WO2022233676A1

WO2022233676A1 - Method of measuring vehicle driver's behaviour towards cyclists

Info

Publication number: WO2022233676A1
Application number: PCT/EP2022/061193
Authority: WO
Inventors: Iulian Podaru
Original assignee: Continental Autonomous Mobility Germany GmbH
Priority date: 2021-05-04
Filing date: 2022-04-27
Publication date: 2022-11-10
Also published as: DE112022002341T5

Abstract

The invention refers to a method of measuring a vehicle driver's behaviour towards cyclists performed by means of at least a data processing hardware using sensor data from a sensor arrangement. The method includes pre-determining a minimal lateral distance between the vehicle and the cyclist during an overtaking manoeuvre as default safe distance. Each time the cyclist is identified in front of the vehicle and an overtaking manoeuvre of the identified cyclist by the vehicle is initiated, the method further includes determining a lateral distance between the vehicle and the identified cyclist from a multitude of lateral distances collected during the overtaking manoeuvre by means of the sensor arrangement and saving the determined lateral distance from a first overtaking manoeuvre as average lateral distance. After each overtaking manoeuvre the method includes incrementing by one unit a count of cyclists' overtaking manoeuvres and saving the count of cyclists' overtaking manoeuvres. The method further includes computing an average lateral distance starting from a second overtaking manoeuvre, the average lateral distance depending on the saved average lateral distance between the vehicle and the cyclist from a previous step of the method, the determined lateral distance and the count of cyclists' overtaking manoeuvres and saving the average lateral distance. The method includes computing a behaviour indicator depending on both the saved average lateral distance and the pre-determined minimal lateral distance. The data processing hardware is configured to present the behaviour indicator to the vehicle driver. ]

Description

Method of measuring vehicle driver's behaviour towards cyclists

Field of the invention

[0001] The present invention relates to a method of measuring behaviour of a vehicle driver when the driver overtakes a cyclist.

Prior art

[0002] On a road, unexpected manoeuvres of avoiding potholes, drainage systems or debris may endanger cyclists and other road users, as well. Also, cyclists may lose their balance due to unexpected mechanical issues or tyre punctures. They may also be affected by air flow caused by large vehicles, especially heavy trailers. All these manoeuvres and events may lead to hazard on the road when a vehicle overtakes a cyclist without having a safe lateral distance. That's why cyclists are considered vulnerable road users. Any accident that involves a cyclist could lead to serious injuries, even deaths. In 2016, cyclists accounted 8% of all road casualties in the European Union. Around 25,000 cyclists were killed between 2004 and 2013 on European roads. The European Union promotes cycling within the context of health, but with emphasis on safe use of the roads. Thus, the European Union recommends overtaking of a cyclist by a vehicle driver at a safe lateral distance of 1 - 1.5 meters, according to existing regulations. [0003] EP 2643829 B1 relates to a method and device for reducing instances of vehicle collisions. Based on parameter data of a first vehicle and road conditions, and parameter data regarding a second vehicle, the first vehicle can undergo warnings or positional adjustments to reduce the likelihood of a collision with the second vehicle or other collision body. The vehicle adjustment can be in the form of braking or steering adjustments whereas the warning can be in the form of a visual, audio, or haptic signal.

[0004] EP 2528765 B1 discloses a method for a vehicle, to identify moving objects on the side and to avoid unnecessary collision warnings, the method includes: a) detecting an object with the aid of a first sensor; b) detecting the object with the aid of a second sensor; c) checking whether the object has left the detection area of the second sensor; and dl) discarding the distance data measured by the first sensor and the second sensor if the object has left the detection area of the second sensor; or d2) determining the position of the object from measured distance data if the object has not left the detection area of the second sensor.

[0005] EP 3257034 B1 provides enhanced warnings of potential future adverse events (e.g., automobile crashes) by tracking the location and motion of multiple vehicles and providing alerts or warnings to the drivers of such vehicles in case that a risk of an adverse event is identified.

