DE102020209202A1 - Ultrasound-based creation of objects - Google Patents

Abstract

Disclosed is a method for determining objects in the immediate vicinity of a vehicle by a control unit, with direct echoes and/or cross echoes being received from a plurality of ultrasonic sensors in the form of measurement data, echo distances being determined from the received measurement data and entered in echo lists, with the echo distances between a Ultrasonic sensor and at least one reflection point are determined, traces of the echo distances are created by tracking the echo distances over time in the echo lists, an object point and/or an object are determined based on the traces created. Furthermore, a control unit, a computer program and a machine-readable storage medium are disclosed.

Description

The invention relates to a method for determining objects in the vicinity of a vehicle. Furthermore, the invention relates to a control unit, a computer program and a machine-readable storage medium.

State of the art

Ultrasonic sensors are currently used primarily to support vehicles parking or leaving a parking space. The ultrasonic sensors are used to send and receive echoes. The echoes received can be used to determine transit times and distances to relevant obstacles in the field of view can be calculated. Then, for example, the driver can be warned if there is a risk of a collision.

The ultrasonic sensors work according to a pulse-echo method, in which the propagation times of sound pulses between an ultrasonic transmitter, a reflection point and an ultrasonic receiver are measured. In this case, the ultrasonic transmitter can also function as an ultrasonic receiver. Ultrasonic sensors distributed evenly over the vehicle thus alternately emit ultrasonic pulses, which are reflected by objects in the area and are received again by the same ultrasonic sensor in the form of direct echoes or by neighboring ultrasonic sensors in the form of cross echoes. The received amplitudes of the echoes that are above a defined threshold are stored in echo lists as amplitudes over time. With the help of a subsequent evaluation, the distances to obstacles can be calculated

Various sensors are used in particular in vehicles that can be operated automatically in order to detect objects in the vehicle environment. The measurement data from ultrasonic sensors have so far only been used to support object recognition. An exclusive object recognition based on ultrasonic sensors has not yet been carried out.

Disclosure of Invention

The object on which the invention is based can be seen as proposing an exclusively ultrasound-based method for detecting objects.

This object is solved by means of the respective subject matter of the independent claims. Advantageous configurations of the invention are the subject matter of the dependent subclaims.

According to one aspect of the invention, a method for determining objects in a close range of a vehicle is provided. The method can preferably be carried out by a control unit.

In one step, direct echoes and/or cross echoes are received from a plurality of ultrasonic sensors in the form of measurement data and echo distances are calculated from the measurement data received. The echo distances between an ultrasonic sensor and at least one reflection point are preferably determined. The calculated echo distances are then entered in echo lists.

The sensors can generate echoes or sound pulses at regular time intervals and emit them into the environment. Direct echoes reflected in the surroundings can be detected by the respective ultrasonic sensors. Direct echoes are received or detected by the ultrasonic sensors and transformed into electrical signals or digital measurement data. The resulting electrical signals or measurement data can then be made available to the control unit for further calculations.

In a subsequent step, traces of the echo distances are created by tracing the echo distances over time in the echo lists. The change over time in the echo distances entered in the echo lists can be determined.

The so-called "tracks" or "traces" are generated by plotting the echo distances over time. As a result, the history of the determined echoes can be used as part of tracking.

An object point and/or an object are then determined based on the traces created. Only stable traces are used in this processing step, which are determined over several measurement cycles.

If, for example, individual traces do not appear in every measurement cycle, such traces are classified as "unstable" and are not further considered.

In particular in the case of vehicles which can be operated in a highly automated or fully automated manner, the detection threshold for generating objects can be set as sensitively as possible. The method can be used independently of other sensors or in support of sensor data fusion.

Based on the tracks created, the best possible estimate for spawning or object generation can be formed, which is initially based exclusively on ultrasound data. As a result, consequential errors such as false alarms can be reduced.

The ultrasonic sensors for collecting measurement data from cross echoes can be used both for generating or emitting sound waves into the environment and for receiving sound waves reflected back from the environment. For example, a first ultrasonic sensor can be set up exclusively for emitting sound waves and a second ultrasonic sensor can be set up exclusively for receiving sound waves. Alternatively or additionally, the ultrasonic sensors can be operated alternately in a transmission mode and in a reception mode.

A distance to at least one reflection point is then determined by triangulation, multilateration and the like of the received measurement data from the direct echoes and the cross echoes of the ultrasonic sensors. A position and/or a height of the reflection point in the area can also be determined. The at least one reflection point can lie, for example, on a surface of an object in the vicinity of the vehicle.

