Driver assistance system with fatigue detection and method for predicting the degree of fatigue

Abstract

A method for predicting a time at which a predetermined level of fatigue is expected to be reached in a vehicle fatigue detection system, comprising the steps of: (i) recording a fatigue profile by continuously determining the current level of fatigue of a driver; (ii) determining a change in fatigue over time within a predetermined time interval of the fatigue profile and / or providing an empirically determined change in fatigue over time, wherein the determination of the change in fatigue in step (ii) is carried out such that the fatigue profile within the predetermined time interval is independently filtered over two PT1 elements with different time constants and the change in fatigue is determined from the difference between the two filtered fatigue profiles;and (iii) determining the time at which a predetermined level of fatigue is expected to be reached, based on the determined and / or provided change in the level of fatigue.

Description

The present invention relates to a method for predicting a point in time at which a predetermined degree of fatigue is expected to be reached in a vehicle system for fatigue detection. The invention further relates to a driver assistance system having a processing unit for carrying out the method.Prior ArtTiredness in road traffic is one of the main causes of accidents. Driver assistance systems (ADAS) offer a possibility of automatically detecting driver fatigue and warning it in the event of driving inability.Driver assistance systems comprise a system for fatigue detection for this purpose. The latter is defined herein as an in-vehicle system that monitors and activates driver behavior if the driver seems impaired by tiredness. The aim is in particular to prevent a second sleep during driving.Through observation-based scaling methods, the fatigue of the vehicle operator may be associated with a certain degree of fatigue. An assessment of tiredness as objective as possible is ensured by the use of standardized scales. Two of the most common scales are from Wierville and Ellsworth (Wierville scale for short) and from the Human Factors Consult GmbH Berlin (HFC scale for short). The Wierville scale" includes four levels of driver fatigue ranging from "1 = non-tired" to "4 = extreme tired.". Each stage is characterized by specific eye parameters and behavioral indicators. The HFC scale includes nine levels of fatigue ranging from "1 = awake" to "9 = sleep". Nine indicators are used to classify a driver state. A further possibility for scaling is offered by the TU-Berlin drowsiness scale (TUBS scale for short).Already known systems for fatigue detection now calculate the degree of fatigue on the basis of sensor data using prediction algorithms. Such a sensory fatigue detection in real time can take into account the available driving performance parameters (steering movements, speed, lane keeping) or characteristic data of the vehicle driver, such as gaze movement or blinking behavior, head position and movement, heart rate or similar biophysiological parameters.However, the perceptibility and sustainability of fatigue detection is low for the end customer with the vehicle assistance systems currently on the market. The reason for this is that the fatigue detection is only noticeable when a predefined fatigue degree is reached. This may be due to the fact that the current fatigue detection systems only provide a value for the current fatigue state of the driver.It is known from DE 102 55 544 A1 to provide a motor vehicle assistance system with a fatigue prediction device for predicting a fatigue state of a driver of a motor vehicle in order to increase the traffic safety.DE 10 2011 121 537 A1 discloses a motor vehicle assistance system and method for generating, controlling and triggering a break recommendation as well as a motor vehicle.DE 10 2004 022 581 A1 discloses a driver assistance system which is provided for displaying a determined item of state information which specifies the degree of fitness or the tiredness of the driver.From the literature PLATHO, Christina; PIETREK, Anna; KOLREP, Harald: Sensing of driver fatigue; Bremen: Carl Schunemann Verlag, 2013 (Reports of the Federal In Streetnik-Vehrentechnik; F89). For page + Impression+ Table of contents+ pp. 1-75. - ISBN 978-3-956606-029-8, behavior observation is known, wherein different evaluation scales are present for detecting the driver fatigue.Disclosure of the InventionOne or more disadvantages of the prior art can be eliminated or at least reduced with the aid of the method according to the invention for predicting a point in time at which a predefined degree of fatigue is likely to be reached in a vehicle system for fatigue detection. The method according to the invention comprises the steps of (i) detecting the fatigue profile by continuously determining the current degree of fatigue of the vehicle driver; (ii) determining a change in the degree of fatigue over time in a predefinable time interval of the fatigue profile and / or providing an empirically determined change in the degree of fatigue over time; and (iii) determining the point in time at which a predefined degree of fatigue is expected to be reached on the basis of the determined and / or provided change in the degree of fatigue.With the aid of the invention, it is possible in a simple and robust manner to make a prediction about a point in time at which the degree of tiredness has reached a specific predefined value. The starting point is the current degree of tiredness. For this purpose, on the one hand, a gradient of the degree of tiredness (change in the degree of tiredness over time) can be determined in a predefined time interval. The time interval can be set, for example, such that the fatigue curve is continuously recorded in the last 10 minutes. Additionally or alternatively, an empirically determined value can also be used for the change in degree of tiredness over time. The empirically determined change in tiredness over time is preferably 0.5 to 1.5 degrees of tiredness, in particular 1 degree of tiredness, according to the HFC scale per hour.According to a preferred variant, the fatigue change is determined in steps (ii) in such a way that the fatigue curve is filtered independently of one another over two PT1 elements having different time constants in the predefined time interval. The change in fatigue degree is then determined from the difference between the two filtered fatigue curves. In this way, the degree of tiredness change can be provided particularly easily and in real time for the further steps of the method.In step (iii) of the method, the time at which the degree of tiredness reaches a predefined threshold value is now calculated on the basis of the current degree of tiredness of the vehicle driver and the determined and / or empirically determined change in the degree of tiredness. Preferably, this predetermined degree of tiredness is in the range of 5 to 7, in particular at 6, of the HFC scale.An essential aspect of the invention is therefore the provision of an algorithm which estimates from the history of the fatigue profile the time for which the driver can still drive his vehicle until he becomes tired--if the current driving style is maintained--and must take a pause. The time generated in the system can then be displayed to the driver for information purposes, for example, or taken into account in another manner in the vehicle assistance system or else a navigation system.According to a further preferred variant of the method according to the invention, an average value is formed from the degree of tiredness change determined in step (ii) and the degree of tiredness change provided. This makes it possible to further improve the prediction of the degree of tire fatigue change starting from the current degree of tire fatigue.Another aspect of the invention is to provide a vehicle assist system having a processing unit. This system and the processing unit are configured to carry out the method described above.Further preferred embodiments of the invention can be found in the following description or the dependent claims.DRAWINGSThe invention is explained in more detail below with reference to exemplary embodiments and an associated drawing. It shows: FIG. 1 shows a fatigue curve according to the HFC scale over time.Embodiment of the InventionA method for predicting a point in time at which a predefined degree of fatigue is expected to be reached in a vehicle system for fatigue detection will be described below.FIG. 1 shows a real course of the degree of tiredness over time up to a time t 1. The degree of fatigue is measured according to the HFC scale, and this value is determined and provided using known fatigue detection systems on the basis of sensory data in the vehicle. Furthermore, a threshold value 10 for the degree of tiredness is stored in the tiredness detection system, here the value 6 of the HFC scale. When this value is reached, a warning signal is conventionally output via the fatigue detection.The present method now provides a prediction for a time t 3, at which the degree of tiredness is expected to reach the threshold value 10.According to a first variant, after a certain minimum driving duration at any desired time t 1 the fatigue profile is analyzed in a certain time interval. The time interval Δt shown in FIG. 1 is limited, for example, by the times t 1 and t 2= t 1- 10 min. Using the gradient formula, the gradient can then be calculated from the degrees of tiredness at the times t 1 and t 2. As soon as the gradient is known, it is finally possible, assuming that the gradient of the degree of tiredness remains approximately constant, to extrapolate to a point in time t 3 at which the degree of tiredness reaches the threshold value 10 (in FIG. 1, intersection point of the upper dashed degrees 12 calculated from the gradient with the threshold value 10). Knowing the time t 3 a time until the warning (time-to-throughput) can now be determined and, if appropriate, indicated to the vehicle driver.According to a further variant, it is possible to calculate the warning time, i.e. the reaching of the threshold value 10, also on the basis of a slope stored in the system. When using the HFC scale, it can be calculated, for example, that the HFC value increases by 1 per hour of travel time. Accordingly, starting from the time t 1 it is possible in turn to determine a time t 3' at which the threshold value 10 is likely to be reached (see the point of intersection of the lower dashed degrees 14 calculated from the stored gradient with the threshold value 10). The gradient stored in the system can be determined empirically in advance.The two variants described above can be combined in particular in such a way that an average value is formed (see FIG. 1 solid degrees 16) and a correspondingly corrected time t provides 3".In a particularly preferred implementation of the variant in which the gradient of the degree of fatigue is determined directly from the detected fatigue profile, two PT1 elements with different time constants are provided as filter elements. The resulting difference between the filtered fatigue curves after the two PT1 terms is used to determine the slope of the degree of fatigue.