[0006] Specified documents provide methods and/or devices for vehicles collision avoidance based on data received from vehicle sensors that provide lateral distances between vehicles. This information is used therefore for warning the vehicle driver.

Disadvantages of the prior art

[0007] Known systems or methods from prior art provide warnings of side collision avoidance for vehicles or even interventions into the braking system or steering system of the vehicles. The downside of a vast majority of these actions of warnings or interventions happen only when a high collision risk is detected. The driver does not become aware of his/her behaviour towards cyclists since there is no recording of how the driver has behaved so far during overtaking manoeuvres, the driver' s behaviour not being tracked or followed historically in time.

[0008] Moreover, even though the European Union guidelines and regulations recommend a minimal lateral distance, unfortunately this is not always put into practice since it is difficult to measure the actual lateral distance in case of an incident.

Problem to be solved by the invention

[0009] The main problem to be solved by the present invention is the lack of awareness of a vehicle driver regarding the danger at overtaking a cyclist at an unsafe lateral distance.

[0010] The objective of the invention is to keep a score based on driver's behaviour towards cyclists and also to warn the driver about his/her behaviour during overtaking manoeuvre.

SUMMARY OF THE INVENTION

[0011] In order to solve the problem, the inventor conceived a computer-implemented method for measuring the vehicle driver's behaviour towards cyclists performed by means of at least a data processing hardware using sensor data from a sensor arrangement of which field of view covers the frontal lateral area in the vicinity of the vehicle. The method comprises the following steps:

- pre-determining a minimal lateral distance between the vehicle and the cyclist during an overtaking manoeuvre as default safe distance;

- each time the cyclist is identified in front of the vehicle and an overtaking manoeuvre of the identified cyclist by the vehicle is initiated: - determining a lateral distance between the vehicle and the identified cyclist from a multitude of lateral distances collected during the overtaking manoeuvre by means of the sensor arrangement; and saving the determined lateral distance from a first overtaking manoeuvre as average lateral distance;

- after each overtaking manoeuvre, incrementing by one unit a count of cyclists' overtaking manoeuvres; and saving the count of cyclists' overtaking manoeuvres;

- computing an average lateral distance starting from a second overtaking manoeuvre, each time the count of cyclists' overtaking manoeuvres is incremented, the average lateral distance depending on the saved average lateral distance between the vehicle and the cyclist, the determined lateral distance and the count of cyclists' overtaking manoeuvres; and saving the average lateral distance;

- computing a behaviour indicator depending on both the saved average lateral distance and the pre-determined minimal lateral distance;

- displaying the behaviour indicator towards cyclists to the vehicle driver.

The advantages of the invention

[0012] By providing a score regarding the driver's behaviour towards cyclists based on saved historical data, the awareness of the driver regarding overtaking cyclists is increased .

[0013] In order to operate the method according to the invention, usually no extra sensors are necessary since most vehicles are already equipped with sensors from Advanced Driver Assistance Systems (ADAS).

Brief description of the drawings

[0014] Next, there are given few drawings to better illustrate the invention, namely:

- Fig. 1 shows a scenario of overtaking a cyclist by a vehicle where recommendations for lateral distance are not met;

- Fig. 2 shows a scenario of overtaking a cyclist by a vehicle at a safe lateral distance;

Fig. 3 shows a schematic diagram of a method of measuring vehicle driver's behaviour towards cyclists, according to invention.

[0015] List of references:

10 - vehicle

11 - sensor C - cyclist

D_LAT - lateral distance - minimal lateral distance

Detailed description of the invention and embodiments

[0016] Referring now to Fig. 1, a scenario of overtaking a cyclist where recommendations for lateral distance are not met is shown. A driver of a vehicle 10 is going to overtake a cyclist C driving on a side part of the road. The driver may present great risk of injuring the cyclist C on the road in a variety of circumstances, such as the vehicle fails to give way to the cyclist. Moreover, in case of any unexpected manoeuvre or event on cyclist's side (for example, lose balance, tyre puncture, unexpected obstacle on the road) the vehicle 10 may endanger the cyclist C in case the vehicle 10 overtakes the cyclist C without a safe lateral distance.