According to a further aspect of the invention, a control unit is provided, the control unit being set up to carry out the method. The control unit can be, for example, a vehicle-side control unit, a vehicle-external control unit or a vehicle-external server unit, such as a cloud system.

In addition, according to one aspect of the invention, a computer program is provided which includes instructions which, when the computer program is executed by a computer or a control unit, cause the latter to execute the method according to the invention. According to a further aspect of the invention, a machine-readable storage medium is provided on which the computer program according to the invention is stored.

According to the BASt standard, the vehicle can be operated in an assisted, partially automated, highly automated and/or fully automated or driverless manner. However, the method is not limited to the automotive sector. In particular, the method can be used in mobile units, such as robots, industrial robots, watercraft, rail vehicles, commercial vehicles, agricultural vehicles, buses, robotaxis, airplanes, drones, helicopters and the like.

The method can preferably be applied to mobile units which can be operated in an assisted, partially automated, highly automated or fully automated manner. Such a mobile unit can preferably be highly automated or fully automated and can be operated completely without the intervention of a driver or an operator.

The aim of the method is to identify all relevant objects in the vehicle environment as quickly as possible and to have a minimum number of ghost objects.

The so-called spawning or the detection of objects based on ultrasound can contribute to a functional improvement and an increase in safety, since each sensor can only collect measurement data from new objects within its field of vision or scanning range. For example, in a traffic situation, an object can only appear in the vicinity of the vehicle. In particular, the vicinity of the vehicle can be effectively monitored by ultrasonic sensors. For example, such a traffic situation can be illustrated by a pedestrian who, concealed by road users, can walk sideways into a field of vision of the vehicle sensor system. In particular, a pedestrian may fall in front of the vehicle or be a child, such that other sensors, such as LIDAR sensors or camera sensors, cannot detect the pedestrian underneath a hood.

A further advantage compared to a system without ultrasonic-based object recognition is that a start-up release for the vehicle that can be operated automatically is made possible. At least two independent sensors that are sensitive in the close-up range, such as a surround vision camera or short-range radar with ultrasonic sensors, can monitor the close-up range and, in an emergency, generate an object on the sensor fusion level. Sensor-specific weaknesses can be compensated for by such a redundant configuration.

In one embodiment, an object point and/or an object is determined from each track created. As a result, a technically particularly simple variant for spawning based on ultrasonic sensors can be provided. Each reflection point formed from echo lists creates a new object. The method can thus be carried out particularly quickly and used, for example, in dynamic traffic situations.

According to a further exemplary embodiment, a speed of the vehicle is determined, with an object point and/or an object being determined on the basis of the created tracks if the determined speed of the vehicle falls below a threshold value. This measure allows the method to be restricted such that the ultrasound-based detection of objects takes place below a defined speed threshold. As a result, the spawning or the determination of objects object points can only be carried out at a speed of 0 km/h or, for example, below 5-10 km/h. At higher vehicle speeds, the vehicle would no longer be able to react reliably due to the low sampling frequency of the ultrasonic sensors, so that the necessary computing power for the ultrasonic-based object recognition is controlled as required.

According to a further embodiment, in order to determine an object, the traces created are generated from echo distances from a basic number of reflection points that are locally adjacent to one another. As a result, a minimum number of reflection points that belong together is necessary in order to determine an object or an object point. The reflection points that belong together can, for example, be arranged spatially adjacent to one another or within a defined environment. For example, the reflection points that belong together can be arranged in a radius of a few centimeters, such as 2-20 cm. The base number of reflection points can be any natural number, such as 3-10. Such a procedure means that objects or object points can only be determined when a defined set of points is present. This allows ghost objects to be avoided.

According to a further exemplary embodiment, the basic number of reflection points that are spatially adjacent to one another is determined by a plurality of measurement cycles. Alternatively or additionally, the basic number of reflection points that are adjacent to one another or that belong together can be obtained in the form of a number of measurement cycles. For example, each reflection point can be provided with an identification number or ID, which is observed over several measurement cycles. If a reflection point with the same identification number appears several times in a row, the basic number can be fulfilled and the object point or object can be determined.

For example, a reflection point with one and the same identification number can be observed over at least three consecutive measurement cycles in order to fulfill a basic number of three. The assignment of the identification number to different reflection points can be implemented via internal clustering, for example.

