TRAINALL 5th Newsletter

Transcription

1 October 2009 Integrated system for driver TRaining and Assessment using Interactive education tools and New training curricula for ALL modes of road transport Editorial Special points of interest: FERSI TRAIN-ALL Seminar. Simulator and Training Guidelines for Simulator Selection Simulators status Future Prospects Welcome to the 5th TRAIN-ALL Newsletter. TRAIN-ALL is approaching at the finishing line as a European funded project, with many interesting outcomes becoming step-by-step available to the Consortium partners and the Driving Simulators community. In this issue special interest is paid to these outcomes as they were presented at the Training Seminar on Simulator based training and Research, organ- ized by TRAN-ALL and FERSI (Forum of European Road Safety Institutes in Thessaloniki on September 14-15, This event was coupled with the TRAIN-ALL Final Workshop and Demonstrations taking place on September 15-16, 2009 at the same place, while the project partners met for the last plenary sessions on September 17-18, 2009 to ensure the timely completion of the re- TRAIN-ALL Week in Thessaloniki Inside this issue: Editorial 1 TRAIN-ALL Week in Thessaloniki 1 Simulators and Training 2 Guidelines for Simulator Selection 3 Motorcycle Simulators 4 Virtual Reality (VR)/ Augmented Reality (AR) Simulators 4 Truck Simulators 5 Emergency Vehicle Simulators 6 Future Prospect s 6 September was a busy month for the TRAIN ALL Consortium members; not only did the partners have a tremendous amount of work for finalising the remaining tasks towards the project end, they were actively engaged in delivering keynote speeches for the FERSI - TRAIN-ALL seminar & for the final TRAIN-ALL Workshop. In detail, the third week of September was totally devoted to TRAIN-ALL featuring: - the FERSI and TRAIN-ALL seminar on Simulator - based Training and Research: current driving simulator technologies, tools and scenarios, adequate for training all types of drivers and riders were presented. The seminar lasted two full days (14-15/09/2009) covering all major issues regarding simulation training. - the TRAIN-ALL 2nd PanEuropean Workshop: it started in the afternoon of September 14, 2009 with the demonstration of the project developments and continued with presentations and discussions on the project outcomes the following day. The above events were a great success, with 75 participants Project Coordinator TRAIN-ALL Copyright Dr. Mary Panou Created by the Project Technical Manager Institute of Communication and Dr. Evangelos Bekiaris Computer Systems (ICCS) Centre for Research & Technology Hellas / Hellenic Institute of Transport TRAIN-ALL Dissemination Manager: maining tasks for the finalization of the project. In this Newsletter you will find a rather detailed overview of the most interesting topics addressed and discussed at the FERSI supported Seminar by the TRAIN-ALL speakers. The final Newsletter will be published at the end of the project (December 2009), focusing on evaluation outcomes with real users, based on the tests carried out in 8 European cities. (55 of which are external to the project). Nearly all key manufacturers in the area attended, in addition to those included in the project Consortium, namely FAROS, Autosim, Virage simulation, OKTAL, KRAUSMAFEI! Also Presidents of the Driving school Associations from Sweden and Finland, many members of the Board of the Greek Driving Association, over 20 driving instructors and police representatives attended. - the TRAIN-ALL final plenary meeting (on 17-18/09/09) gathering for the last time all partners, aiming at finalizing the remaining tasks and deliverables of the project. Dr. Angelos Amditis (a.amditis@iccs.gr) {mpanou, abek}@certh.gr, url: Co funded by the European Commission

2 Page 2 Simulators and Training Driving is a daily activity that given the circumstances can be considered from very complex and demanding, which might implicate safety issues arising, to boring and simple. It cannot be doubted that it requires dedicated training and licensing for all levels of drivers, starting from novice vehicle and motorcycle drivers and going further to professional drivers. There is a rich literature over the years about the goals for driver education, the anticipated driver abilities, the tasks to be mastered and how all these correspond to the actual driver training and licensing. The situation becomes even more complex, when taking into consideration that traffic participants are a heterogeneous population and so are their training needs, but there are still commonalities. These commonalities call for the use of new modular tools and the establishment of new integrated curricula for their training. Simulation training presents a number of characteristics that turn it in a very appealing tool not only for researchers but also for driving schools: a large number of scenarios, traffic conditions and circumstances (even hazardous ones) can be simulated, driver behavior can be accessed and analyzed any time, feedback can be also provided and as technology advances even more tools will be provided for the instructors to assist them in their tasks. But how did everything start? The first simulators were invented for flying purposes back in 1910, while the first electronic ones date back to the 1940s. Evidence of the first driving simulators used for group training in driving schools can be placed in 1960, while the first ones featuring video scenes and landscape modeling were commercially used in the 1970s. A decade later the first dynamic driving simulators featuring motion base and wide video projection appeared in research laboratories. The past years simulators have evolved to state-of-the-art installations that often substitute the manufacturing of car prototypes (i.e. the Toyota Advanced Driving Simulator). The current requirements for simulator based training as derived by school owners and researchers opt for: - A training curricula that integrate driving simulators as training tools - Scenarios relevant for different phases in the training curricula - Instructions to the learner driver for each scenario - Criteria of correct behaviour for the assessment - Accurate data for the criteria based assessment - Feedback based on deviation from correct behavior - Tools for the instructor to supervise and assess the learner driver - Tools to keep track of training progress Based on those needs has been developing its modules and simulators.

