Radar measuring device with multiple radar chips

Abstract

Radar system (159) with a radar measuring device (100), set up for process automation in an industrial or private environment, comprising the radar measuring device (100): a first radar chip (120), configured to emit a first radar measurement signal; a second radar chip (130), configured to emit a second radar measurement signal; wherein the first radar chip (120) and the second radar chip (130) are arranged apart from each other; a lens (155) which is arranged in the beam path of the first radar chip (120) and the second radar chip (130), such that the first radar measurement signal emitted by the first radar chip (120) is focused by the lens (155) in a first direction and the second radar measurement signal emitted by the second radar chip (130) is focused by the lens (155) in a second direction which is different from the first direction; wherein the radar system (159) includes a circuit (161) which includes a machine learning-based evaluation module (162) which is set up to evaluate the radar measurement signals detected by the radar measuring device (100).

Description

Technical field

[0001] The present invention relates to industrial process measurement technology. In particular, the present invention relates to a radar measuring device set up for process automation in an industrial or private environment, a radar system with such a measuring device, a training method for training, by means of machine learning, an evaluation module of a circuit of such a radar system, an evaluation method for evaluating first and second radar measurement signals, a computer program product, training data for training, by means of machine learning, an evaluation module of a circuit of such a radar system, and several uses. background

[0002] Radar systems are known that use multiple radar chips with a common local oscillator clock (synchronized) and perform signal processing using digital beamforming.

[0003] DE 10 2022 128 393 A1 describes a radar-based level measuring device for determining location-specific level values ​​with a collecting lens.

[0004] DE 10 2007 039 834 A1 describes a radar sensor device for measuring the speed of a vehicle.

[0005] DE 10 2013 208 719 A1 describes a radar sensor with a radar probe device and a radar lens, wherein the individual transmitting antennas are simultaneously designed as receiving antennas.

[0006] DE 10 2004 059 332 A1 describes a radar transceiver comprising at least one oscillator that can be tuned with a control voltage. Summary

[0007] Against this background, it is an objective of the present disclosure to provide an alternative radar measuring device, in particular for three-dimensional radar measurement, which is characterized by a simplified design.

[0008] This problem is solved by the subject matter of the independent claims. Further developments of the present disclosure are set forth in the dependent claims and the following description of embodiments.

[0009] A first aspect of the present disclosure relates to a radar system with a radar measuring device, which is set up for process automation in an industrial or private environment. The radar measuring device has a first radar chip and a second radar chip. The first radar chip is configured to emit a first radar measurement signal in the direction of a product surface / scenery. The second radar chip is configured to emit a second radar measurement signal in the direction of the product surface / scenery.

[0010] The two radar measurement signals are reflected at the surface of the fill material / scenery and received by the corresponding receiving antennas of the respective radar chips.

[0011] The first and second radar chips are arranged at a distance from each other. A lens is provided, positioned in the beam path of both the first and second radar chips, such that the first radar measurement signal emitted by the first radar chip is focused by the lens in a first direction, and the second radar measurement signal emitted by the second radar chip is focused by the lens in a second direction, which is different from the first direction.

[0012] By positioning several radar chips on a common surface, in particular on a surface that is rotationally symmetric to the optical axis of the lens, each combination of transmitting and receiving antennas of a radar chip results in a different antenna characteristic, so that direction-dependent distance information of the scene in front of the lens, determined by the position of the radar chip, can be obtained.

[0013] The described radar measuring device has a simplified design compared to radar systems whose multiple chips are synchronized with a common local oscillator clock and perform signal processing using digital beamforming.

[0014] The term "process automation in industrial environments" refers to a subfield of engineering that encompasses measures for operating machines and systems without human intervention. One goal of process automation is to automate the interaction of individual components within a plant in industries such as chemicals, food, pharmaceuticals, petroleum, paper, cement, shipping, or mining. A wide variety of sensors can be used for this purpose, specifically adapted to the requirements of the process industry, such as mechanical stability, resistance to contamination, extreme temperatures, and extreme pressures. Measurement data from these sensors is typically transmitted to a control room where process parameters such as fill level, limit level, flow rate, pressure, and density are monitored, and settings for the entire plant can be adjusted manually or automatically.