[0017] Fig. 2 shows a scenario of overtaking a cyclist by a vehicle at a safe lateral distance, for example 1 - 1.5 meters. Vehicle 10 is equipped with a sensor arrangement comprising at least one sensor 11 of short-range radar sensor, camera sensor, ultrasonic sensor or lidar sensor, namely sensors of Advanced Driver Assistance Systems (ADAS) technologies, for example. The at least one sensor 11 may be positioned in a frontal corner of the vehicle 10, for example, but other arrangements/positions of the sensor may be adopted as well, the reason being that the sensor's field of view has to cover the frontal lateral area in the vicinity of the vehicle 10. The at least one sensor 11 is provided with a data processing hardware (for example, a controller, processor or microcontroller) and a memory hardware in communication with the data processing hardware. The memory hardware stores instructions that when executed by the data processing hardware cause the data processing hardware to perform lots of functions, such as object detection, object classification (vehicle, cyclist, pedestrian, obstacle), motion status of the detected object (including detecting that the cyclist is moving or is stationary), free space detection in front of the vehicle, road boundary detection, speed and relative lateral distance measurement, lane markers detection and so on. For more plausible results in performing the specified functions, the sensor arrangement may comprise a combination of sensors, such as short-range radar sensor used in combination with a camera or a lidar sensor, for example. Data about relative lateral distances between the vehicle and the cyclist are collected from the at least one sensor 11 of the sensor arrangement during overtaking manoeuvre. Another piece of information that could be used depends on the region/area/country travelled and if there are regulations regarding a minimal lateral distance D_MLAT. In case there are corresponding regulations, then this minimal lateral distance D_MLAT could be used as threshold, otherwise a default/pre determined safe distance is used as threshold.

[0018] Fig. 3 shows a schematic diagram of a method 100 of measuring a vehicle driver's behaviour towards cyclists, according to the invention. The method 100 comprises the following steps realised by means of at least a data processing hardware using sensor data from a sensor arrangement of the vehicle:

SO- pre-determining a minimal lateral distance between the vehicle and the cyclist during an overtaking manoeuvre as default safe distance;

51 - each time the cyclist is identified in front of the vehicle and an overtaking manoeuvre of the identified cyclist by the vehicle is initiated:

Sl.l - determining a lateral distance between the vehicle and the identified cyclist from a multitude of lateral distances collected during the overtaking manoeuvre by means of the sensor arrangement; and saving the determined lateral distance from a first overtaking manoeuvre as average lateral distance;

52 - after each overtaking manoeuvre, incrementing by one unit a count of cyclists' overtaking manoeuvres; and saving the count of cyclists' overtaking manoeuvres;

53 - computing an average lateral distance starting from a second overtaking manoeuvre, the average lateral distance depending on the saved average lateral distance between the vehicle and the cyclist from a previous step of the method, the determined lateral distance and the count of cyclists' overtaking manoeuvres; and saving the average lateral distance;

54 - computing a behaviour indicator depending on both the saved average lateral distance and the pre-determined minimal lateral distance;

55 - displaying the behaviour indicator towards cyclists to the vehicle driver.

[0019] At step SO the pre-determined minimal lateral distance between the vehicle and the cyclist may be

received from the data processing hardware of the at least one sensor from the sensor arrangement every time an overtaking manoeuvre is performed or may be read from the memory of the at least one sensor 11 . This means that the minimal lateral distance DML_AT has been saved in the memory of the at least one sensor 11 and it remains unchanged until a new value for the minimal lateral distance D_MLAT is notified and then saved in the memory. This happens when the vehicle enters in a region/area/country with new regulations about the minimal lateral distance DML_AT towards cyclists.

[0020] At step SI the data processing hardware of the at least one sensor 11 detects the presence of an object in front or front side of the vehicle, in a proximity covered by the field of view of the at least one sensor 11 and identifies the object as cyclist. The object is preferably identified as a cyclist before the vehicle passes by the cyclist. The overtaking manoeuvre of the identified cyclist may be assumed when the frontmost point of the car gets in the rearmost point of the cyclist, for example.

