CZ308105B6 - Method of evaluating the positive transfer of training on a simulator and equipment for it. - Google Patents

Abstract

A method of evaluating the positive transfer of training on a simulator, in which the reference person carries out the prescribed actions on the simulator and the reference person's interaction with the simulator is monitored and recorded, and a trained person carries out the same prescribed actions on the same simulator, the interaction of both persons with the simulator means is compared and the degree of consistency of the trained person's interactions with the reference person's interactions is determined. A reference dynamic model is determined from the interaction of the reference person with the simulator , a unique dynamic model is determined from the interaction of the trained person with the simulator, both dynamic models create Bode diagrams, which are then compared to determine the matching degree of the persons.The device for evaluating the positive transfer of training on the simulator comprises a motion system, a system of status and position sensors of the simulator controls (1) and a control system (2) adapted to carry out high-fidelity simulation functions. The control system (2) has a block (20) for projecting a predefined exercise task onto the instrument panel of the simulator (1), and also has a data logger (3) for recording and storing the simulation progress data. The device further comprises an evaluation unit (4) adapted to create a dynamic model of a person carrying out a predefined task on the simulator.

Description

Method of evaluation of the positive transfer of training on the simulator and equipment for its implementation

Technical field

The invention relates to a method of evaluating a positive transfer of training on a simulator, wherein the reference person performs prescribed actions on the simulator, during which the reference person's interaction with the simulator means is monitored and recorded, and the trained person performs the same prescribed actions on the same simulator while monitoring the interaction trained persons with simulator resources, whereupon the interaction of both persons with simulator resources is compared and the degree of consistency of the trained person's interactions with the reference person's interactions is determined.

The invention also relates to an apparatus for evaluating a positive transfer of training on a simulator comprising a motion system, a system of state and position sensors of the simulator controls and a control system adapted to perform high-fidelity simulation functions, the control system having a block for projecting a predefined exercise task. a simulator dashboard, and is further provided with a data logger for recording and storing simulation progress data, the apparatus further comprising an evaluation unit adapted to create a dynamic model of the person performing a predefined task on the simulator.

BACKGROUND OF THE INVENTION

Loss of control is currently the primary cause of catastrophic aviation accidents. Training based solely on exercise of attention and prevention proved to be inadequate. Faithful simulation, along with realistic cockpit processing, provides crews with an environment to practice contact with flight conditions during unusual flight cases, and can help enhance the learning process, primarily to prevent critical modes from entering and preventing them. Faithful simulation, however, requires not only a corresponding simulation device, the so-called simulator faithfully corresponding to the behavior of the simulated device, but also requires an assessment of the effectiveness of training, respectively. evaluation of the positive transfer of training to the trained person, ie the user of the simulator, evaluation of his / her progress in training and, where appropriate, evaluation of his / her abilities. It is required that this assessment be objective, ie devoid of subjective influences, which can be introduced into the assessment by a human factor in the form of an evaluator - instructor, etc.

Learning and evaluating learning outcomes are known from EP 3104360.

Although various approaches for assessing the positive transfer of training are known in the art, it appears that these techniques bring about a subjective human factor in the assessment, or do not have a sufficient predicative value, in particular in terms of clarity, demonstrability, comparability, etc.

The object of the invention is to eliminate or at least minimize the disadvantages of the prior art, in particular to increase the informative value, unambiguity, demonstrability, comparability, etc. of the automatic evaluation of the positive transfer of training.

SUMMARY OF THE INVENTION

The aim of the invention is achieved by the method of evaluation of the positive transfer of training, which is based on the fact that the interaction of the reference person with the simulator resources determines the reference dynamic model, the interaction of the trained person with the simulator resources determines a unique dynamic model. Bode diagrams, which are then compared to determine the degree of match between the reference person's interactions and the trained person's interactions.

- 1 GB 308105 B6

The essence of the device for evaluating the positive transfer of training on a simulator is that the evaluation unit is further adapted to generate Bode diagrams from dynamic human models and to compare these Bode diagrams in the frequency domain, wherein the control unit is further adapted to determine the Pearson correlation coefficient between comparing Bode diagrams in the frequency range from 0.1 to 10 rad / s.

The invention makes it possible to objectively evaluate the degree of conformity of a trained person with a reference person in controlling a particular device, which enables to replace responsibly and effectively the expensive, time-consuming and often complex training of the trained person directly on the device with training on a corresponding simulator. the ability of the trained person to control the actual equipment and, in particular, to assess the progress made by the trained person.