[0015] A subfield of process automation in industrial environments concerns the logistics automation of plants and supply chains. Using distance and angle sensors, logistics automation automates processes inside or outside a building or within a single logistics facility. Typical applications include baggage and freight handling at airports, traffic monitoring (toll systems), retail, parcel distribution, and building security (access control). What these examples have in common is that the respective application requires presence detection combined with precise measurement of the size and position of an object.For this purpose, sensors based on optical measurement methods using lasers, LEDs, 2D cameras or 3D cameras that detect distances according to the time-of-flight (ToF) principle can be used.

[0016] Another subfield of process automation in industrial settings concerns factory / production automation. Applications for this can be found in a wide variety of industries, such as automotive manufacturing, food production, pharmaceuticals, and packaging in general. The goal of factory automation is to automate the production of goods using machines, production lines, and / or robots, i.e., to allow it to proceed without human intervention. The sensors used here and the specific requirements regarding measurement accuracy for capturing the position and size of an object are comparable to those in the previous example of logistics automation.

[0017] The terms used in the claims should be interpreted in such a way as to give them the broadest possible reasonable interpretation in accordance with the foregoing description. For example, the use of the article "a" or "the" when introducing an element should not be interpreted as excluding a multitude of elements. Likewise, the mention of "or" should be interpreted as including a multitude of elements, so that the mention of "A or B" does not exclude "A and B" unless it is clear from the context or the preceding description that only one of A and B is meant.Furthermore, the phrase "at least one of A, B, and C" is to be understood as one or more elements from a group of elements consisting of A, B, and C, and not as requiring at least one of each of the listed elements A, B, and C, regardless of whether A, B, and C are related as categories or otherwise. Moreover, the mention of "A, B, and / or C" or "at least one of A, B, or C" should be interpreted as encompassing each individual unit of the listed elements, e.g., A; each subset of the listed elements, e.g., A and B; or the entire list of elements A, B, and C.

[0018] According to a further embodiment of the present disclosure, the first radar chip and the second radar chip are arranged on a common substrate.

[0019] According to a further embodiment of the present disclosure, the first radar chip and the second radar chip are arranged on a common plane.

[0020] According to a further embodiment of the present disclosure, the first radar chip and the second radar chip are arranged on a curved surface.

[0021] According to a further embodiment of the present disclosure, the second radar chip is arranged at an angle to the first radar chip. For example, the first radar chip and the second radar chip are each aligned with the center of the lens.

[0022] According to a further embodiment of the present disclosure, the first radar chip and the second radar chip each have a transmitting antenna and at least one receiving antenna.

[0023] According to a further embodiment of the present disclosure, the first radar chip and the second radar chip are not synchronized by a common local oscillator signal from which the respective transmit or receive signal is derived.

[0024] According to a further embodiment of the present disclosure, the first radar chip and the second radar chip are spaced no more than 10 mm apart. Here, the distance is understood to be the distance between the centers of the first and second radar chips.

[0025] According to another embodiment, the radar measuring device has at least one additional radar chip besides the first radar chip and the second radar chip.

[0026] The radar chips can be arranged in a two-dimensional array, for example, 2 x 2 chips, 3 x 3 chips, or 4 x 4 chips. More radar chips are also possible. Furthermore, the radar chips can be arranged linearly, along a line, or within a circle (not within a square). They can also be arranged rotationally symmetrically around the optical axis of the lens.

[0027] In the case of a linear arrangement of radar chips, the lens can be a cylindrical lens positioned parallel to the line above it. This allows the radar chips to scan an entire line on the surface of the product / scene.

[0028] Optionally, a so-called Time-of-Flight (ToF) sensor, which operates in the optical wavelength range, can be provided, which supplies reference data for training a neural network based on the radar signals.

[0029] The radar measuring device described above can also be called a "radar camera" and, especially using the time-of-flight sensor, it can learn how to combine the measurement signals acquired by the individual radar chips to gain further insights.

