DE102017223570B3 - A method for enriching a cloud with acoustic vehicle environment data - Google Patents

Abstract

A method for enriching a cloud with vehicle interior acoustic environment data of at least one vehicle, the method comprising: detecting vehicle interior acoustic environment data by means of vehicle side sensors; searching for relevant patterns in the detected vehicle interior acoustic data and when a relevant pattern has been detected: classifying the identified relevant pattern as an event; Communicating the event and position data of the event to a cloud; cloud-side enriching of cloud data with the event and position data; cloud-side transmission of the enriched cloud data to other vehicles at a predetermined distance from the location data, the enriched Data in addition to the event and the position data additionally contain a time stamp.

Description

TECHNICAL AREA

The invention relates to the field of driver assistance systems. More particularly, the invention relates to a method for enriching a cloud or backend with vehicle acoustic environment data.

TECHNICAL BACKGROUND

Modern vehicles have a multiplicity of different driver assistance functions which assist the driver in planning the route or in guiding the vehicle. ADAS (Advanced Driver Assistance Systems) are designed to increase comfort, efficiency, safety and driver satisfaction while driving. These are electronic auxiliary devices in motor vehicles, which support the driver in certain driving situations. ADASs include, for example, adaptive headlight settings, adaptive cruise control, lane departure warnings, cornering warnings, speed limit warnings. For this purpose, ADAS intervene partly autonomously or autonomously in the drive, control or signaling devices of the vehicle and warn the driver shortly before or during critical situations by means of suitable man-machine interfaces. Some ADAS use a range of sensors such as RADAR, infrared, ultrasound, and optical sensors such as digital video cameras and LIDAR.

Some ADAS use digital map data. In particular, the digital map provides information about the road network, road geometry, road conditions and terrain around a vehicle. Digital map data provides valuable information that can not be detected by sensors such as curvature, grade, undetached speed limits, lane boundaries, etc. In addition, a digital map has a certain predictive function that goes beyond the ability of sensors to detect or view the driver.

ADAS, which require digital map data, typically use a geographic database associated with the vehicle's navigation system. This database of the navigation system contains data for mapping a road network in a region, in particular the geographical data including height of roads, intersections, intersection crossing restrictions, street names, speed limits. Exceeding digital map data can be obtained in real time from a cloud or a backend. However, these map data do not contain any acoustic data from which relevant events relevant to ADAS can be deduced.

From the DE 10 2012 016 819 A1 a method is known in which a background noise in the vicinity of a motor vehicle is detected, the detected background noise is compared with at least one danger-indicating noise and a perceptible by a driver notice is generated when the background noise contains a proportion that a sufficient match with the has danger-indicating noise.

From the DE 10 2015 214 575 A1 For example, a system is known which uploads state information and position information acquired in a vehicle to a cloud and sends the state information to another vehicle when that other vehicle is moving toward a position encoded by the position information. Subsequently, an assistance system is activated, which warns the driver of the information receiving vehicle.

DISCLOSURE OF THE INVENTION

It is therefore an object of the invention to provide a method for enriching a cloud with acoustic vehicle environment data, which is reliable and safe.

This object is achieved by a method according to the independent method claim. The dependent claims relate to advantageous embodiments.

One aspect relates to a method for enriching a cloud with vehicle environment data of at least one vehicle. The method includes detecting vehicle acoustic environment data by means of vehicle-mounted sensors. The method also includes searching for relevant patterns in the acquired acoustic environment data, and when a relevant pattern has been detected, classifying the detected relevant pattern as an event, and communicating the event to a cloud and cloud-enriching cloud data with the event , The relevant patterns may relate to typical patterns for a particular acoustic signal, such as a siren. The search can be done with a classification method implemented in one or more on-board processors. For this purpose, a comparison of the detected acoustic environment data or the acoustic signal can be done with a database containing a variety of relevant patterns. A relevant pattern may be assigned one or more events which are determined by the classification method can. The cloud data to be enriched relate in particular to vehicle environment data in the form of digital map data. If, for example, a siren is recognized as an event, the digital map data with the siren or a vehicle that has activated a siren can be added in real time. In particular, it may be provided that the method includes determining the distance and / or direction of the event with respect to the vehicle for determining position data of the event, transmitting the event with associated position data to the cloud and cloud-enriching cloud data. Data with the event and associated position data includes. Thus, for example, a specific spatial position on a digitized roadway can be enriched with a vehicle with activated siren horn.

