Multi-input, multi-control-output (MIMSO) radar

Abstract

The present invention provides a radar for a road vehicle comprising a beamforming network having a first number of input ports and a second number of beamforming ports, the first and second numbers being greater than one, a first number of antennas for receiving signals, each antenna of the first number of antennas being connected to a respective input port of the beamforming network, a second number of receivers, at least two of which are simultaneously operable, each receiver being connected to a respective beamforming port of the second number of beamforming ports, a third number of antennas for transmitting signals, the third number being greater than one, and processing means configured to determine a position of an object relative to the radar based on amplitudes of signals received by the receivers.

Description

[Technical Field]

[0001] This disclosure relates to radar systems and methods for electronically sensing the surrounding environment of a vehicle. In particular, this disclosure is directed toward the improvement of multiple input multiple output (MIMO) radars by using beamforming networks such as Rottman lenses. [Background technology]

[0002] Traffic accidents kill an average of 1.35 million people and injure more than 50 million worldwide every year. Economically alone, traffic accidents are projected to cost $24 trillion by 2030. The majority of traffic accidents are caused by human error, such as lack of caution, poor foresight, and driver inattention. By mitigating human error, more than 90% of traffic accidents can be avoided. From this perspective, automotive radar technology has become widespread in order to eliminate the element of human error from traffic fatalities. In a radar system, an object's location is determined by the time it takes for a signal to be emitted from the radar and reflected back to the radar.

[0003] Disclosures relating to radar systems (e.g., EP3497473A1, EP2527871B1, US9070972B2, US9869762B1, US7924215B2, EP1666914B1, US9041596B2, JP3256374B2, JP2002228749A, JP2004226158A, US6795012B2, JP2017215328A, EP1742081B1, DE10354872B4, US6275180B1, US8686894B2, CA2901610C, WO2016146666), more specifically disclosures relating to MIMO radar systems (e.g., "Spatial Smoothing for Direction of Arrival Estimation of Coherent Signals (On Numerous patents exist on this topic, including "Spatial smoothing for direction-of-arrival estimation of coherent signals," "Coherence, compressive sensing, and random sensor array" by L. Carin and D. Liu, and "Design of a linear non-uniform antenna array for a 77-GHz MIMO FMCW radar" by C. Schmid et al. (US9664775B2, US2018 / 0267555, US2015 / 0253419A1, US9541639B2). U.S. Patent Publication 2020 / 096626 further describes radar antenna arrays for obstacle detection. However, there is a need for systems that mitigate the effects of human error in automobiles by accurately estimating the position of objects within a suitable time frame without using expensive sensors or processing hardware. [Overview of the Initiative] [Problems that the invention aims to solve]

[0004] This disclosure is directed toward improving the angular resolution for a given number of transmitters and / or receivers. By using this, the number of transmitters and / or receivers required to obtain a given angular resolution can be reduced. Furthermore, this disclosure is directed toward reducing the computational complexity of a MIMO system designed for a given angular resolution. This advantageously reduces the delay time of object detection in MIMO radar. In addition, by reducing the complexity of the system, the complexity of manufacturing is reduced compared to radars with other lenses.

[0005] Another advantage provided by this disclosure is that the system described herein has a flat plate structure that facilitates integration into automotive systems, thereby increasing the applications in which the system can be used and reducing integration costs. Furthermore, the proposed system can be optionally adjusted to place greater emphasis on the areas where it is most needed, for example, providing high angular resolution at boresight (directly in front) while providing low angular resolution at angles greater than ±45°. In addition, while prior art MIMO radar systems generally have a wide field of view by their nature, the proposed system can be optionally adjusted to have a narrow field of view depending on the requirements of the radar system. The field of view (FOV) of a radar is the cone that forms an angle recognizable by the sensor at a given moment. [Means for solving the problem]

[0006] The present invention is directed to a radar for road vehicles, the radar comprising: a beamforming network including a Rotman lens having a first number of input ports and a second number of beamforming ports, wherein the first and second numbers are greater than 1; a first number of antennas for receiving signals, each of which is connected to the respective input port of the beamforming network; a second number of receivers, at least two of which are capable of operating simultaneously, each of which is connected to the respective beamforming port of the second number of beamforming ports; a third number of antennas for transmitting signals, wherein the third number is greater than 1; and processing means configured to determine the position of an object relative to the radar based on the amplitude and phase of the signals received by the receivers. The beamforming network is configured to combine signals received by at least two of a first number of antennas from at least two of a third number of antennas, and the beamforming network is further configured to apply an additional phase term to each signal being combined to form a beam at a desired angle.