[0030] For example, the evaluation module can include training data based at least on the acquired first and second radar measurement signals. Accordingly, the acquired radar measurement signals and other measurement signals can be used to train the evaluation module using machine learning.

[0031] For example, the evaluation module can be a module trained locally on the circuit using machine learning. Alternatively or additionally, the evaluation module can be pre-trained, particularly on another circuit or computer that has been used in a radar system. An advantage of initially or subsequently training the evaluation module on the radar system's circuit is that the training can be specific to the radar's operating conditions. This avoids deviations or errors arising from variations between different operating conditions, or due to differences in design, size, quality, etc., which could potentially reduce the precision of the evaluation module's analysis. Operating conditions can include, for example, the total operating time, the operating time per period (e.g., per day), the ambient temperature, ambient pressure, etc.encompass and thus have an influence on the radar measurement signals and therefore on the evaluation.

[0032] For example, the circuit can be configured to store the evaluation module's training data in the circuit's own memory. The circuit can also be configured to compress the training data and / or delete it after a predetermined time and / or after new training data has been generated. This significantly reduces the memory requirement on the circuit, enabling local training of the evaluation module, even if the circuit is, for example, a low-cost microprocessor with limited memory. When compressing the training data, identical data can be grouped together to avoid duplicate data storage. New training data can replace previously stored training data after it has been deleted or can directly overwrite it.Alternatively, the training can also take place in a cloud, whereby the training data can be transferred to the cloud, and, for example, a partially or fully trained ML model can be made available on the evaluation module.

[0033] According to the present invention, the radar system has a circuit which includes a machine learning-based evaluation module which is set up to evaluate the radar measurement signals detected by the radar measuring device.

[0034] Another aspect of the present disclosure relates to a training method for training, by means of machine learning, an evaluation module of a circuit of a radar system described above and below.

[0035] Another aspect of the present disclosure relates to an evaluation method for evaluating first and second radar measurement signals which were detected by a radar measuring device of a radar system described above and below.

[0036] Another aspect of the present disclosure relates to a computer program product, comprising instructions which, when the program is executed by a computer, cause it to perform the training procedure and / or the evaluation procedure described above.

[0037] Another aspect of the present disclosure concerns training data for training, by means of machine learning, an evaluation module of a circuit of a radar system described above and below, wherein the training data are based at least on captured first and second radar measurement signals.

[0038] Another aspect of the present disclosure relates to the use of a radar measuring device, described below, for level measurement or area monitoring.

[0039] Further embodiments of the present disclosure are described below with reference to the figures. Where the same reference numerals are used in the following figure descriptions, they denote identical or similar elements. The representations in the figures are schematic and not to scale. Brief description of the characters Fig. Figure 1 shows a radar system according to an embodiment of the present disclosure. Fig. Figure 2 shows a flowchart of a process according to an embodiment of the present disclosure. Detailed description of embodiments

[0040] Fig. Figure 1 shows a radar system 159, which includes a radar measuring device 100 and a circuit 161 with an evaluation module 162. The circuit 161 is, for example, a server or other computer equipment that is connected to or embedded in a cloud 160, either via cable or wirelessly. Alternatively, the circuit 161 can be a microprocessor on which the evaluation module 162 runs. The evaluation module can be a conventional [missing information - likely a specific type of microprocessor]. A cloud connection is not strictly necessary. It is sufficient if the evaluation unit evaluates the radar signals and makes the acquired data available via radio, cable, or as a switching output. For example, for area monitoring: persons in the danger zone / no persons in the danger zone.

[0041] The radar measuring device 100 has several radar chips 120, 130, 140, above which is a lens 155 which focuses the radar measurement signals emitted by the individual radar chips in a different direction and reflects the radar measurement signals back to the corresponding radar chip 120, 130, 140.

[0042] The radar system 159 therefore has at least one lens 155, which can be convex, plano-convex, cylindrical or otherwise designed, and at least two radar chips 120, 130, each of which has at least one transmitting antenna 121, 131 and one receiving antenna 122, 132 and can be designed as Antenna on Chip (AoC) or Antenna in Package (AiP).