3 Page 3 Guidelines for Simulator Selection One of the main problems when it comes to selecting/ purchasing a simulator is what to look for, depending on the use it is aimed for. Since this event was not only aimed at researchers, manufacturers and academia, but also driving school representatives, it was essential to address this issue. The success level of training depends highly on the scenario selection and curricula (the order of tasks, the duration of the sessions, the sequence of the sessions etc). Different scenarios are needed depending on the user group to be trained as the needs per user group (novice drivers, elderly drivers, professional drivers, motorcyclists and emergency vehicle drivers) are very different. The scenarios are, however, strongly linked to the technical requirements of each simulator. For example scenarios involving high speed on a motorway need motion simulation of at least the longitudinal axis and small FOV, while scenarios on urban road need motion simulation of at least the lateral axis and wide FOV. Bearing that in mind as well as the fact that all levels of the GADGET matrix should be covered, the simulator requirements are crucial in order to study the complete driving tasks and understand when the driver's reaction and manoeuvrability of the vehicle play an important role in the system. Therefore, for such types of simulators it is necessary to have a certain design approach, which should include the following characteristics: The simulator should work in real time; The descriptive model of the vehicle should be complete, and all the subsystems like the motor, controls, transmission, steering system, suspensions, etc. should be modeled; The descriptive model of the characteristics of the vehicle should be able to represent a wide spectrum of handling situations; The model of the vehicle should also be able to reproduce the effects between the vehicle and the highway, e.g. sliding surfaces or highways with gravel; The simulator should have a visualization system with a wide angle of vision; The simulator should have a dynamic base, for the simulation of the inertial forces. It is crucial that the time of delay introduced by the simu- lator is short, compared with the delays that are given in real vehicles ( ms). That is to say, the driver should not experience any unnatural delay between a certain manoeuvre and the corresponding change in the visual scene. The main selection criteria for a simulator, which is appropriate for driving training, that one should look for is primarily sufficient reliability to represent each traffic scenario, then possibility for upgrade within reasonable cost, of course a high costbenefit ratio, not causing simulation sickness to the drivers and of course proper scenarios for the specific driver groups, to which a school is aiming. Specific guidelines were presented regarding the visualization system (resolution, display time, FOV, frame rate), the controls as real vehicle controls are necessary-, the vehicle models and their parameters, the motion system and even the sound systems that are necessary. The session concluded with the reminder that simulators should and must be used in addition to road training and not as a substitute, to replace specific parts of the training (e.g. introduction to gear shifting). Simulators can be both great and dangerous tools, in particular due to wrong usage or selection. Final Project event: visit to CERTH /HIT Simulator & demo of the TRAIN-ALL modules (Thessaloniki, September 09)

4 Page 4 Motorcycle Simulators Motorcycle drivers are often involved in hazardous accidents, as traffic scanning errors prevalence is very high among them. A large number of motorcycle drivers have no experience on using ABS and ASR. Due to the nature of motorcycle driving, bend negotiation, positioning, overtaking as well as filtering require specified training such as for example for steering and countersteering techniques. Though the design of riding simulators still poses a number of challenges to be solved: ers practices large to be simulated difficulty in rendering of the roll physics are almost impossible to render because of kinesthetic and visual deficiencies ( a c c el e ra t io n a n d breaking, resolution, luminance etc.) visual rendering causes delays, latencies, lack of accuracy simulator sickness is always an open question large variability in rid- For the time being their main usage lies in virtual prototyping of vehicles and devices, to simulate ergonomics and vehicle dynamics, ECA FAROS EF-Bike University of Padova while it is also used for human factor experiments and studies when riding. It is gradually being embedded in riding training, however there are still underlying issues such as finding a proper solution to the steering and counter-steering problem, the motion complexity, use of the simulator within the curriculum, combination of real and virtual riding and of course the price. Within TRAIN-ALL INRETS has been working successfully to advance fundamental research and work on riding simulators by improving for example the steering/countersteering issue through pressure sensors and also by developing and testing a number of the TRAIN-ALL modules and embedding them to their riding simulator. Virtual Reality (VR)/ Augmented Reality (AR) Simulators VR/AR Simulators are based on the Virtual Reality technology, where two pictures - one for the left and one for the right eye - are created, superimposed, so for our brain to turn the difference into depth. VR has a number of characteristics that turn it into a very promising technology when it comes to driving simulators: realtime, interactivity, multi- modal interaction and immersion. Augmented reality on the other hand is an advanced derivate of VR, where images of reality are superimposed with virtual objects and the user can interact with the real world being supported by the information or images superimposed by the computer. Many vehicle manufacturers are using and evolving VR/AR Simulators (e.g. AUDI, VW, CRF, Peugeot, Renault, Toyota). These systems were launched to cover needs relevant to virtual prototyping, vehicle architectures, ergonomics and human factors studies, such as for example to test safely driver assistance functions in critical yet reproducible traffic situations. The future in this type of simula-