[0021] The lateral distance between the vehicle and the identified cyclist from sub-step Sl.l and referred hereinafter as DL_AT is determined per overtaking manoeuvre from a multitude of relative lateral distances between the vehicle and the cyclist. The multitude of relative lateral distances is collected/measured during each overtaking manoeuvre by means of the at least one sensor 11 preferably only when the vehicle gets in the proximity or vicinity of the cyclist, wherein said proximity or vicinity corresponds to a detected distance between the cyclist and the vehicle being less than a threshold distance. This has the effect of reducing the amount of collected relative lateral distances. The lateral distance DL_AT may be determined as a minimum of the multitude of relative lateral distances, for example. Another possibility for determining the lateral distance DL_AT is to measure it during overtaking manoeuvre when the frontmost point of the car gets in the rearmost point of the cyclist. The determined lateral distance D_LAT from the first overtaking manoeuvre (at a first use of the method) is saved as an average lateral distance in the memory of the at least one sensor 11 since there aren't at least two lateral distances so far to be able to compute an average with them.

[0022] The count of cyclists' overtaking manoeuvres from step S2 is incremented by one unit and saved in the memory of the at least one sensor 11 of the sensor arrangement after each overtaking manoeuvre.

[0023] At step S3 a new average lateral distance is computed starting from the second overtaking manoeuvre (from a second use of the method). At the first use of the method, the average lateral distance is the first lateral distance D_LAT received during the first cyclist' overtaking manoeuvre (from the first use of the method). Starting with the second cyclist' overtaking manoeuvre, the lateral distance D_LAT determined at step Sl.l is used to compute the new average lateral distance with the following formula, for example, but not limited to: NewAverageLatDist = (OldAverageValue * OldNumberOfOvertaking + ActualMeasuredLateralDistance) / (OldNumberOfOvertaking†1) wherein,

OldAverageValue is the average lateral distances stored/saved already in the memory of the at least one sensor at the moment of overtaking manoeuvre;

OldNumberOfOvertaking is the number of cyclists' overtaking manoeuvres saved or stored already in the memory of the at least one sensor at the moment of overtaking manoeuvre;

- ActualMeasuredLateralDistance - is the lateral distance DLAT determined during cyclist' overtaking manoeuvre.

The new average lateral distance is saved in the memory of the at least one sensor 11 of the sensor arrangement after each overtaking manoeuvre. This average lateral distance will be OldAverageValue from the formula when compute NewAverageLatDist at a next overtaking manoeuvre.

[0024] The behaviour indicator from step S4 may be sent to an infotainment system or a vehicle Instrument Cluster and may be expressed in two ways: as a colour code and/or as a number less or equal to 100.00 with two decimals computed using a formula, for example. If the travelled country regulations regarding lateral distance are respected, more exactly if the driver overtakes the cyclist at the minimal lateral distance D_MLAT then a value of 100% is considered as respected. The colour codes may be classified using a comparison between the average lateral distance and a percentage of the minimal lateral distance D_MLAT from the travelled country regulations as follows, for example:

- green, if the average lateral distance >= 100%, i.e. the average lateral distance is greater than or equal to the minimal lateral distance from regulations;

- yellow, if the average lateral distance >= 75%, but < 100%, i.e. the average lateral distance is greater than or equal to 75 percent and less than 100 percent calculated from the minimal lateral distance from regulations;

- red, if the average lateral distance >= 50%, but < 75%, i.e. the average lateral distance is greater than or equal to 50 percent and less than 75 percent calculated from the minimal lateral distance from regulations;

- purple, if the average lateral distance < 50%, i.e. the average lateral distance is less than 50 percent calculated from the minimal lateral distance from regulation.