[0043] The at least two radar chips are essentially arranged in a surface rotationally symmetric to the optical axis 150 of the lens 155 and are not synchronized by a common local oscillator signal from which the respective transmit or receive signal is derived.

[0044] The area 160 mentioned above can be a flat surface or a curved surface (for example, convex or concave). Furthermore, the transmitting and receiving antennas of the radar chips used are, for example, no more than 5 mm apart.

[0045] Due to the different positioning of the individual radar chips 130, 140, and 120, each combination of transmitting and receiving antennas of a radar chip results in a different antenna characteristic. This means that the distance information of the scene in front of the lens is directional and dependent on the position of the radar chip. Therefore, it is possible to calculate the topography of a bulk material surface by processing the individual direction-dependent distance information. Furthermore, area monitoring is also possible.

[0046] Since the received signal for a given angular direction does not need to be synthesized from a multitude of phase-stable received signals using digital beamforming, as is the case with 3D radar systems, the proposed system is less prone to errors. Furthermore, the processing power requirements of the microprocessor used for signal processing are not as high as for radar systems that employ digital beamforming.

[0047] The acquired radar measurement signals can be evaluated using a neural network, artificial intelligence, or machine learning, for example, in a cloud or locally on the connected circuit 161 with the evaluation module 162. In particular, more than two radar chips can be used. To improve the resolution of the measurements, the number of radar chips used can be increased.

[0048] The in Fig. The radar system 159 shown in Figure 1 has at least two radar chips 120, 130, 140, each with at least one transmitting antenna 121, 131, 141 and at least one receiving antenna 122, 132, 142. The transmitting and receiving antennas are essentially positioned on a surface 160 that is rotationally symmetrical with respect to the optical axis 150 of the lens 155. In the simplest case, the surface is a plane.

[0049] In a first rough geometric approximation, the achievable angular deflection θ 165 to the optical axis of the lens is: θ=tan−1xf where x corresponds to the lateral offset 170 of the center point of the transmitting and receiving antennas to the optical axis of the lens and f corresponds to the focal length 175 of the lens.

[0050] Due to the different positioning of the individual radar chips, each results in a different effective direction-dependent antenna characteristic 125, 135, 145 through the combination of the corresponding transmit and receive characteristics (123, 133, 143 or 124, 143, 144).

[0051] This allows each radar chip to capture at least one direction-dependent distance measurement of the scene in front of lens 155, with the direction being determined by the position of the radar chip. Thus, by processing the individual direction-dependent distance measurements, it is possible to determine the topography of the bulk material surface. Furthermore, area monitoring is also possible, for example.

[0052] To create a signal processing model, an additional sensor 180, for example an optical Time-of-Flight (ToF) sensor, can be used at least temporarily, which enables the measurement of direction-dependent distance information and can serve as a reference for the model to be created.

[0053] The radar system 159 comprises a circuit 161 with a storage unit (not shown) and an evaluation module 162. The circuit 161 may include a processor (not shown) that executes and / or is part of the evaluation module 162. The radar measuring device 100 and other sensors (not shown) may be wirelessly or wired connected to the circuit 161, enabling the evaluation module 162 to evaluate the detected radar measurement signals and the other measurement signals from the other sensors directly or indirectly (based on derived data).

[0054] The evaluation module 162 is based on machine learning. The acquired radar measurement signals and other measurement signals (or data derived from them) can serve as training data for the evaluation module 162 or one of its machine learning models. The machine learning model of the evaluation module 162 can be a pre-trained model and / or a model trained on the radar system 159. Accordingly, the machine learning model may, for example, have been pre-trained on one or more other radar systems 159 and then transferred to the radar system 159 shown. It can also be further trained on the radar system 159 using the training data. Alternatively, the machine learning model may not be pre-trained and may only be trained on the radar system 159 using the training data.

[0055] A training procedure can be performed on the radar system 159 shown, whereby the evaluation module 162 is trained within the framework of the training procedure to evaluate the acquired radar measurement signals and any other acquired measurement signals. The evaluation module 162 can be trained, in particular, to generate condition data, predictive maintenance data, and / or other data.

