligent entities.

We contribute to the improvement of safety and comfort (as far as possible) of vehicles dynamics and new transport systems (electric vehicles and hybrids). All-terrain vehicles, such as tractors, agricultural machines and tools in general (for which mobility is not always the essential function). The ActiSurTT ANR project is dedicated to the improvement of safety for all-terrain vehicles (ATV) such as agricultural machinery and in particular tractors. A new Tire - Ground contact model is developed and control assistance is proposed for safety improvement for ATV. The control systems use the concept of isoarchy, and the holonic paradigm. Its design use multicriteria methods by complete aggregation allowing decision making. This is obtained through AHP (Analytic Hierarchy Process) and ANP (Analytic Network Process).

5. Perspectives

The all-terrain vehicles (ATV) and mobile robots or agricultural heavy machines take an important place in the prospects of our activities. We will endeavour in an immediate future to integrate our proposals in mechatronic systems (multidisciplinary) to prove the feasibility of effective onboard of security systems. Diagnostic tools and decision-

2

making aid will be integrated into the embedded system. The extension of the cooperative project, on multi energy sources systems for building, in a Mediterranean and European project is in progress. http://rmei.info/index.php?option=com_content&task=view&id=194&Itemid=1 A study for an application to tractors has been initiated and is supported by an ANR project which will last for 3 years starting from the beginning of 2011. The project is called ActiSurTT, whose title is Active safety system for ("off road" vehicles) ATVs. As far as multi energy sources systems are concerned, we are involved in the assembly of a network of co-operation on a Mediterranean scale (project in progress). As for control of production systems, in addition to the continuation of the partnership research projects in progress (InterVascular, Antalios), various projects are being set up, as FUI or as direct partnerships (CIFRE) or international projects (Canada, GB).

6. Experimental setups in SASV

We use a driving simulator, a wind test bed and a production platform to validate our propositions.

Driving Simulator Vehicles Scaner Studio

For the experimental validation of our observers, driving and diagnosis aids systems, we use the environment proposed by Oktal Scaner Studio (www.oktal.fr). This simulator, which is now the reference car manufacturers Renault and PSA among others, allows to study and develop active devices for the safety of vehicles in all-terrain environment.

It will be used to host the simulator of the harvester tractor studied in ANR ActiSurTT. This 36-month project began in the middle of December 2010. This simulator can also be used for prototyping any integrated function into a vehicle.

Test bench Renewable Energy: Wind – Solar A test bench consisting of a 2-bladed wind cone, of a data acquisition system and of data processing, is available at LSIS-Marseilles. Simulation tools (DSP-D space) complement this material to allow the study of design and virtual prototyping. The generator is simulated with Matlab Simulik. We shall associate to this system, photovoltaic panels and an energy storage system of electrical type.

3

4

Versatile Production Platform The Versatile Production Platform is a research tool that reproduces all the features of a modern manufacturing factory: it is a VSAS. It allows the investigation of all emerging problems in the field and related to emerging production control systems modes. This factory is built around a production system incorporating all current technologies (transitic, robotics, storage and retrieval machine, RFID traceability ...) associated with an integrated management system of type ERP (Enterprise Resource Planning). The production system is organized around a transitic Montrac Chaos system allowing to manage automatically and flexibly all product flow and all production management possible modes. The factory is designed to produce small objects. For this, 11 workstations are distributed around three transportation loops served by self-propelled Montrac shuttles. All production operations are subject to close monitoring obtained by reading the RFID tags attached to products and containers moving between the different workstations. The chosen ERP, to support the information and management system, is Sage X3. The Versatile Production Platform, in the core of the holonic paradigm, allows the integration of mechatronics and infotronic technologies, that allow to control the variability of VSAS. This allows to consider an approach of 'ambient automatic’ type. For example, the validation of the holonic, isoarchic and multicriteria control approach, will be done, on one hand, by experiments using a distributed simulation environment, and on other hand, in interaction with the real system 'in the loop', with operators. In the two cases, the focus will be on the system performance evaluation.