[0007] Multiple time-delay equilibrium lines can preferably be used to connect the beam port of a beamforming lens to a receiver. As used herein, the term receiver is used to encompass any hardware suitable for processing signals received from an antenna, which may, for example, be included as part of a transceiver set.

[0008] Similarly, multiple time-delay balance lines may preferably be used to connect a transmitter to each of a third number of antennas. As used herein, the term transmitter is used to encompass any hardware suitable for generating signals to transmit to an antenna, which may, for example, be included as part of a transceiver set.

[0009] Preferably, the first number of antennas are arranged in a line with a predetermined distance d between each antenna, and the third number of antennas are arranged in a line with a distance of N × d between each antenna, where N is the first number.

[0010] As used herein, the term beamforming network refers to any analog network used for beamforming (such as lenses) and any conventional equivalents, such as dielectric lenses, RKR lenses, and Luneberg lenses. However, Rotman lenses have been found to be particularly advantageous beamforming networks for the purposes of this disclosure.

[0011] When in use, the Rottman lenses form a virtual array of Rottman lenses through their use with a third number of antennas, and the Rottman lenses are configured so that the individual beams formed by the Rottman lenses of the virtual array overlap.

[0012] Preferably, the individual beams formed by the Rotman lenses of the virtual array overlap at half the peak value of each beam. Preferably, the beamwidth of the individual beams formed by the virtual array of Rotman lenses is narrower than the beamwidth of the Rotman lens. More preferably, the beamwidth is multiplied by a coefficient inversely proportional to the third number of antennas.

[0013] Optionally, the Rottman lens may be configured so that the beam port is concentrated around the radar's boresight. This is advantageous because it improves the radar's angular resolution around the radar's boresight (i.e., the radar's straight-ahead direction) where angular accuracy is critical. Preferably, the beam formed by the Rottman lens covers a wide field of view (e.g., 120°, i.e., ±60° from the array boresight direction), although other narrower fields of view may be desirable depending on the application.

[0014] Preferably, all antennas are configured to operate with millimeter-wave signals.

[0015] The present invention is also directed toward a road vehicle radar, the radar comprising: a beamforming network having a first number of input ports and a second number of beamforming ports, wherein the first and second numbers are greater than 1; a first number of antennas for receiving signals, each of which is connected to the respective input port of a beamforming lens; a second number of receivers, at least two of which are capable of operating simultaneously, each of which is connected to the respective beamforming port of the second number of beamforming ports; a third number of antennas for transmitting signals, wherein the third number is greater than 1; and processing means configured to determine the position of an object relative to the radar based on the amplitude and phase of the signals received by the receivers. The beamforming network is configured to combine signals received by at least two of the first number of antennas from at least two of the third number of antennas.

[0016] Furthermore, the present application is also directed to a method for determining the position of an object with respect to a radar for a vehicle traveling on a road. This method includes transmitting signals from a plurality of transmitting antennas, receiving signals with a plurality of receiving antennas, providing received signals from at least two receiving antennas to an input port of a beamforming network including a Lotman lens, wherein the beamforming network has a plurality of beamforming ports, the beamforming network is configured to combine the signals received at its input port and provide the combined signal to its beamforming ports, the beamforming network is further configured to apply additional phase terms to each of the signals to be combined so as to form a beam at a desired angle, providing the received signals, obtaining an output from the beamforming network with a plurality of receivers respectively connected to each of the beamforming ports of the beamforming network, and processing the obtained output to determine the position of the object at each receiver based on the amplitude and relative phase of the combined signal provided from the beamforming ports of the beamforming network to the receivers.

