Pilot alertness calculation model based on biological mathematical model

Abstract

The invention discloses a calculation model for an airline company to evaluate the alertness of a pilot. Starting from a three-process model, extension and expansion are carried out. The method comprises four parts, the influence of internal alertness S, sleep inertia W, a 24-hour phase process C and a 12-hour phase process U on the prediction effect, namely accuracy is comprehensively discussed, all components are summed to obtain a comprehensive alertness score, the comprehensive alertness score can be used for predicting the score of a Carlin Sca sleep scale, calculation is carried out under various combination conditions, and the prediction accuracy is improved. And the function of alertness prediction can be realized by inputting the working time arrangement of the personnel. Through comparison and consideration of various conditions, it is finally determined that the optimal model combination is S + C + U, and the optimal performance can be shown.

Description

Technical Field

[0001] The present invention is based on a biomathematical model, relates to the technical field of the aviation industry, and applies it to the scenario of pilot alertness assessment. Background Art

[0002] In modern aviation-related industries, the physiological state of pilots is crucial to the transportation process, among which sleepiness and fatigue are two key risk factors. Therefore, biomathematical models of sleep fatigue are becoming increasingly important in the fatigue risk management system (FRMS) of air transportation. Air transportation is the safest mode of transportation per kilometer. However, once an extremely rare traffic accident occurs, the consequences are often devastating. Accidents in the aviation industry rarely have a single cause, and most accidents are man-made disasters.

[0003] In road transport, several studies have shown a link between human error and fatigue. The main causes of sleepiness and fatigue are: 1) circadian phase; 2) duration of wakefulness; and 3) amount of previous sleep. In addition, task time may induce fatigue when sustained attention is involved. Individual differences are likely to play a role in accidents related to sleepiness and fatigue, driving performance, and affect performance during different sleep lengths and sleep deprivation experiments.

[0004] In the aviation industry as well, human errors and inappropriate decisions are also affected by sleepiness and fatigue. Overall, irregular working hours, working at inconvenient times, and frequent time zone crossings on long-haul flights are characteristics of work life in the aviation industry. All three of these factors have a negative impact on alertness and may increase the risk of accidents.

[0005] One way to combat aircrew fatigue is through restrictions on flying and duty time. However, regulators are currently investigating how to incorporate sleep and performance science directly into their fatigue risk management systems through biomathematical approaches to modelling crew sleepiness and fatigue.

[0006] In view of this, this invention is proposed. Summary of the invention

[0007] The purpose of the present invention is to provide a method for assessing the alertness of flight crew members based on a biomathematical model, which can construct a model to assess the alertness of personnel under the observed sleep conditions and sleepiness data. The specific implementation steps are as follows: Step 1: Internal alertness S. The S process runs through all moments of a person's sleep / wakefulness. The wakefulness stage is characterized by the consumption of alertness, and the sleep stage is characterized by the recovery of alertness. At the same time, the recovery state is further divided into two situations according to the sleep situation; Step 2: Sleep inertia W. The W process is the sleep inertia of the human body, which acts at the initial moment of the wakefulness stage, and its effect on alertness decays over time; Step 3: 24-hour phase process C. This is the body's own physiological regulation of work and rest, which is only related to the time of day. Regardless of sleeping or waking, it naturally adjusts alertness; Step 4: 12-hour phase process U. Same as step 3, except that the adjustment effect acts on 12 hours; Wherein, step 1 comprises the following steps: Step 1.1: Establish the function of S in the awake state with respect to the awake time, which includes parameters such as low-level asymptotes and attenuation coefficients.

[0008] Step 1.2: Establish a function of S in the sleep state with respect to sleep time, which includes a high-level asymptote. Modify the original process since 1990, taking S at the time of falling asleep as input, and divide the process into two sleep parts: high-pressure and low-pressure.

[0009] Step 2 includes the following steps: Step 2.1: Establish a function of W in the awake state with respect to the awake time, which includes an initial decrease and an exponential recovery coefficient.

[0010] Step 3 includes the following steps: Step 3.1: This process is only related to the time of day, the period of C is 24 hours, and it contains the default day and night phase, amplitude and muon parameters.

[0011] Step 4 includes the following steps: Step 4.1: This process has a cycle of 12 hours and has the same functional form as step 3, but the parameter values ​​are different.

[0012] First, let’s talk about the calculation process of S. In the function of internal alertness, there are two constant asymptotes, a higher level (ha=14.3) and a lower level (la=2.4). When we are awake, alertness decreases, and when we sleep, alertness recovers.

