Procedure for commissioning a multi-sensor network

Abstract

Method for commissioning a multi-sensor network (1), in particular a vehicle (2), wherein, depending on measurement data from a plurality of sensor units (3, 4, 5, 6) of the multi-sensor network (1), installation positions of the sensor units (3, 4, 5, 6), in particular on the vehicle (2), are determined, wherein different measurement data of the sensor units (3, 4, 5, 6) are generated by acquiring at least one different measurement target (7, 8, 9, 10) for each sensor unit (3, 4, 5, 6), wherein the at least one measurement target (7, 8, 9, 10) is designed as a calibration object for calibrating the sensor units (3, 4, 5, 6).

Description

The invention relates to a method for commissioning a multi-sensor network. Furthermore, the invention relates to a control device and to a vehicle with a corresponding control device. The invention also relates to a system for commissioning a multi-sensor network and to a corresponding computer program. Methods for commissioning multi-sensor networks are known from the prior art. DE 10 2014 017 917 B3 discloses a method for configuring at least one radar sensor installed at one of several installation positions in a motor vehicle with respect to its installation position and a motor vehicle. DE 10 2020 210 158 A1 discloses a sensor arrangement of a vehicle. A method for commissioning a multi-sensor network, particularly a vehicle, is proposed. In this method, the installation positions of the sensor units, especially on the vehicle, are determined based on measurement data from a plurality of sensor units of the multi-sensor network. The vehicle is preferably designed as an automated guided vehicle (AGV). An "automated guided vehicle" is understood to mean, in particular, a vehicle with one of the automation levels 1 to 5 of the SAE J3016 standard. Specifically, the automated guided vehicle has the technical equipment required for these automation levels. This technical equipment includes, in particular, environmental sensors for detecting the vehicle's surroundings, such as radar sensors, lidar sensors, cameras and / or acoustic sensors, control units, or the like. Preferably, the sensor units of the multi-sensor network include the environmental sensors or at least partially incorporate them. Alternatively or additionally, it is conceivable that the sensor units of the multi-sensor network include, or at least partially incorporate, interior sensors for detecting the vehicle's interior. Preferably, the automated vehicle is configured as a land vehicle. In particular, the automated vehicle can be configured as a passenger car, a truck, a construction vehicle, an agricultural vehicle, or any other vehicle that would be considered appropriate by a specialist. Alternatively, the automated vehicle can also be configured as an aircraft, for example, a drone, an airplane, a helicopter, a vertical take-off and landing aircraft, or the like, or as a watercraft, for example, a ship, a submarine, or the like. Preferably, the vehicle includes the multi-sensor network.Alternatively, it is conceivable that the multi-sensor network is designed separately from a vehicle, for example as a monitoring system for monitoring a building, as a traffic monitoring system, as an early warning system for detecting natural events, or as another multi-sensor network that would seem useful to an expert. Preferably, the sensor units of the multi-sensor network are configured analogously to one another. In particular, the sensor units are configured as the same type of sensor, for example, all sensor units as radar sensors, all sensor units as lidar sensors, or all sensor units as ultrasonic sensors. Alternatively, it is conceivable that the sensor units are configured differently from one another, for example, one sensor unit as a radar sensor, another sensor unit as a lidar sensor, and an additional sensor unit as an ultrasonic sensor. Preferably, different multi-sensor networks, especially those of a vehicle, can be put into operation using the method, for example, first all radar sensors of the vehicle and then all lidar sensors of the vehicle. Preferably, the multi-sensor network comprises, in particular, wired and / or wireless data connections that connect the sensor units to each other and / or to a control device, especially of the vehicle, via signal transmission. In particular, the data connections are designed for bidirectional communication. For example, control signals can be transmitted to the sensor units via the data connections, and the sensor units can transmit measurement data generated by sensing their environment. In particular, a vehicle control unit, e.g., an electronic control unit, can include or at least partially constitute the control or regulation device. A control unit prepares sensor data as input signals, processes these using the control or regulation device, for example, a programmable logic device, an FPGA or ASIC chip, or a computer platform, and provides logic and / or power levels as a control or regulation signal. The control or regulation signal can, for example, also be used to control or