This enriched cloud data, i. in particular the event enriched digital map model can be made available to other road users. Accordingly, the method may include the cloud-side transmission of event-enriched cloud data to at least one other vehicle or at least one further road user. Thus, road users can be informed of the event before they can perceive it acoustically. Thus, the acoustic range of a siren can be extended without the actual volume must be increased. Road users can therefore anticipate to set free a roadway. This is especially relevant before traffic lights, as it is often not possible to avoid the red light, without ignoring the red light and rescue vehicles are obstructed.

Apart from the event itself, it may also be appropriate to transmit the underlying acoustic environment data to the cloud. For example. Missing engine noise may indicate a traffic jam. On the other hand, missing engine noise can also indicate a, for whatever reason, little busy road. Final allocation can be done by matching with other data in the cloud. Accordingly, the method may include communicating environmental event data underlying the event to the cloud and cloud-enriching cloud data with the environmental event data underlying the event and the event.

The acoustic environment data underlying the event may be sent from or to the cloud by road users. Accordingly, the method may include the cloud-side transmission of the event and the acoustic environment data underlying the event to at least one other vehicle or at least one further road user.

The acoustic environment data can be acoustically reproduced in the interior of a vehicle, in particular at an increased volume. The cloud-related vehicle acoustic environment data may be acoustically reproduced in the interior of other vehicles. On the basis of the original data, the driver himself can derive the underlying in the current context.

One aspect relates to improving the perception of a reproduction of the acoustic environment data in the interior of a vehicle. For this purpose, the volume of another acoustic playback bspw. A radio announcement or music playback can be reduced. Accordingly, the method may include adjusting down the volume of another acoustic replay in the interior of the at least one other vehicle.

The relevant patterns in the acoustic vehicle environment data may in particular be one of the following: siren, horn, shouts of passers-by, silence, drive noise, disc clanking, thunder, patter of rain on the road. Events can be assigned to the relevant patterns. For example. Due to the decided noise of an accelerating motorcycle, this event can be attributed to the event "motorcycle accelerated". If this information is distributed via the cloud to various nearby road users, they can adjust to this and refrain from own overtaking maneuvers, which could lead to a collision with the motorcycle, soon, so that the chances of survival of the motorcyclist may increase.

One aspect relates to improving the classification of the relevant patterns in the vehicle environment data. This enhancement may be accomplished by considering other non-acoustic, e.g., vehicle optical environment data. For example. For example, for the relevant pattern for silence, the events traffic jam or low traffic road can be verified by evaluation of optical vehicle environment data. The vehicle optical environment data may be captured by imaging sensors and provided to and processed by the cloud or backend.

One aspect relates to a device having a processor configured to perform a method as described above. The method may be embodied in program instructions that the processor executes.

list of figures

• 1 ein Blockschaltbild eines bevorstehenden Überholvorgangs; und
• 2 ein Flussdiagramm einer Ausgestaltung des Verfahrens

The invention will be described below with reference to embodiments with reference to figures. Show it

• 1 a block diagram of an upcoming overtaking process; and
• 2 a flowchart of an embodiment of the method

DETAILED DESCRIPTION

1 shows a block diagram 100 an overtaking process, being a vehicle 108 two more trucks 104 . 106 want to overtake, which in front of the vehicle 108 on the same lane 102 drive. The road surface is relatively narrow, so that in particular the preceding truck 104 relatively close to the middle of the road. Noticed the driver of this truck 104 an overtaking vehicle is not in time in the rearview mirror and therefore does not lead on time to the right, so this may cause the overtaking vehicle first decelerate, which delays the overtaking process and unnecessarily endanger the driver of the overtaking vehicle. An upcoming overtaking of vehicle 108 Announces acoustically by turning the engine and possibly. Previous downshifting on. Can this sound or this information to the driver of the preceding truck 104 be transmitted, so this can adjust to a passing process and continue to the right at least in the middle of the right lane 102 go, leaving the vehicle 108 if it is on the left lane not by the truck 104 is hampered.