[0013] During wakefulness, this process can be represented as S, which takes as input the S level during wakefulness, S(sw), and computes the decay of alertness (d = -0.0353) as a function of the time spent awake (taw).

[0014] During sleep, there are three functions that describe the internal alertness process. Let the first one be S'. This is the original process, which takes the S level (ss) when falling asleep as input.

[0015] Here, the S' process is extended and corrected to the "braking function". Equation (2) can be divided into S ’ 1 and S ’ 2 , where the former is for the high-pressure sleep part. When S' < bl = 12.2, it is in a high-pressure state; the latter is for the low-pressure sleep part.

[0016] In the above equations, parameters are used to calculate the recovery of alertness as a function of sleep time (tas), and the calculation of the braking point (bt) of sleep time is used to determine when to switch from S ’ 1 to S ’ 2 .

[0017] The above is the complete calculation process of internal alertness. As mentioned before, the value of sleep inertia also needs to be calculated. The sleep inertia function reduces the alertness at the beginning of waking, but this reduction effect gradually decays as a function of wake time (taw) (W c = -5.72, W d = -1.51).

[0018] Finally, there are two processes related to the time of day (tod). They are called the 24-hour and 12-hour phase processes respectively. Among them, the period of process C is 24 hours, with a default circadian rhythm phase (p = 16.8), amplitude (Ca = 2.5), and midpoint (Cm = 0).

[0019] The form of process U is similar to the above equation, but the period of (8) is 12 hours, with an amplitude (Ua = 0.5) and a midpoint (Um = -0.5).

[0020] Adding the above four model components together can generate a comprehensive alertness score, and this score can be used in a linear transformation function (a = 10.6, b = -0.6) to predict the score of the Karolinska Sleepiness Scale (KSS), as shown in the following formula: The model also includes a sleep generator that can be used to automatically insert sleep in imperfect data when no observed sleep is available. The sleep generator is an implicit part of TPM and in its current implementation is based on alertness scores and proximity to duty and sleep. The threshold for falling asleep is set to S+C+U<8.38 by default, and the threshold for waking up is set to S+C+U>11.38 by default. At the same time, in order to prevent the tossing behavior of repeatedly falling asleep and waking up in certain situations, some restrictions are added: you cannot sleep within 1 hour after being on duty, you cannot sleep again within 2 hours after getting up, and you will not wake up in the first hour after falling asleep.

[0021] That’s all about the model.

[0022] The relevant pictures are as follows. BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 It is a schematic diagram of the algorithm structure of the present invention. Figure 2 It is a schematic diagram of S+C+0+w in the present invention. Figure 3 It is a schematic diagram of S+C+U in the present invention. Figure 4 It is a schematic diagram of S+C in the present invention.

Claims

1. A method for calculating pilot alertness based on a biomathematical model, including: Step 1, internal alertness S. The S process runs through all moments of a person's sleep / wakefulness. The wakefulness stage is characterized by the consumption of alertness, and the sleep stage is characterized by the recovery of alertness. At the same time, the recovery state is further divided into two situations according to the sleep situation; Step 2, sleep inertia W. The W process is the sleep inertia of the human body, which acts at the initial moment of the wakefulness stage, and its effect on alertness decays over time; Step 3, 24-hour phase process C. This is the body's own physiological regulation of work and rest, which is only related to the time of day. Regardless of sleeping or waking, it naturally adjusts alertness; Step 4, 12-hour phase process U. Same as step 3, the only difference is that the adjustment effect acts on 12 hours.

2. The method for calculating pilot alertness based on a biomathematical model as claimed in claim 1, characterized in that: In step 1, internal alertness is consumed during wakefulness and restored during sleep, which can be expressed as a function of time; the sleep process includes a high static pressure process and a low static pressure process, and the recovery of alertness is different in different states; the specific stress state is determined based on the braking of the sleep time.

3. The method for calculating pilot alertness based on a biomathematical model as claimed in claim 1, characterized in that: In step 2, the sleep inertia function only takes effect at the initial stage of the wakefulness state, and its effect will rapidly decay over time.

4. The method for calculating pilot alertness based on a biomathematical model as claimed in claim 1, characterized in that: In step 3, the phase process is related to the time of the day, has a cycle of 24 hours, and includes a default day and night phase.

5. The method for calculating pilot alertness based on a biomathematical model as claimed in claim 1, characterized in that: In step 4, the phase process is related to the time of the day, has a cycle of 12 hours, and includes a default day and night phase.

6. According to the method for calculating pilot alertness based on a biomathematical model according to any one of claims 1 to 5, a comprehensive alertness score can be obtained by adding the above components, which can be used to predict the score of the Karolinska Sleepiness Scale (KSS).

Images