5 tors lies in the development of VR simulators for other types of vehicles, such as agricul- Page 5 tural vehicles and earth moving machinery, as well as integration of realtime photo- realistic rendering for quality evaluation of vehicle s interiors. which adds up to 500,000 hours of training, 12,5 25 billions of kilometers and 2million tons of carbon dioxide emitted. Therefore, there is a potential for the truck/bus simulator market to grow, so as to cover the training needs. Some of the companies worldwide that are engaged in producing such simulators are DORON precision system, ECA FAROS, INDRA, AUTOSIM, OKTAL SE, CORYS T.E.S.S., RHEINMETALL Defence and of course THALES, which is member of the TRAIN-ALL Consortium. Thales has developed the TRUST simulator, which enables efficient training in realistic road and urban environments, where all normal and dangerous traffic and weather conditions can be simulated and reproduced continuously until the relevant skills and reactions are attained by the trainees. TRUST features among others an inhouse developed hexapod system to take into account the different weights of the simulator on-board elements, real driving gears, steering wheel reaction as well as pedals and gear lever to enhance simulation quality. Truck Simulators Truck/ bus simulators are not only extremely developed today, but they are also mobile, as there can be a simulator set -up inside a truck travelling around. The essential purpose of truck simulator training (and in general simulation training) is to reduce road accidents, maneuvering damages and to allow economical driving with respect to the environment and transportation costs. In France alone 125,000 truck drivers (including bus drivers and public organizations truck drivers) per year out of a total of 620,000 truck drivers are involved in continued training

6 Page 6 Emergency Vehicle Simulators Since an emergency drive includes much more than standard driving tasks and it is considered very dangerous to be practiced in public road traffic, police cadets are in need of alternative ways of practicing and learning how to master cooperative driving in hazardous and stressful situations. Simulation has proven to be a solution to these problems, as there is the possibility of including all components of emergency driving into the training sequences. This fact coupled with the possibility to repeat training sequences, training of critical traffic situa- tions and individualization increases the quality and efficiency of police officers, when they are called to handle complex situations in real traffic. Police officers of the Bavarian Police follow a training programme that combines simulator and computer-based training with practical driving exercises and group training, so as to master a number of capabilities: danger recognition, orienteering and navigating while communicating on the radio, tasks division in the car etc. What is essential for this type of training, is to be able to train also the co-driver within the same simulation. Given the problems arising when a second individual is placed in the simulator (inadequate visualization, motion sickness etc), the co-driver is placed outside the cabin at a second screen viewing and interacting with the driver in realtime, while the instructor can monitor both simultaneously. Results in BPP have shown not only greater acceptance by the trainees but also a verifiable significant learning success, as significant increase of safe driving strategies was reported. exists. Further to that, the future of simulation is said to lie in distributed simulations, where an on-line/off-line exchange of data between the physical and the virtual world shall exist. Real geographic traffic and driving data will be integrated in the simulation, as well as real vehicles interiors. Another major problem in simulators penetration in the market is the lack of a standard architecture, which leads to highly customized solutions without any interoperability and thus to increased costs and to limited choices by the customers. Therefore standardization of the architecture, road networks, environment generations are foreseen to be the next steps. Could there be other uses for the simulators such as infrastructure auditing for example for tunnels? What is more, research should focus more also on simulator sickness aversion as well as on the long term traffic safety effects of simulator training. Future Prospects By the end of the seminar, based on the presentations and the discussions, the vital question was raised: What shall happen next and what are the customers to expect by the simulator developers and companies? One big issue which is under research concerns simulator validity. In other words researchers are trying to figure to which extent a standard correlation between the most significant driving behaviour parameters in real traffic and driving simulation conditions

7 October 2009 Integrated system for driver TRaining and Assessment using Interactive education tools and New training curricula for ALL modes of road transport P r o j e c t Co o r d i n a t o r Dr. Mary Panou Centre for Research & Technology Hellas / Hellenic Institute of Transport mpanou@certh.gr, url: P ro je c t Te chni c al Manage r Dr. Evangelos Bekiaris Centre for Research & Technology Hellas / Hellenic Institute of Transport abek@certh.gr, url: Created by ICCS T his proj ec t is Co- funded by th e Europea n Commi ssion und er the 6th F ram ework Pro gramm e. F or m ore inf ormation v isit the pro jec t s we bsite: TRAIN-ALL Copyright This newsletter was created by the Institute of Communication and Computer Systems (ICCS) TRAIN-ALL Dissemination Manager: Dr. Angelos Amditis (a.amditis@iccs.gr)