[0017] Preferably, the method includes configuring the receiving antennas and / or the transmitting antennas into a linear array consisting of at least four antennas, wherein the distance between at least the central antennas of the array is d, and the distance between the outermost antenna and each adjacent antenna is greater than d.

[0018] The term "and / or" used above represents an inclusive "or". That is, A and / or B means either A or B or both A and B.

[0019] By providing a predetermined linear array with unequal intervals between each antenna, the linear array can be configured so as not to affect the characteristics of the main beam, and the number of antennas used can be reduced.

[0020] Alternatively, the method includes configuring the receiving antennas to be arranged on a first line having a predetermined distance d between each receiving antenna, and configuring the transmitting antennas to be arranged on a second line having a distance of N×d between each transmitting antenna, where N is the number of receiving antennas.

[0021] Preferably, the method configures the receiving antennas on the first line and the transmitting antennas on the second line, and the first line is substantially parallel to the second line.

[0022] In an alternative embodiment, the method configures the receiving antennas on the first line and the transmitting antennas on the second line, and the first line is not parallel to the second line. In this embodiment, the first line is preferably arranged at an angle greater than 45° with respect to a parallel line of the second line. More preferably, the first line is orthogonal to the second line.

[0023] Preferably, the beamforming network is a Rotman lens. The Rotman lens forms a virtual array of the Rotman lens by using a plurality of transmitting antennas. The Rotman lens is preferably configured such that the individual beams formed by the virtual array of the Rotman lens overlap.

[0024] Optionally, the Rotman lens is configured such that the beam ports are concentrated around the radar boresight.

[0025] Preferably, the radar signals transmitted and received by the antennas and processed by the Rotman lens are millimeter-wave radar signals.

[0026] Furthermore, this application is also directed toward a method for determining the position of an object relative to a radar for road vehicles. This method includes transmitting signals from a plurality of transmitting antennas, receiving signals with a plurality of receiving antennas, and providing received signals from at least two receiving antennas to the input ports of a beamforming network, wherein the beamforming network has a plurality of beamforming ports and is configured to synthesize the signals received at its input ports and provide the synthesized signals to its beamforming ports; acquiring outputs from the beamforming network with a plurality of receivers, each connected to a beamforming port of the beamforming network, and processing the acquired outputs to determine the position of an object at each receiver based on the amplitude and relative phase of the synthesized signals provided to the receivers from the beamforming ports of the beamforming network. [Brief explanation of the drawing]

[0027] The present invention will be more clearly understood from the following description of its embodiments, which are given only as examples, with reference to the accompanying drawings. [Figure 1] Figure 1 shows prior art for a conventional 8-receiving element antenna configuration. [Figure 2] Figure 2 shows a comparison of a 4-element antenna beam pattern and an 8-element antenna beam pattern. [Figure 3] Figure 3 shows a prior art MIMO configuration that corresponds to the conventional 8-receiving element antenna configuration shown in Figure 1. [Figure 4] Figure 4 shows an example of TDM used on the transmitting side of an FMCW MIMO system where the number of transmitters is M>2. [Figure 5] Figure 5 shows prior art for MIMO arrays used in radar systems that perform 3D mapping. [Figure 6] Figure 6 shows the normalized received array radiation pattern scanned to 0 degrees for a radar having one transmitting element and four receiving elements. [Figure 7] Figure 7 shows the normalized virtual receiver array emission pattern scanned to 0 degrees for a MIMO radar having two transmitting elements and four receiving elements. [Figure 8] Figure 8 shows the chirp of the FMCW radar and the results of estimating the resulting distance and velocity using a two-dimensional FFT. [Figure 9] Figure 9 shows the result of performing an FFT on the distance and velocity domains for all receivers in order to calculate the angle in a conventional MIMO radar. [Figure 10] Figure 10 shows the configuration of the MIMSO according to this disclosure. [Figure 11] Figure 11 shows prior art for a typical Lotman lens configuration. [Figure 12] Figure 12 shows a comparison of the target detection output level against the vehicle's angular position for one output port of a Rotman lens design with 10 receiving antenna elements, comparing the case using one transmitting antenna with the case configured as a MIMSO radar system using two transmitting antennas. [Figure 13] Figure 13 shows the MIMSO angle processing configuration according to this disclosure. [Figure 14] Figure 14 shows an example of a Rotman lens with a reconfigurable beam, high angular accuracy towards the center of the radar, and lower accuracy towards the edges of the field of view. [Modes for carrying out the invention]

