Using Embedded Systems to Improve Performance of Assessment in Virtual Reality Training Environments. Page 1. Using Embedded Systems to Improve Performance of Assessment in Virtual Reality Training Environments Ronei Marcos de Moraes. Liliane dos Santos Machado. Abstract Several techniques for training assessment in virtual reality environments have computational complexity incompatible with commodity computers performance. Even advanced computers could be unable to process, simultaneously and in real- time, 3. D stereo graphics, haptical tasks, users interactions and assessment system.
Advanced programming techniques, using specific hardware, can decrease costs of processing, allowing the utilization of low end computers. However, such strategies can be insufficient to guarantee the main features of virtual reality systems.
In this paper it is proposed the use of embedded systems to compute specific tasks, to prevent those problems and allowing the use of low end computers. Consequently, embedded systems could contribute to minimize the lose of information and improve performance of assessment in virtual reality training environments. Index Terms Embedded Systems, Virtual Reality, Training Assessment, Computational Performance. Ronei Marcos de Moraes, Department of Statistics, Federal University of Paraíba. Cidade Universitária s/n, Castelo Branco, João Pessoa- PB. Brazil. ronei@de. Liliane dos Santos Machado, Department of Informatics, Federal University of Paraíba.
Cidade Universitária s/n, Castelo Branco, João Pessoa- PB, Brazil. INTRODUCTION Military war games, problem fixes in outside space, rescue operations into risk areas and medical surgeries are examples of tasks in which there is risk for human being.
Realistic environments based on virtual reality have been developed to immerge users into a virtual world, in which critical situations can be simulated. Due to cost reduction, security and facilities of execution, several kinds of training are already performed in these virtual reality environments. In general that simulators demand high performance computers to provide realistic stereoscopic visualization, deformation, textures, etc. However, it is important to know users performance in those training in order to correct deficiencies and improve skills and psychological conditions to execute the real task. In several kinds of training in the medical area expert physicians assess trainees through videotapes. Such modality characterizes an offline assessment. In this kind of assessment the trainee will receive a feedback with critics and suggestions only some days after the training.
2002 TimeSys Corporation. The Concise Handbook of Real-Time Systems [EB/OL].[2005-06-20]. Sha L, Lehoczky J P. A Periodic Task Scheduling for hard Real-Time Systems[J].The Journal of Real-Time Systems, 1989,1: 27–60. The Concise Handbook Of Real-Time Systems TimeSys Corporation Version 1.3 ©2002 TimeSys Corporation Pittsburgh, PA www.timesys.com. We consider pipelined real-time systems that consist of a chain of tasks executing on a distributed platform. The processing of the tasks is pipelined: each processor executes only one interval of consecutive tasks.
This fact offers a great inconvenience for his learning, because the user will not remember several details of that training. The first papers in training assessment probably belongs to Dinsmore and his collaborators [1. They proposed a virtual reality environment for detecting subsurface tumors in which the trainee is assessed by a training quiz. The trainee is asked to identify the location and hardness of tumors and to provide a diagnosis.
. simultaneously and in real-time, 3D stereo graphics. has tactile properties similar to real tissues. Assessment systems can be used coupled to any virtual. TimeSys Co.; The Concise Handbook of Linux for Embedded. The Concise Handbook Of Real-Time Systems TimeSys Corporation Real-Time… Real SolutionsTM Version 1.1 ©2000 TimeSys Corporation www.timesys.com. The Concise Handbook Of Linux for Embedded Real-Time Systems TimeSys Corporation Version 1.1 1-888-432-TIME ©2002 TimeSys Corporation Pittsburgh, PA www.timesys.com.
Similarly, Wilson et al. MIST) in which each task could be programmed to deliver several degrees of difficulty. The trainee’s performance could be recorded and saved for later replay to be used by the supervisor or for statistical analysis. In parallel, the assessment of surgical skills has been subject of several authors. Some of them use statistical methods to create models .
TimeSys Corporation. The concise handbook of Linux for embedded real-time systems. Proceedings of the 11th Euromicro Conference on Real Time Systems. 1999. A group priority earliest deadline first scheduling algorithm. REAL-TIME SYSTEMS. Misconceptions about. The Concise Handbook of Real-Time Systems, version 1.3, Timesys Corporation Real-Time and Embedded Systems.
Others use statistical methods to show that virtual reality systems could help to distinguish between experienced surgeons and inexperienced surgeons [4. It was shown also that surgeons trained using virtual reality can obtain better results [1. The assessment of psychomotor skills inside a virtual reality simulator using haptic devices can quantify surgical dexterity with advantages over traditional subjective evaluation, as provide unbiased and objective measurement of surgical precision . For these attributes, it was presented the assessment of psychomotor abilities as part of the future of medical teaching and training [2. In this decade, the research area of training assessment received new impulses, where could be mentioned relevant papers of Rosen et al. Machado et al. [2.