According to a further embodiment, an amplitude is determined using the received measurement data, with at least one reflection point being assigned to a highly reflective object based on the determined amplitude. Preferably, the required basic number of reflection points that are locally adjacent to one another is reduced in the case of reflection points of a highly reflective object.

A highly reflective object usually has a higher number of reflection points per measurement cycle and therefore only needs to be observable with the same ID over significantly fewer measurement cycles before an object is generated or spawned. For example, a highly reflective object or an object with surfaces that can reflect ultrasonic waves particularly effectively can be generated from the very first measurement cycle.

The measurement data can preferably be in the form of amplitudes of the reflected echoes. The amplitudes can be received by the ultrasonic sensors in analog or digital form and represent the strength of the reflected ultrasonic echoes.

In the case of large amplitudes of the reflected echoes that have led to a reflection point or have been reflected from a reflection point, a changed or adapted basic number of reflection points may be required for spawning. For example, the basic number can be halved for highly reflective objects.

Alternatively or additionally, these and other possible echo parameters can be used to calculate a point significance. If this significance exceeds a threshold value, spawning takes place.

According to a further exemplary embodiment, the object is created several times over a number of measurement cycles and associated with objects that have already been created, with a successfully associated object being transmitted to a sensor data fusion. An object can be successfully associated if it appears in 90-100% of the measurement cycles. As a result of this measure, an internal validation is carried out in a first step, which is limited to the ultrasound system. Only after the ultrasound system has validated the object can the object be forwarded to a sensor data fusion. the Sensor data fusion is independent of the type of sensor system in question.

A possible implementation would be, for example, that initially the determination of objects is only activated internally or without being passed on to the sensor data fusion. In this case, the objects can already be determined based on an existing reflection point. However, this object is initially only kept internally and is not passed on. An attempt is then made to associate the internally determined object over a defined number of measurement cycles, for example 10 measurement cycles each lasting 100 ms, using the echo data or other reflection points. If this is successful in a defined number of cases, for example 90-100%, the object is also forwarded externally to the sensor data fusion.

For example, in the case of very low-reflection objects, no or only non-continuous reflection points can be determined. Such a hybrid spawning allows the internal objects or hybrid objects to be determined on the basis of one or more reflection points and the created echo lists. In this case, an object or an object point can be triggered in a technically simple manner by at least one reflection point. This object or object point is initially handled internally within the ultrasound system as a new object. A subsequent association of the hybrid object over several measurement cycles can result in a general release of the object for a sensor data fusion in the form of a validation.

According to a further embodiment, a height of the reflection point is determined, with an object point and/or an object being created based on the generated traces of the reflection point if the height of the reflection point is above a minimum height. The minimum height can be set, for example, in such a way that it corresponds to an object that cannot be driven over. Such a for example can be above 6 cm. If measurement data from direct echoes and/or cross echoes are evaluated by a number of ultrasonic sensors and an object or an object point is detected above a subsurface, an object can be determined immediately, since higher objects are particularly important for enabling the vehicle to start moving.

The respective configurations of the method can be implemented in the form of a serial or parallel query when evaluating the measurement data from ultrasonic sensors.

• 1 eine schematische Draufsicht auf eine Anordnung gemäß einer Ausführungsform und
• 2 ein schematisches Ablaufdiagramm zum Veranschaulichen eines Verfahrens gemäß einer Ausführungsform.

Preferred exemplary embodiments of the invention are explained in more detail below on the basis of greatly simplified schematic representations. show here

• 1 a schematic plan view of an arrangement according to an embodiment and
• 2 a schematic flow diagram for illustrating a method according to an embodiment.

the 1 1 shows a schematic plan view of an arrangement 1 according to an embodiment. The arrangement 1 has a vehicle 2 . An object 4 is arranged in an environment U of the vehicle 2 . In particular, the object 4 is arranged in a close range of the vehicle 2 .

The vehicle 2 has a number of ultrasonic sensors 6, 8, which are set up to generate sound pulses and to emit them into the environment U. The ultrasonic sensors 6, 8 can be used to emit and receive direct echoes and/or cross echoes. In the 1 direct echoes are shown as an example.

Direct echoes reflected in the environment U can be detected by the respective ultrasonic sensors 6 , 8 . Direct echoes are received or detected by the ultrasonic sensors 6, 8 and transformed into electrical signals or digital measurement data. The resulting electrical signals or measurement data can then be received by a control device 10 .