[0028] Next-generation in-vehicle radar used in automobiles, buses, trucks, and other vehicles is required to detect the angle of objects with a resolution of 1 degree. Preferably, the system mounted on or integrated with the vehicle must also completely cover the 360 ​​degrees around the vehicle. This is thought to be achieved by using multiple radar systems on a single vehicle, with each radar system covering a different angular range, and by combining the multiple radar systems to cover the entire 360 ​​degrees around the vehicle.

[0029] To resolve the angular position of an object, a technique known as "beamforming" can be used. Beamforming uses multiple antennas arranged in an array (in a transmitter (commonly abbreviated as TX) or receiver (commonly abbreviated as RX), or in both a transmitter and a receiver). By manipulating the coherence of the signal waves transmitted from the antennas using the array of antennas, the energy transmitted or received by each antenna is concentrated into a narrow beam, increasing the peak output of the beam and decreasing the output in all other directions. By adjusting the amplitude and phase of the signal at each antenna in the array, beam characteristics such as beam steering angle and beam width can be manipulated. Increasing the number of antennas in the array can narrow the resulting beam width. In the case of a transmitting array, increasing the number of antennas in the array increases the focusing of the transmitted signal and increases the gain of the array.

[0030] Beamforming radar can be broadly classified into two types: analog and digital.

[0031] In the first type of analog radar, the radar's angular performance is determined by one or more physical antenna beamforming structures. Analog radars also include radars in which beam manipulation is completed within these antenna beamforming structures. Antenna systems implemented in this way have been used in numerous applications, including base stations, military and space radar systems. In conventional electron beam steering antenna systems, which electronically move the direction of the main lobe of the antenna beam pattern, radio frequency (RF) phase shifters are commonly used to dynamically adjust the relative phase between antenna elements. For this reason, RF phase shifters are an important component in beam steering antenna systems for steering the narrow beam pattern of a phased array antenna.

[0032] Another analog approach is to use a lens supply structure as a beamforming component. In particular, a lens supply structure (e.g., a Rotman lens) has multiple supply ports configured to supply received signals to a beamforming component. When in use, 1) each supply port is operably connected to a switched port, and switching between the supply ports of the lens supply structure is used to connect one transceiver, transmitter, or receiver to the appropriate supply port for the desired beam direction. 2) Each supply port is operably connected to multiple transceivers, transmitters, or receivers, and the hardware connected to the supply port for the desired beam direction is activated, while the hardware connected to the other supply ports is deactivated. 3) Each supply port is operably connected to multiple transceivers, transmitters, or receivers, and the hardware connected to two or more supply ports is activated to enable reconstruction of the transmitted emission pattern or simultaneous observation of multiple received beam directions. Or, 4) each supply port is operably connected to a combination of a switched port and multiple transceivers, transmitters, or receivers. By scanning the transmitting and receiving beams, the angular position of the target object can be determined.

[0033] Analog radars are generally large (i.e., they contain many RF components, which increases the risk of component failure), but they can determine the angular position of an object with minimal computational processing.

[0034] Digital radar uses complex computational algorithms to digitally "beamform" the radar's transmitted and / or received signals. These radars are generally small and have fewer RF components, but require more complex computational processing, which can increase latency depending on the system's CPU processing power.

[0035] Therefore, the arrangement of the TX antenna (also called the "transmitting element" or "transmitting antenna element") and the RX antenna (also called the "receiving element" or "receiving antenna element") in a radar system can take many different forms. Both analog and digital radar systems have advantages and disadvantages, and they are not incompatible with each other. Nevertheless, radar architectures generally follow one of these two approaches.