regulate actuators for longitudinal and / or lateral guidance of the vehicle in order to keep the vehicle in its lane and / or predict a trajectory. The control unit is preferably integrated into the vehicle's electrical system, for example, a CAN bus. The electrical system can, in particular, include or at least partially constitute the data connections of the multi-sensor network.The control unit is, for example, an electronic control unit for automated driving functions, called a Domain ECU in English. In particular, the control unit can be an ADAS (advanced driver assistance system) / AD (autonomous driving) Domain ECU for assisted to fully automated, i.e., autonomous, driving. The control device is implemented, for example, as a system-on-a-chip with a modular hardware concept. This means that all, or at least a large part, of the functions are integrated onto a single chip and can be expanded modularly. The chip can be integrated into the control unit. The control device includes, for example, a multi-core processor and memory modules. The multi-core processor is configured for signal / data exchange with storage media. For example, the multi-core processor includes a bus system. The memory modules form a working memory. Examples of memory modules are RAM, DRAM, SDRAM, or SRAM. In a multi-core processor, several cores are arranged on a single chip, that is, a single semiconductor device.Multi-core processors achieve higher computing power and are more cost-effective to implement on a single chip compared to multi-processor systems where each individual core is located in a processor socket and the individual processor sockets are arranged on a motherboard. According to one aspect of the invention, the control device comprises at least one central processing unit, abbreviated CPU. The control device preferably includes at least one graphics processing unit (GPU). GPUs have a special microarchitecture for parallel processing. According to one aspect of the invention, the GPU includes at least one processing unit specifically designed to perform tensor and / or matrix multiplication. Tensor and / or matrix multiplication are the core computational operations for deep learning. According to another aspect of the invention, the control device also includes hardware accelerators for artificial intelligence, such as deep learning accelerators. According to a further aspect of the invention, a classifier is provided using the CUDA programming language. This allows the GPU to directly execute sections of the classifier's software code.Preferably, the control device or control unit is configured to be modularly expanded with several, for example at least four, such chips. Preferably, the control device is designed to at least partially execute the commissioning method for the multi-sensor network. "Designed" is understood to mean, in particular, specifically programmed, specially equipped, and / or specially designed. The phrase "designed" means, in particular, that the object performs the function in at least one operating state. Specifically, the commissioning method for the multi-sensor network is at least partially computer-implemented. Preferably, the sensor units are arranged at different installation positions, particularly on the vehicle. For example, it is conceivable that a first sensor unit is arranged at the front of the vehicle, a second sensor unit at the rear, a third sensor unit on one side of the vehicle, and a fourth sensor unit on another side of the vehicle facing away from the first. In normal operation, the control device is preferably designed to control or regulate functions, particularly autonomously, based on measurement data generated by the sensor units. For example, the control device can be designed to initiate emergency braking of the vehicle based on the detection of a pedestrian by a sensor unit in front of the vehicle.To effectively control or regulate the functions, the control device must know, in particular, the installation position of each sensor unit. Specifically, the control device can then meaningfully evaluate and process the measurement data from a sensor unit depending on its installation position. For example, if a pedestrian is detected by a sensor unit at the rear of the vehicle, it is not necessary to initiate emergency braking when the vehicle is moving forward. Preferably, the multi-sensor network, and in particular the sensor units, is / are ready for regular operation or further calibration after commissioning. Preferably, measurement data, especially test measurement data, is requested from the sensor units, particularly by the control device. Preferably, depending on the request, the sensor units acquire data about their surroundings in order to generate the measurement data. The measurement data is preferably transmitted from the sensor units to the control device. Preferably, the measurement data is evaluated, particularly by the control device, especially with regard to device identifiers of the sensor units and / or objects detected by the sensor units. Preferably, depending on the evaluation, the installation positions of the sensor units