Echo lists with echo distances 12 can be created based on the received measurement data. The echo distances 12 form distances between the ultrasonic sensors 6, 8 and reflection points 14 at which the direct echoes on the object 4 were reflected. Two reflection points 14 are illustrated as an example. The reflection points 14 are points on a surface of the object 4 facing the vehicle 2.

In the 2 1 is a schematic flowchart illustrating a method 16 according to an embodiment. The method 16 is used to determine objects 4 in a close range of the vehicle 2. In particular, the method 16 can be carried out by the vehicle-side control unit 10.

In a first step 18, a speed of vehicle 2 is determined. In this case, an object point 5 and/or an object 4 is determined on the basis of the tracks created if the speed of the vehicle 2 is below a threshold value. The method 16 can preferably be carried out when the speed of the vehicle 2 is below an exemplary threshold value of 10 km/h.

In a further step 20, direct echoes and/or cross echoes are received from a plurality of ultrasonic sensors 6, 8 in the form of measurement data. Echo lists with echo distances of 12 are created from the measurement data received.

Subsequently, traces of the echo distances 12 are created by tracking the echo distances 12 of the echo lists over time. In this case, the object 4 can be tracked, for example, over a number of scanning cycles. In particular, in step 22 the object 4 can be assigned to a plurality of reflection points 14 .

If such an assignment 24 takes place, the object 4 can be perceived by the control unit 10 as having been determined. The object 4 can thus be recognized by the control device 10 .

If no object 4 can be assigned to the determined reflection points 14, a further stage 26 of the method 16 is possible. In this case, the objects 4 can be assigned to a basic number of reflection points 14 with an increased amplitude or significance over a number of scanning cycles. If the assignment is carried out successfully, the object 4 is spawned or recognized 24. Otherwise, a next stage 28 of the method 16 can be carried out.

In the next stage 28 of the method 16, a height of the path reflection points 14 can be determined based on the determined measurement data. If a minimum height is exceeded, the object 4 and/or the object point 5 can be determined or spawned 24 .

If the reflection points 14 have a lower height than the minimum height, the reflection points 14 can be tracked 30 over a number of measurement cycles or scanning cycles and an object 4 or an object point 5 can be assigned in each measurement cycle. If the assignment is successful, the reflection points 14 can be assigned to one and the same object 4 in almost all measurement cycles. Such a successful assignment of the reflection points 14 also results in a determined object 4.

Claims (11)

Method (16) for determining objects (4) in a close range of a vehicle (2) by a control unit (10), wherein - direct echoes and/or cross echoes are received from several ultrasonic sensors (6, 8) in the form of measurement data, - Echo distances (12) are determined from the received measurement data and entered in echo lists, the echo distances (12) being determined between an ultrasonic sensor (6, 8) and at least one reflection point (14), - Traces of the echo distances (12) are created by tracing the echo distances (12) over time in the echo lists, - An object point and/or an object (4) are determined based on the traces created. procedure after claim 1 , An object point and/or an object (4) being determined from each track created. procedure after claim 1 or 2 , wherein a speed of the vehicle (2) is determined, based on the tracks created an object point and / or an object (4) are determined when the determined speed of the vehicle (2) falls below a threshold value. Procedure according to one of Claims 1 until 3 , wherein for the determination of an object (4) the traces created are generated from echo distances (12) from a basic number of reflection points (14) that are spatially adjacent to one another. procedure after claim 4 , wherein the basic number of reflection points (14) that are spatially adjacent to one another is determined by a plurality of measurement cycles. procedure after claim 4 or 5 , wherein an amplitude is determined on the basis of the received measurement data, wherein based on the determined amplitude at least one reflection point (14) is assigned to a highly reflective object, the required basic number of locally adjacent reflection points (14) being reduced for reflection points (14) of a highly reflective object will. Procedure according to one of Claims 1 until 6 , wherein the object (4) is created several times over a number of measurement cycles and associated with objects that have already been created, with a successfully associated object being transmitted to a sensor data fusion. Procedure according to one of Claims 1 until 7 , wherein a height of the reflection point (14) is determined, with a height of the reflection point (14) above a minimum height based on the generated traces of the reflection point (14) an object point and / or an object (4) is created. Control unit (10), wherein the control unit (10) is adapted to the method according to one of Claims 1 until 8th to execute. Computer program comprising instructions which, when the computer program is executed by a computer or a control unit (10), cause the latter to carry out the method (16) according to one of Claims 1 until 8th to execute. Machine-readable storage medium on which the computer program according to claim 10 is saved.

Images