The formula for computing the behaviour indicator (or shortly named score) as a number may be, for example:

AverageLatDist

Score = Min(100.00, - -- * 100)

RegulationValue wherein,

- Min is the minimum between 100.00 (which is considered the maximum score) and the average lateral distance reported to the minimal lateral distance from the travelled country regulations;

- AverageLatDist is the average lateral distance that was saved in the memory of the at least one sensor; at the first use of the method, the AverageLatDist is the first lateral distance measured during the first overtaking;

- RegulationValue is the minimal lateral distance DMT.¾T provided by the data processing hardware according to travelled region/area/country regulations.

[0025] At step S5 the data processing hardware is configured to present the behaviour indicator to the vehicle driver. There are several methods to inform the driver about his/her behaviour towards cyclists:

- passive information - the data processing hardware is configured to present the behaviour indicator to the vehicle driver whenever the driver wants to know his/her behaviour towards cyclists. Thus, the behaviour indicator can be found in a vehicle menu of the infotainment system, for example as colour code and/or as a numerical value when the driver searches it. In this way, the driver is not annoyed by a multitude of warning messages; the behaviour indicator may be present in the vehicle menu from the first use of the method and in this case the behaviour indicator is based on the first lateral distance measured at the first cyclist overtaking, or it may be available in the vehicle menu after ten overtaking manoeuvres, for example; active information - whenever the driver overtakes one cyclist C, the driver is informed with a short pop-up message that appears on the vehicle cluster, for example, revealing how friendly he/she was with the cyclist C during the overtaking manoeuvre. The message may be accompanied by the colour code as previously described, but in this case, the colour code is classified using a comparison between the lateral distance determined during the overtaking manoeuvre and the percentage of the minimal lateral distance from the travelled country regulations, for example.

[0026] By informing the driver about his/her behaviour towards cyclists based on the behaviour indicator, the driver is encouraged to respect the safe lateral distance at overtaking cyclists, the awareness of the drivers regarding the danger at overtaking cyclists at an unsafe lateral distance will be increased, and cyclists will be safer on the roads. [0027] In an embodiment, the idea of the invention can be used when investigating road accidents in which cyclists are involved. There might be a mechanism like Event Data Recorder which is monitoring the distances between the vehicle and vulnerable road users.

Claims

Patent claims

1. Method of measuring a vehicle driver's behaviour towards cyclists, the method being performed by means of at least a data processing hardware using sensor data from a sensor arrangement of the vehicle, characterized by that the method comprises the following steps:

(SO)- pre-determining a minimal lateral distance between the vehicle and the cyclist during an overtaking manoeuvre as default safe distance;

(51) - each time the cyclist is identified in front of the vehicle and an overtaking manoeuvre of the identified cyclist by the vehicle is initiated:

(Sl.l) - determining a lateral distance between the vehicle and the identified cyclist from a multitude of lateral distances collected during the overtaking manoeuvre by means of the sensor arrangement; and saving the determined lateral distance from a first overtaking manoeuvre as average lateral distance;

(52) - after each overtaking manoeuvre, incrementing by one unit a count of cyclists' overtaking manoeuvres; and saving the count of cyclists' overtaking manoeuvres;

(53) - computing an average lateral distance starting from a second overtaking manoeuvre, the average lateral distance depending on the saved average lateral distance between the vehicle and the cyclist from a previous step of the method, the determined lateral distance and the count of cyclists' overtaking manoeuvres; and saving the average lateral distance;

(54) - computing a behaviour indicator depending on both the saved average lateral distance and the pre-determined minimal lateral distance;

(55) - displaying the behaviour indicator towards cyclists to the vehicle driver. 2. Method according to claim 1, wherein the sensor data comprises at least one of short-range radar sensor data, lidar sensor data, ultrasonic sensor data or camera sensor data.

3. Method according to claim 1, wherein the sensor data are processed by at least the data processing hardware of at least one sensor from the sensor arrangement.

4. Method according to claim 1, wherein the behaviour indicator is expressed as a colour code, wherein the colour code is determined using a comparison between the updated average lateral distance and the minimal lateral distance.

5. A computer program product comprising computer-readable instructions which when executed on a processor of data processing hardware cause the processor to perform the method according to any of the preceding claims.