are determined, particularly by the control device. The inventive design of the method for commissioning a multi-sensor network advantageously enables a particularly user-friendly commissioning process. Advantageously, the need for different sensor variants and / or sensor connectors can be eliminated. Advantageously, a commissioning method for the multi-sensor network can be provided that exhibits low susceptibility to errors and requires minimal coordination. Furthermore, it is proposed that the measurement data include a unique device identifier assigned to the sensor unit generating the measurement data, wherein the measurement data are assigned to the sensor units depending on the device identifiers. In particular, the sensor units have different device identifiers from one another. Preferably, the device identifiers are transmitted together with the measurement data. In particular, the device identifier of a sensor unit is transmitted in a data packet with the measurement data generated by the sensor unit, especially in a header of the data packet. The device identifiers are preferably designed as metadata, in particular as MAC addresses (Media Access Control addresses) of the sensor units. Preferably, the measurement data is assigned to the sensor units by the control device depending on the device identifiers. Advantageously, a user-friendly assignment of the measurement data can be enabled. According to the invention, it is proposed that different measurement data from the sensor units are generated by detecting at least one different measurement target per sensor unit. In particular, each sensor unit is designed to detect at least one measurement target that differs from the measurement targets of the other sensor units. The measurement targets can differ, for example, in size, distance to the sensor unit, speed of movement relative to the sensor unit, or the like. The measurement targets can be configured, in particular, as corner reflectors, target simulators, or as other measurement targets that would appear useful to a person skilled in the art. Depending on the different detected measurement targets, the generated measurement data of the sensor units differ from one another.Preferably, the different measurement targets are assigned to the various installation positions of the sensor units, in particular the control device. Preferably, the measurement data from the sensor units, especially from the control device, are evaluated to determine the respective measured target. In particular, the installation position of the respective sensor unit is deduced based on the measured target, especially from the control device. Advantageously, this enables the generation of unique measurement data for each sensor unit and the determination of the sensor unit installation positions. According to the invention, it is proposed that at least one measurement target is designed as a calibration object for calibration, in particular end-of-line calibration, of the sensor units. Specifically, the sensor units can be calibrated during or after commissioning of the multi-sensor network, particularly using the calibration objects. Commissioning of the multi-sensor network can be carried out at a calibration station, in particular on a vehicle production line. Advantageously, user-friendly multiple use of the measurement targets can be enabled. Furthermore, it is proposed that unique, and in particular predetermined, network identification identifiers be assigned to the sensor units depending on their installation positions. These network identification identifiers are assigned to the sensor units, in particular by the control device. Specifically, the network identification identifiers are transmitted to the sensor units. Preferably, the network identification identifiers are configured as IP addresses. In particular, the network identification identifier of a sensor unit differs from the network identification identifiers of all other sensor units in the multi-sensor network. Preferably, unambiguous communication with the sensor units is possible via the network identification identifiers. Preferably, the network identification identifiers are predefined, in particular stored in a memory unit of the control device.In particular, the predetermined network identification identifiers can already be assigned to sensor units in specific installation positions and then assigned to these sensor units depending on the determined installation positions. Alternatively or additionally, it is conceivable that the network identification identifiers, especially by the control device, are generated depending on the determined installation positions and assigned to the sensor units. Advantageously, this enables unambiguous communication within the multi-sensor network. Furthermore, a control device, particularly for a vehicle, is proposed. This control device is designed to determine the installation positions of at least one of the aforementioned multi-sensor networks, specifically on the vehicle, based on measurement data from multiple sensor units. Advantageously, this allows for user-friendly determination of the