Clarification of drawings

The invention is schematically shown in the drawing, wherein FIG. 1 shows an example of spatial determination of a predefined flight trajectory exercise task; FIG. 2 an example of projecting a predefined flight trajectory exercise exercise on a primary flight display and evaluating the deviation Fig. 3 shows a diagram of the dynamic pilot model operation, Fig. 4 another example of spatial determination of a predefined flight trajectory observation exercise, Fig. 5 examples of the type and quality of signals to identify a dynamic pilot model, Fig. 6 Fig. 7 shows an embodiment of a system for carrying out the method according to the invention; Fig. 8 a sequence of operations to express the level of training of the pilot.

DETAILED DESCRIPTION OF THE INVENTION

The invention will be described by way of example in flight training on faithful simulation hardware in combination with sophisticated software for flight dynamics simulation, scene visualization, etc. Equivalently, the simulation can be used in ground vehicle training, etc.

In the considered flight training example, to evaluate the positive transfer of training, a comparative sample is first described that describes the performance of the trained pilot (instructor) on the simulator. The formation of this pilot trained comparator consists in providing the trained pilot with the task of performing a predefined flight trajectory tracking exercise on the simulator, for example as shown in Figures 1 and 4. This predefined training pilot exercise is a training flight, with relevant stimulation of stress and fatigue levels during the task. A typical pre-defined exercise task is an experimental flight simulator trajectory in a space where the trajectory contains a compact set of drill maneuvers such as take-off, climb, horizontal flight, horizontal bends at different heeling angles, ascending and descending bends, approach to landing, landing, etc. for simulated aircraft. Part of the definition of the training flight trajectory parameters is also the specification of flight variables (speed, height, course) at the entrance to individual partial maneuvers of the assembly, adding fast events (dynamic changes) to the trajectory, such as turbulence, etc. about the superposition of the dynamic change signal (fast action) on the basic trajectory of the maneuver. The pilot performs the task by displaying the image of the given trajectory and the intentional image of the position of the training aircraft represented by the simulator on the flight display in the simulator, eg on the primary flight display or head-up display. Typically, a simulated environment is projected onto a flight display with a point determining the desired position of the airplane (simulator), and the actual position of the airplane (simulator) is projected onto the flight display as a "sight pattern". By controlling the simulator controls, in this case, the control lever, foot pedals, balance controls, and engine control lever, the pilot attempts to minimize the visual deviation between the desired trajectory (the desired point) and the actual actual position of the airplane indicated by the "sight pattern". For example, this is realized on the primary

-2E 308105 B6 similar to the artificial flight horizon and the usual sign to show the actual position of the airplane relative to the artificial horizon. Alternatively, this can be displayed on the main display of the simulator, to which the image of the surroundings of the simulated airplane is otherwise projected. During the assigned task, the simulator reads data describing the reactions of the trained pilot during the task, eg forces acting on the simulator control system, position of simulator controls etc., possibly also the physical functions of the trained pilot, eg blood pressure, heart rate etc. This procedure gives a comparative sample that corresponds to the behavior of the trained pilot on the simulator, assuming that the trained pilot is capable of performing the tasks at a high level, with high accuracy, and is therefore expected to train pilots to achieve comparable results as already trained pilot.

Furthermore, in the considered flight training example, to evaluate the positive transfer of training, a current sample is created that describes the performance of the currently trained pilot on the same simulator. Creating this actual sample of the currently trained pilot is that the currently trained pilot is tasked to perform the same predefined flight trajectory tracking exercise on the simulator, as shown in Figures 1 and 4, as the pilot already trained and described in more detail above. The actual task is then performed by the currently trained pilot as well as the already trained pilot, while data describing reactions of the currently trained pilot during the task, such as forces acting on the simulator control system, position of simulator controls etc. pilot, such as blood pressure, heart rate, etc. This procedure obtains a current sample that corresponds to the behavior of the currently trained pilot on the simulator.

In essence, the above-described procedure creates a so-called dynamic model of a trained pilot (comparative sample) and a dynamic model of the currently trained pilot (actual sample), both on the same simulator, both dynamic models being acquired while performing the same task. exposed to the same external influences, eg visual, movement or haptic etc. The dynamic pilot model is basically a description of the pilot's reactions to external stimuli and the corresponding reactions of the pilot to these external stimuli are manifested by the pilot's influence on the simulator / airplane controls. simulator control lever positions, etc. Each dynamic pilot model is characterized by transmission function parameters that are recorded on the simulator while performing a predefined flight trajectory tracking exercise.