[0036] Both analog and digital beamforming receivers utilize the fact that when a signal is incident on multiple antennas, there is a time delay corresponding to the angle of incidence before the signal reaches each element.

[0037] For example, Figure 1 shows an array of eight RX antennas with an angle θ i The wavefront is shown when incident on a given element, and the antennas are spaced d apart. In the case of an array of n antennas where the distance between each RX antenna is d, the wavefront will be (n-1)d × sin(θ) until it reaches the nth RX antenna in the array, compared to the first element the wavefront reaches. i This causes an extra movement of (n-1)φ at the nth RX antenna. i A phase lag occurs.

number

[0038] The angle at which the signal is emitted can be calculated by performing a mathematical operation known as the Discrete Fourier Transform (DFT) on the amplitude and phase outputs from the array, which can be calculated using any Fast Fourier Transform (FFT) algorithm known in the art. The resolution of prior art systems is typically improved by increasing the number of RX antennas in the receiving array.

[0039] When the signals received by each element of the array are combined, the amplitude of the resulting signal depends on the angle at which the signal arrives and the complex weighting (with amplitude and phase) of the signal received by each element. Mathematically, this focusing effect of the array can be expressed by multiplying the radiation pattern of one element in the array by the array factor.

[0040] The array factor is a complex-valued representation of the far-field radiation pattern of an array of isotropic radiators (i.e., theoretical antennas that are non-directional and radiate equal energy in all directions). The one-dimensional array factor for a linear array is calculated for each discrete angle θ as follows:

number

number

[0041] Figure 2 shows the array factor 215 of the 4-element array 210 and the array factor 225 of the 8-element array 220. Since each array factor in Figure 2 is calculated without adding an additional phase to each element, it reaches its maximum at θ s = 0°. By changing φ s , the peak of the array factor can be scanned across the FOV of the radar. The value of φ s used for array weighting is equal to the value of φ i which is the phase delay between array elements due to the angle of incidence of the received signal. When they are equal, the signals coherently combine to form a peak. The formula for the array factor is similar to the following DFT.

Equation

[0042] The Fourier transform is also used in a Frequency Modulated Continuous Wave (FMCW) radar system to estimate the distance and speed of a target and can be calculated using the FFT. The array factor can be converted to decibels using the following formula.

Equation

[0043] Calculating the array factor using the physical antenna positions in the physical array and multiplying it by the radiation pattern of one physical antenna in that array (i.e., a real antenna with its own composite radiation pattern), the resulting radiation pattern will be a good approximation of the radiation pattern of the physical array (without including effects such as mutual coupling between elements in the physical array that can change the pattern). Thus, in decibel form, the gain of the array is given by the following formula.

Equation

[0044] If beamforming consists of circuits arranged in a specific way so that calculations are performed by analog circuits between the RX antenna and the remaining signal processing elements of the receiver (e.g., "addition" or "subtraction" of RF signals by circuit elements such as hybrid couplers, addition of time delay by lenses such as Rottmann lenses or Luneberg lenses, or phase shifting by phase shifter circuit elements), then such arrangement falls under the aforementioned definition of an analog type radar. In the case of analog beam steering, the analog circuits perform calculations similar to DFT. In the case of conventional phased arrays, the scanning angle θ s To change this, a phase shifter is used to apply complex weighting to each element in the array w n Phase φ s Change it.

[0045] In digital beam steering, one or more transmitting antennas are used in combination with multiple receiving antennas. If such a combination circuit is not available, the signal from each antenna is converted from the analog domain to a digital signal before being input to the receiver's digital signal processing element. In this type of system, beamforming can be performed in the receiver using digital circuits, either to perform further mathematical operations such as addition and subtraction, or by introducing phase delays between elements, as previously described for analog systems. In the case of MIMO radar, which will be discussed in more detail later, one of several steps that can be performed is an "angle FFT" performed on the complex received signal from each receiver across the array to resolve the angular position of the target. The angle FFT is a means of calculating the DFT, and the formula for the array factor is similar. The angle FFT is performed across the region of interest of the MIMO radar and outputs a peak at the angle where the target is located. The peak of the angle FFT is the shape of the beam, or main lobe. By using more receiving elements, the beam width can be narrowed, increasing the resolution between closely spaced objects.