sensor unit installation positions. Furthermore, it is proposed that the control device include at least one storage unit, in particular the aforementioned one, in which different measurement targets are stored and assigned to the installation positions of the sensor units. Advantageously, the control device can determine the installation positions of the sensor units independently. Furthermore, a vehicle comprising at least a multi-sensor network and at least one control or regulating device according to the invention is proposed. Advantageously, a vehicle that is user-friendly and easy to operate can be provided. Furthermore, a system for commissioning a multi-sensor network is proposed. The system comprises at least one control device according to the invention and a plurality of different measurement targets. Preferably, the system can be used for calibrating the sensor units, either alternatively or additionally, in addition to commissioning the multi-sensor network. In particular, the system can be designed as a calibration system. Advantageously, a system can be provided that enables user-friendly commissioning of the multi-sensor network. Furthermore, a computer program product for commissioning a multi-sensor network is proposed. The computer program product comprises execution commands which, when executed by a control device according to the invention, cause it to execute a method according to the invention. Advantageously, a computer program product can be provided that enables user-friendly commissioning of the multi-sensor network. The invention is illustrated by an embodiment in the following figures. Figure 1 shows a schematic perspective view of a system according to the invention for commissioning a multi-sensor network, Figure 2 shows a schematic view of a vehicle according to the invention, and Figure 3 shows a schematic flowchart of a method according to the invention for commissioning a multi-sensor network. Fig. 1 shows a perspective schematic representation of a system 13 for commissioning a multi-sensor network 1. A vehicle 2, in particular an automated vehicle, comprises the multi-sensor network 1. The vehicle 2 is exemplified as a land vehicle, in particular a passenger car. The system 13 comprises a control device 11 and a plurality of different measurement targets 7, 8, 9, 10. The multi-sensor network 1 comprises a plurality of sensor units 3, 4, 5, 6. The sensor units 3, 4, 5, 6 are arranged at different installation positions on the vehicle 2. In the present embodiment, a first sensor unit 3 is arranged at the front of the vehicle, a second sensor unit 4 at the rear of the vehicle, a third sensor unit 5 on one side of the vehicle, and a fourth sensor unit 6 on another side of the vehicle facing away from the first side. Due to the perspective view, the fourth sensor unit 6 is not visible in Fig. 1. Each sensor unit 3, 4, 5, 6 is assigned a measurement target 7, 8, 9, 10. In the present embodiment, the first sensor unit 3 is assigned a first measurement target 7, the second sensor unit 4 a second measurement target 8, the third sensor unit 5 a third measurement target 9, and the fourth sensor unit 6 a fourth measurement target 10. The measurement targets 7, 8, 9, 10 differ from one another, in the present embodiment, for example, with regard to size, shape, and distance from the sensor units 3, 4, 5, 6. Each sensor unit 3, 4, 5, 6 is designed to generate measurement data by detecting the respective measurement target 7, 8, 9, 10. The measurement targets 7, 8, 9, and 10 are configured as calibration objects for calibration, in particular end-of-line calibration, of the sensor units 3, 4, 5, and 6. Alternatively or additionally, the system 13 can be used to calibrate the sensor units 3, 4, 5, and 6 for commissioning the multi-sensor network 1. The system 13 can be configured as a calibration system. Fig. 2 shows the vehicle 2 in a schematic representation. The multi-sensor network 1 comprises, here by way of example, wired data connections 14, which connect the sensor units 3, 4, 5, 6 to each other and / or to the control device 11 for signal transmission. The data connections 14 are designed for bidirectional communication. The control device 11 is designed to execute, at least partially, a method for commissioning the multi-sensor network 1, in particular a method that is at least partially computer-implemented. The control device 11 is designed to determine the installation positions of the sensor units 3, 4, 5, 6, particularly on the vehicle 2, based on the measurement data from the majority of sensor units 3, 4, 5, 6 of the multi-sensor network 1. The control device 11 comprises at least one storage unit 12 in which the different measurement targets 7, 8, 9, 10 are stored and assigned to the installation positions of the sensor units 3, 4, 5, 6. Fig. 3 shows a schematic flowchart of the commissioning procedure for the multi-sensor network 1. In this procedure, the installation positions of the sensor units 3, 4, 5, 6, particularly on the vehicle 2, are determined based on the measurement data from the majority of sensor units 3, 4, 5, 6 of the multi-sensor network 