Linux/RT is a tool for developing and testing embedded and non-embedded real-time applications. TimeSys. TimeSys Corporation. E-Mail: info@timesys. The booklet titled “The Concise Handbook of Real-Time Systems. The Concise Handbook Of Real-Time Systems : / Version 1.3 / TimeSys Corporation. 14th IEEE International Conference on Embedded and Real-time Computing Systems and Applications. Newnes electrical engineer's handbook /.
Rosen et al. [4. 4] and Moraes and Machado [2. In those works, the assessment system used Hidden Markov Models and rule- based fuzzy expert systems. Additionally, it was defined two modalities of assessment systems: the traditional offline, in which the information collected during the training are analyzed later, and the online, in which the assessment is accomplished immediately at the end of the simulation. After 2. 00. 3, online assessment received several contributions and new techniques proposed by Moraes and Machado: Gaussian Mixture Models [2.
Fuzzy Gaussian Mixture Models [2. Maximum Likelihood [3. Fuzzy Bayes Rule [3.
Nowadays, they proposed also assessment models by Internet [3. INTERTECH March 0.
São Paulo, BRAZILInternational Conference on Engineering and Technology Education. Page 2continuous assessment [3. However, advanced techniques that could be used as basis of assessment systems have high computational complexity and are incompatible computers. Even medium port computers could be unable to simultaneously run in real time 3. D stereo graphics, haptic loops, users interactions and assessment systems . Advanced techniques of programming combined to dedicated hardware could result in lower costs of computers and processing, allowing the use of low end computer systems.
Thus, in this paper it is proposed the use of embedded systems [4. This approach is justified by the increase of complexity of training systems based on virtual reality in which it is necessary to monitor simultaneously a large number of variables.
METHODOLOGY Methods for Training Assessment Advances of training systems based on virtual reality demanded methods to assess users to improve psychomotor skills, psychological conditions and to correct deficiencies in order to prevent mistakes when they execute a real task. As an illustrative example of that necessity [2. Figure 1 shows a simulator to training new doctors to execute the bone marrow harvest, one of the stages of the bone marrow transplant.
FIGURE. 1 THE VIRTUAL REALITY BASED SIMULATOR FOR BONE MARROW HARVEST TRAINING IN USE. The procedure is done blindly, performed without any visual feedback, except the external view of the donor body, and the physician needs to feel the skin and bone layers trespassed by the needle to find the bone marrow and then start the material aspiration. The simulator uses a robotic arm that operates with six degrees of freedom movements and provides force feedback to give to the user the tactile sensations during the penetration of the patient’s body [2. In the system the robotic arm simulates the needle used in the real procedure and the virtual body visually represented has tactile properties similar to real tissues. Assessment systems can be used coupled to any virtual simulator. In particular, in medical area, invasive procedures simulated in virtual reality can easely demonstrate the advantages and justify the necessity of assessment systems [2.
Assessment system can offer to trainee a report about his performance during the training and with this results, the trainee can improve his dexterity. The assessment is proceeded by the comparison of patterns with user’s parameters obtained during the execution of training. In order to do that, patterns are previously stored by experts, according to M classes of performance. The number of classes is defined by experts, which also supplies their meanings. For example, for M=3 classes of performance, the meanings could be: 1) correct procedures, 2) acceptable procedures, 3) badly executed procedures. So, the classes of performance for a trainee could be: "you are well qualified", "you need some training yet", "you need more training". However, that number can be defined according to the application.
The assessment subsystem must provide also a measure of adaptation of trainee performance for a specific class or for all the possible classes of performance. FIGURE. 2 ARCHITECTURE OF AN ASSESSMENT SYSTEM FOR TRAINING BASED ON VIRTUAL REALITY. ADAPTED FROM MORAES AND MACHADO [3. A training simulator based on virtual reality and its assessment subsystem are interdependent subsystems, which works simultaneously. All subsystems of a simulator (visualization and interaction) and assessment subsystem works together inside of a same computer system. Thus, general performance of simulation must guarantee always to run in real time .
Then, the assessment subsystem must obey an important restriction: it must work without compromise the simulator performance and its realism. This way, the research of assessment systems for complex training, in which it is necessary to monitor a large number of variables simultaneously, is strongly dependent of four factors: a) computational complexity of the virtual reality © 2. INTERTECH March 0.
São Paulo, BRAZILInternational Conference on Engineering and Technology Education. Page 3environment; b) computational complexity of the assessment method; c) accuracy of method, once validated its basic assumptions for the specific problem and c) the computer system available to execute training environment and assessment subsystem. The balance of that four factors is not easy to solve. Very realistic virtual reality environments demand high computational costs, as well as may demand interactions using haptic devices and the use of deformable models. These requirements have priority over other subsystems as the assessment. A solution is use methods of low computational cost for assessment. Unfortunately, this solution can cause negative impact over assessment subsystem accuracy.