One method of doing this will be further discussed with reference to 2 described. Part of the process steps is in the truck 106 , another part in the cloud 110 and further part in the truck 104 executed. This is indicated by the corresponding reference numerals in FIG 2 and the dashed vertical lines indicated.

Lorry 106 detects the acceleration sound of vehicle 108 as vehicle acoustic environment data with one or more acoustic sensors, see step 202 , In one or more processors one in the truck 106 When the analysis unit is installed, the vehicle acoustic environment data are searched for relevant patterns, see step 204 , It checks if a relevant pattern has been detected, see step 206 So here's a decided acceleration sound. If this is the case, then in step 208 assigned an event to the pattern, here the event "imminent overtaking". Lorry 106 is equipped with a transceiver, which is connected to the analysis unit and also communicatively to a transceiver in the cloud 110 is coupled. The transceiver transmits the identified event together with the underlying acoustic signal and the detected position of the sound source with respect to the position of the truck 106 as well as its position to the cloud. In addition, also trucks 104 communicative with the cloud 110 connected, either by a dedicated communication link or by receiving broadcast signals.

The cloud 110 receives the event with associated data and thereby enriches its own database in which a digital map model is stored, see step 210 , The cloud 110 detects all vehicles at a predetermined distance around the vehicle 106 and transmits the event with position and timestamp to these vehicles. Alternatively, the cloud transfers 110 the event broadcast. On the basis of the transmitted position, the receiving vehicles can decide on the basis of their own position whether the event is of importance to them. Here receives lorries 104 the event. Then controls a control unit in trucks 104 the volume of a music playback down, see step 212 , About a speaker in the cab or a display in the dashboard or in the windshield is then the driver of the truck 104 warned of a possible overtaking, see step 214 , Then the driver of the truck 104 First check in the rearview mirror if there is an overtaking process by vehicle 108 actually occur and react accordingly. Alternatively or additionally, it is also possible that due to the received event directly in the vehicle control of trucks 104 is intervened and thus this on the center or the right edge of the right lane 102 is steered.

Claims (9)

A method for enriching a cloud with vehicle acoustic environment data of at least one vehicle, the method comprising: detecting vehicle acoustic environment data by vehicle-side sensors; Searching for relevant patterns in the acquired acoustic vehicle environment data, and when a relevant pattern has been detected: classifying the identified relevant pattern as an event; Transmitting the event and location data of the event to a cloud; Cloud-enriching cloud data with the event and location data; Cloud-side transmission of the enriched cloud data to other vehicles at a predetermined distance to the position data, wherein the enriched data in addition to the event and the position data additionally includes a time stamp. Method according to Claim 1 , the method further comprising: transmitting environmental acoustic data underlying the event to the cloud; and cloud-side provisioning of cloud data with the environmental event data underlying the event and the event. Method according to Claim 1 or 2 where the cloud data is vehicle environment data or a digital map model. Method according to one of Claims 1 to 3 wherein the method further comprises: determining the distance and / or direction of the event with respect to the vehicle to determine the positional data of the event; Transmitting the event with associated location data to the cloud; and cloud-side enriching of cloud data with the event and associated location data. Method according to one of Claims 1 to 4 the method further comprising: providing cloud-side cloud data enriched with the event and the environmental event data underlying the event to at least one other vehicle. Method according to one of Claims 2 to 5 wherein the method comprises, as a function of the event, audible reproduction of acoustic vehicle environment data underlying the event in the interior of the vehicle and / or in the interior of the at least one further vehicle. Method according to Claim 6 wherein the method further comprises: lowering the volume of another acoustic reproduction in the interior of the vehicle or in the interior of the at least one other vehicle. Method according to one of Claims 1 to 7 , where the relevant pattern is one of the following: siren, horn, shouts of passers-by, silence, drive noise, disc clanking, thunder, patter of rain on a road surface. Method according to Claim 8 , the method further comprising: detecting non-acoustic vehicle environment data with at least one other sensor; and classifying the vehicle acoustic environment data by matching with the vehicle optical environment data.

Images