1. In a first procedure step 15, the measurement data, particularly test measurement data, are requested from the sensor units 3, 4, 5, 6, especially by the control device 11. In a second procedure step 16, the measurement targets 7, 8, 9, 10 are acquired by the sensor units 3, 4, 5, 6, depending on the request, in order to generate the measurement data. The measurement data is then transmitted from the sensor units 3, 4, 5, 6 to the control device 11. The measurement data includes a unique device identifier assigned to each of the sensor units 3, 4, 5, 6 that generate the measurement data. The sensor units 3, 4, 5, 6 have different device identifiers. In a third process step 17, the measurement data are assigned to the sensor units 3, 4, 5, 6 depending on the device identifiers, in particular by the control device 11. The measurement data from sensor units 3, 4, 5, 6 differ depending on the different measurement targets 7, 8, 9, 10 being measured. The control device 11 is aware of the assignment of these different measurement targets 7, 8, 9, 10 to the various installation positions of the sensor units 3, 4, 5, 6 due to the information stored in the memory unit 12. In a fourth process step 18, the measurement data from sensor units 3, 4, 5, 6 are evaluated, particularly by the control device 11, to determine the respective measurement target 7, 8, 9, 10 being measured. Based on the measurement target 7, 8, 9, 10 being measured, the installation position of the respective sensor unit 3, 4, 5, 6 is then determined, particularly by the control device 11. In a fifth process step 19, unique, in particular predetermined, network identification identifiers are assigned to the sensor units 3, 4, 5, 6 depending on their installation positions, in particular by the control device 11. The network identification identifiers are transmitted to the sensor units 3, 4, 5, 6. Unambiguous communication with the sensor units 3, 4, 5, 6 is possible via the network identification identifiers. A computer program product for commissioning the multi-sensor network 1 includes execution commands which, when the program is executed by the control or regulating device 11, cause it to execute the procedure. Reference sign 1 Multi-sensor network 2 Vehicle 3 Sensor unit 4 Sensor unit 5 Sensor unit 6 Sensor unit 7 Measurement target 8 Measurement target 9 Measurement target 10 Measurement target 11 Control device 12 Storage unit 13 System 14 Data connection 15 Process step 16 Process step 17 Process step 18 Process step 19 Process step

Claims

Method for commissioning a multi-sensor network (1), in particular a vehicle (2), wherein, depending on measurement data from a plurality of sensor units (3, 4, 5, 6) of the multi-sensor network (1), installation positions of the sensor units (3, 4, 5, 6), in particular on the vehicle (2), are determined, wherein different measurement data of the sensor units (3, 4, 5, 6) are generated by acquiring at least one different measurement target (7, 8, 9, 10) for each sensor unit (3, 4, 5, 6), wherein the at least one measurement target (7, 8, 9, 10) is designed as a calibration object for calibrating the sensor units (3, 4, 5, 6). Method according to claim 1, wherein the measurement data comprise a unique device identifier assigned to one of the sensor units (3, 4, 5, 6) generating the measurement data, wherein the measurement data are assigned to the sensor units (3, 4, 5, 6) depending on the device identifiers. Method according to one of the preceding claims, wherein the sensor units (3, 4, 5, 6) are assigned unique, in particular predetermined, network identification identifiers depending on the installation positions of the sensor units (3, 4, 5, 6). Control or regulating device, in particular of a vehicle (2), which is designed to determine, depending on measurement data from a plurality of sensor units (3, 4, 5, 6) of at least a multi-sensor network (1), installation positions of the sensor units (3, 4, 5, 6), in particular on the vehicle (2), wherein different measurement data of the sensor units (3, 4, 5, 6) are generated by detecting at least one different measurement target (7, 8, 9, 10) for each sensor unit (3, 4, 5, 6), wherein the at least one measurement target (7, 8, 9, 10) is designed as a calibration object for calibrating the sensor units (3, 4, 5, 6). Control or regulating device according to claim 4, comprising at least one storage unit (12) in which different measurement targets (7, 8, 9, 10) are stored and assigned to the installation positions of the sensor units (3, 4, 5, 6). Vehicle comprising at least a multi-sensor network (1) and at least one control or regulating device (11) according to claim 4 or 5. System for commissioning a multi-sensor network (1) comprising at least one control or regulating device (11) according to claim 4 or 5 and a plurality of different measurement targets (7, 8, 9, 10). Computer program product for commissioning a multi-sensor network (1), comprising execution commands which, when the program is executed by a control or regulating device (11) according to claim 4 or 5, cause it to execute a method according to one of claims 1 to 3.

Images