Claims

Method for predicting a point in time at which a predefined degree of fatigue is likely to be reached in a vehicle system for detecting fatigue, wherein the method comprises the steps of: (i) detecting a fatigue profile by continuously determining the current degree of fatigue of a vehicle driver; (ii) determining a change in the degree of fatigue over time in a predefinable time interval of the fatigue profile and / or providing an empirically determined change in the degree of fatigue over time, wherein the determination of the change in the degree of fatigue in step (ii) is carried out in such a way that the variation in fatigue in the predefined time interval is filtered independently of one another via two PT1 elements with different time constants and the change in the degree of fatigue is determined from the difference between the two filtered fatigue profiles; (iii) determining the point in time at which a predetermined degree of tiredness is predicted to be reached on the basis of the determined and / or provided change in the degree of tiredness.The method of claim 1, wherein an average is formed from the tiredness degree change determined in step (ii) and the tiredness degree change provided.Method according to one of the preceding claims, in which the predetermined degree of tiredness lies in the range from 5 to 7, in particular at 6, of the HFC scale.Method according to one of the preceding claims, in which the empirically determined degree of tiredness change over time is 0.5 to 1.5 degrees of tiredness according to the HFC scale per hour.Driver assistance system having a processing unit for carrying out a method according to one of the preceding claims.

Images