[0046] In conventional digital beamforming architectures, a dedicated receiver is incorporated for each physical antenna (i.e., M receivers for M x M RX antennas). When low angular accuracy is required, a wide beam is sufficient, which can be achieved using fewer antenna elements. Therefore, the number of receivers used can be minimized. However, when high angular accuracy is required, a narrower beam is needed. Consequently, additional antennas are required, as shown in the difference between antenna beam patterns 215 and 225 in Figure 2. As a result, the overall cost and size of the radar increases. Therefore, applications requiring high angular accuracy, such as automotive radar, have been expensive until now. When low cost is required, radars using conventional digital beamforming MIMO can only use a minimum number of elements, thus limiting them to applications with wide beams.

[0047] In systems that do not use digital beamforming, a similar general relationship holds true: higher angular accuracy requires more antenna elements. However, analog beamforming networks often use beam switching to reduce the number of receivers needed to process signals received from the RX antenna array. As a result, analog beamforming networks are considerably larger than digital ones, and their costs increase proportionally. A common compromise between pure analog and pure digital beamforming networks is to use a digitally controlled switching network with analog beamformers, where connections to different antenna elements are sequentially connected or switched to a single receiver. In such systems, while receiver costs can be reduced, the beamforming network and associated radar costs can be very large because phase balance between antenna elements must be maintained.

[0048] Beamforming and beam steering typically both use scaling (also known in this art as tapering or weighting) to shape the beam pattern and different antenna elements (w n The overall shape of the coupled beam is manipulated by amplifying or attenuating the signal intensity between the transmitter and receiver. When beam steering is performed on the transmitter side, the beam pattern is generally manipulated so that only one radiation pattern can exist at a time. However, on the receiver side, digital beamforming can form multiple beam patterns simultaneously and in parallel using different element scaling algorithms, resulting in multiple outputs.

[0049] In one known embodiment of digital radar, multiple transmitters and multiple receivers work together to form a “digital” type MIMO radar. There are two main types of MIMO radar. The first type is called statistical MIMO, in which antennas (including both TX and RX antennas) are placed far apart from each other to provide different fields of view of objects, or “scenes.” The second type of MIMO is called beamforming (or proximity placement) MIMO, in which antennas are placed close together and work together to form a “virtual” beamforming array. This disclosure is directed toward an improvement of this second type of MIMO radar.

[0050] In the case of E-band automotive radar systems, the operating frequency is preferably within a first bandwidth of 76-81 GHz. (However, for specific applications depending on the radar characteristics (e.g., long-range, medium-range, short-range, etc.), it is not necessary to use the full bandwidth of the frequency band.) This first bandwidth presents a significant challenge, as anyone familiar with the technology will know. For example, it is difficult to ensure that frequency-invariant beamforming does not degrade as a result of performance variations of individual components at different frequencies within the first bandwidth.

[0051] Many conventional systems achieve broadband beamforming by using a filter bank that operates either in the time domain using a two-dimensional filter bank, or in the frequency domain (after FFT), filtering each band individually. Essentially, the broadband signal is divided into a predetermined number of subbands, each individually narrowband filtered.

[0052] The system of this disclosure overcomes this challenge by replacing the aforementioned digital angular FFT calculation within the MIMO system using a Lotman lens, while simultaneously operating with minimal degradation across all usable frequencies. In this respect, a compromise is found between purely "digital" and purely "analog" type systems.

[0053] In conventional MIMO systems, the amplitude and phase of signals received by N antennas (multiple inputs) from each of the M transmitters (multiple outputs) in the system can be used to form a "virtual" receiving array that is larger than the physical receiving array (i.e., the virtual receiving array has more "virtual" RX antennas than the N antennas in the physical antenna array). This can improve the angular resolution of objects.