[0056] Accordingly, an evaluation procedure can be executed by the trained machine learning model of evaluation module 162 to evaluate newly acquired radar measurement signals and / or other acquired measurement signals, thereby generating condition data, predictive maintenance data, and / or other data. Furthermore, these newly acquired radar measurement signals and / or other measurement signals can be used to further train evaluation module 162.

[0057] Fig.Figure 2 shows a flowchart of a method according to an embodiment of the present disclosure. In a first step 200, the at least one direction-dependent distance information from the first radar chip is determined. In a second step 201, the direction-dependent distance information from the other radar chips is determined. In a next step 202, a topography is calculated based on the acquired signals. Alternatively, a change in the scenery is transmitted for area monitoring by comparison with previously acquired distance information. In a final step 203, the topography is output, or further steps for area monitoring are initiated, such as the output of an alarm.

Claims

[1] Radar system (159) with a radar measuring device (100), set up for process automation in an industrial or private environment, comprising the radar measuring device (100): a first radar chip (120), configured to emit a first radar measurement signal; a second radar chip (130), configured to emit a second radar measurement signal; wherein the first radar chip (120) and the second radar chip (130) are arranged apart from each other; a lens (155) which is arranged in the beam path of the first radar chip (120) and the second radar chip (130), such that the first radar measurement signal emitted by the first radar chip (120) is focused by the lens (155) in a first direction and the second radar measurement signal emitted by the second radar chip (130) is focused by the lens (155) in a second direction which is different from the first direction; wherein the radar system (159) includes a circuit (161) which includes a machine learning-based evaluation module (162) which is set up to evaluate the radar measurement signals detected by the radar measuring device (100). [2] Radar system (159) according to claim 1, wherein the first radar chip (120) and the second radar chip (130) are arranged on a common substrate (160). [3] Radar system (159) according to claim 1 or 2, wherein the first radar chip (120) and the second radar chip (130) are arranged on a common plane. [4] Radar system (159) according to claim 1 or 2, wherein the first radar chip (120) and the second radar chip (130) are arranged on a curved surface. [5] Radar system (159) according to one of the preceding claims, wherein the second radar chip (130) is arranged tilted relative to the first radar chip (120). [6] Radar system (159) according to one of the preceding claims, wherein the first radar chip (120) and the second radar chip (130) are each aligned to the center of the lens (155). [7] Radar system (159) according to one of the preceding claims, wherein the first radar chip (120) and the second radar chip (130) each have a transmitting antenna (121, 131) and at least one receiving antenna (122, 132). [8] Radar system (159) according to one of the preceding claims, wherein the first radar chip (120) and the second radar chip (130) are not synchronized by a common local oscillator signal from which the respective transmit or receive signal is derived. [9] Radar system (159) according to one of the preceding claims, wherein the first radar chip (120) and the second radar chip (130) are spaced no more than 10 mm apart. [10] Radar system (159) according to one of the preceding claims, wherein the evaluation module (162) comprises training data which is based at least on the radar measurement signals. [11] Radar system (159) according to one of the preceding claims, wherein the evaluation module (162) is an evaluation module (162) trained by machine learning. [12] Training method for training, by means of machine learning, an evaluation module (162) of a circuit (161) of a radar system (159) according to one of the preceding claims. [13] Evaluation method (300) for evaluating first and second radar measurement signals which were detected by a radar measuring device (100) of a radar system (159) according to one of the preceding claims. [14] Computer program product comprising instructions which, when the program is executed by a computer, cause it to perform the training method according to claim 12 or the evaluation method according to claim 13. [15] Training data for training, by means of machine learning, an evaluation module (162) of a circuit (161) of a radar system (159) according to one of claims 1 to 11, wherein the training data are based at least on detected first and second radar measurement signals. [16] Use of a radar measuring device (100) of a radar system (159) according to any one of claims 1 to 11 for level measurement. [17] Use of a radar measuring device (100) of a radar system (159) according to any one of claims 1 to 11 for area monitoring.

Images