The whole idea of training is based on the fact that the trainee dives into the virtual world that copies his actual working environment. This foresees the following modeling requirements:

• image quality allows observing current status of an object without drawing much attention to the object itself. This allows the existent decision-making algorithm to be worked out and life-like experience to be gained
• the quality of virtual environment rendering should create participation effect, as if the trainee is placed in real environment

The learning process is built on the usage of realistic 3D simulation of processes, objects, environment, and interaction between all of them:

• the functionality of each object in the environment depends on the way how each participant will interact with it
• the complexity of simulation depends on the task and vary from simple interaction with single objects to realistic simulation of work of complex technology equipment
• 3d simulation may be based on mathematic model (that is scientifically grounded), if needed; model type is chosen by customer

This type of training means that the trainee specifically develops a formalized interaction procedure with certain object environment. The whole procedure can be described by scenario that includes the following details:

• scenario includes distinct sequence of actions to be taken by each trainee
• scenario sets time intervals for every interaction with each object
• it may foresee the priorities of different variants of actions
• it supposes that all actions end up with a specific event
• candidate's performance is evaluated according to the correspondence of his actions to the pre-set educational scenario

Our product supports up to 250 users working simultaneously. This means that you can train the whole team and improve team work. Its key features include:

• orientation on cooperative actions of the students
• role assignments for each participant (or group)
• personal identification system for each trainee, which makes him unique and recognizable
• built-in communication module: each participant can communicate with any other during the training
• evaluation results for each trainee depends on the consistency of the team work

In order to get experience close to the hand-on skills, trainees needs to have all kinds of interaction with all types of objects, that they can come across during the real work. Therefore the level of interaction 3D environment will include:

• all objects necessary for the teaching course
• bots with certain level of AI (Artificial Intelligence), which imitate people not involved into training process (visitors, passengers, etc.)
• characters controlled by other participants

To increase the efficiency of virtual 3D trainings there is a need to implement the following three stages of training:

• introduction: the participant passes the whole training in a demo-mode
• learning: the trainee goes through the whole set of training tasks with the help of tips
• testing: the trainee passes the whole training without any support and will be evaluated according to the formalized criteria

During the training two parallel processes are happening: ongoing analysis and the analysis based on the results of the final testing. This makes the process of training a controlled one – and enables objective evaluation of the results:

• all intermediate information about results is stored in a database on the server and can be compared (such approach can show the dynamics of the training progress for each participant)
• possibility to compare the results of different trainees (allows to assess the group dynamics)
• tools to analyze current results allows to implement some task changes directly into the training process
• real-time results monitoring for the trainers/supervisor
• priorities-based results analysis for getting the final score and following permission for real work or transfer to the stage of practical training