For the actual creation of dynamic model of trained pilot (comparative sample) and dynamic model of currently trained pilot is used model ARMAX (auto regressive, moving average with exogenous input), which defines parametric linear model with stochastic component, where it is advantageous to use linear, time invariant an approach that takes into account stochastic disturbances in the pilot's response to visual stimulus while performing a pre-defined exercise trajectory tracking task. The visual deviation is defined as the difference between the reference trajectory displayed in the pilot's field of view (head-up display or primary flight display) and the current position of the aircraft, as the input signal for the pilot model identification. FIG. 2. To generate consistent output data, a reference signal determined, e.g., by the tilt angle, is preferably used. To evaluate the fulfillment of the predefined flight trajectory observation exercise, the elevation of the rudder that is recorded on the simulator for this purpose during the fulfillment of the predefined flight trajectory observation exercise is considered as the output signal in the pilot model identification process. The prediction error method is then used to estimate the parameters of the ARMAX model based on the I / O data. The task of estimating the parameters of the ARMAX model is to correctly determine the order of polynomials of the transfer function, whose inaccurate determination could cause problems in model validation. Determination of the correct structure of the model is ensured by an optimization algorithm whose evaluation criterion is defined, for example, as the mean quadratic deviation of the model output and the measured input, for example the elevation of the rudder. For the lateral movement of the airplane are considered similar structures of the model, but with the outputs determined by the aileron and rudder deflection.

-3E 308105 B6

The result of creating a dynamic pilot model using the ARMAX model is then subjected to frequency domain analysis, which is shown in Figure 6 as the so-called Bode diagram. The bandwidth on the x-axis of the Bode diagram in Fig. 6 corresponds to the range of overall human perception of motion, which is a range of frequencies from 0.1 to 10 rad / s for a typical sports airplane pilot. Due to the possible superposition of the dynamic change signal to the basic maneuver carrier signal, a more accurate description of the pilot's dynamic model is achieved in frequency analysis.

For the evaluation of the positive transfer of training on the simulator, the degree of agreement between the Bode diagrams for the dynamic model of the trained pilot (comparative sample) and the dynamic model of the currently trained pilot is evaluated, for example using the Pearson correlation coefficient. The implemented evaluation function is described so that the correlation coefficient R represents the degree of linear correlation between two variables, where the value 1 represents total agreement and the value 0 represents no correlation. From this, it can be determined that the calculated correlation rates between the Bode diagrams for the dynamic trained pilot model (comparative sample) and the currently trained pilot dynamic model for their frequency domain transfer functions with a frequency range of 0.1 to 10 rad / s show the correlation rate between dynamic a pilot-trained model (comparative sample) and a dynamic pilot-trained model (actual sample), which expresses an objective mimic of the practical skills of the currently trained pilot in performing a predefined flight trajectory observation approach to the trained pilot in performing the same predefined training exercise the task of tracking the flight trajectory on the same simulator.

Claims (1)

An example of the degree of correlation between the two dynamic models is shown in the following table. VELIČINA THE CORRELATION COEFFICIENT AMPLITUDE 0.968 PHASE 0.993 Based on a positive correlation between the dynamic model of the trained pilot (comparative sample) and the dynamic model of the currently trained pilot obtained in the frequency domain while performing the same predefined flight trajectory tracking task on the same simulator, the degree of training of the currently trained pilot is reliably and objectively determined. The system for evaluating the positive transfer of training on the simulator in the illustrated embodiment of Fig. 7 comprises an airplane simulator 7 with a mechanical (direct) flight control system, ie an airplane without affecting the dynamic pilot model by active control elements such as flight-by-wire systems. ", Stabilization systems, etc. An example of such an airplane with a mechanical (direct) flight control system is a small sports airplane or training airplane, etc. The simulator 1 comprises a motion system, a system of status and position sensors of the simulator 1 controls, a small sports airplane dashboard, etc. The simulator 1 is further provided with a control system 2, eg a control computer, or a group of computers, with software for performing high-fidelity simulation of a sports airplane. The control system 2 is provided with a block 20 for projecting a predefined flight trajectory monitoring task to the instrument panel of the simulator 1, in particular to the main flight display or head-up display, etc. The simulator 1 is further provided with data logger 3 which records and stores progress data. simulation, the position and states of each simulator control over time, etc. Simulator 1 further comprises an evaluation unit 4, which is typically a computer with software to identify (create) a dynamic pilot model while performing a predefined flight trajectory tracking exercise and comparison of dynamic models, resp. their transmission functions, in the frequency domain, and to determine the Pearson correlation coefficient between the compared Bode diagrams of dynamic models of persons in the frequency domain from 0.1 to 10 rad / s. -4E 308105 B6 It is clear from the above that the method of the invention can be used to assess the positive transfer of training on a simulator of not only a small sport aircraft, but it can be adapted to virtually any type or type of simulator where a dynamic operator model can be identified. , the whole crew, helicopter pilot 5, etc. Industrial applicability The invention is useful in evaluating the training of persons on simulators of systems with a mechanical control circuit, especially in those areas of human activity, where training on real equipment, eg aircraft, helicopter, vehicle, etc., is financially and time consuming or possibly risky etc. .