[0054] MIMO multiplexing schemes include various methods aimed at orthogonalizing these signals so that multiple signals transmitted from individual TX transmitters can be separated on the RX side. Examples of such schemes, though not limited to them, include Time Division Multiplexing (TDM), Frequency Division Multiplexing (FDM), and Binary Phase Multiplexing (BPM).

[0055] Referring to Figure 3 for MIMO in the TDM method, the transmitters transmit signals one by one sequentially (for example, TX1 transmits a signal, and TX2 transmits a signal after a predetermined delay). As a result, the receiving element receives the signals transmitted from each transmitter sequentially in time. This is shown in Figure 4 for an FMCW radar with a total of M transmitters where M>2. Therefore, the signals transmitted from each transmitter can be separated according to which transmitter the signal originated from (for example, the signal received from TX1 can be separated from TX2).

[0056] Other methods can be used to separate signals received by MIMO according to which transmitter they originated from. For example, in BPM MIMO, the phase of the transmitted signal is given a unique orthogonal code consisting of a sequence of weights of +1 (i.e., 0° phase shift) and -1 (i.e., 180° phase shift), and each receiver applies these codes to multiplex the signal, so that the signal is decomposed into separate signals depending on which transmitter each code corresponds to. In FDM MIMO, the transmitters operate in different frequency slots rather than time slots, and the transmitted signals are separated at the receiver using frequency shifting and filtering operations.

[0057] Because the signal from the TX antenna can be separated by the receiver, by appropriately setting the distance between the TX and RX antennas, the signals received by each RX antenna can be rearranged to correspond to the source transmitter. As a result, the phase difference of the received radiated waves will appear to have a considerably larger number of receiving elements than actually exist, creating a "virtual" antenna. For example, as shown in Figure 3, by separating four RX antennas by a distance d and the TX antenna by a distance 4d, a virtual array of 2 × 4 RX antennas is formed. More generally, in a MIMO system with M × (M) TX antennas and N × (N) RX antennas, it is possible to obtain a virtual array of M × N RX antennas using only M + N physical antennas. Therefore, the angular resolution of the virtual array is much finer than the angular resolution of the physical array when not used in the MIMO implementation. Thus, MIMO technology is effective in improving the angular resolution of radar with a given number of transmitters and receivers, or reducing the number of transmitters and receivers required for a given angular resolution.

[0058] Furthermore, referring to Figures 3 and 4, when using a MIMO system in the TDM (Time Deposition Monitoring) method, it is not possible to measure each virtual element in the MIMO system simultaneously due to the time delay between transmissions required to obtain TDM. In the example shown in Figure 3, to obtain signals for the virtual array of eight antennas, it is necessary to repeat the measurement of the RX antenna twice at each "real" RX antenna (i.e., once at each TX antenna). On the other hand, the system in Figure 1 (which has eight physical antennas) can obtain measurements from each RX antenna simultaneously. In terms of accuracy of direction and distance, both systems yield the same results, but the calculation is faster in the system in Figure 1.

[0059] In the MIMO configuration of the system shown in Figure 3, the radar can make decisions with good angular resolution only on one axis (using a linear array of multiple elements), but it cannot make any decisions on the other axis (using a single element).

[0060] For comparison, Figure 5 shows an alternative arrangement to cover 2D angles. Here, the transmitting elements are arranged in a first linear array 501 in one orientation, and the receiving elements are arranged in a second linear array 502 in a second orientation, the second orientation being at an angle with respect to the first linear array. The resulting virtual array 503 can determine the angular position of an object with respect to both the first and second orientations, and as a result, the absolute position of an object can be calculated. Preferably, the first and second linear arrays are oriented orthogonal to each other.

[0061] EP3497473A1 discloses that angular resolution can be further improved by using 3D-printed Luneberg lenses. The advantage of using lenses is that by changing the shape of the lens and its geometric arrangement relative to the antenna elements, multibeam radars with various FOVs and gain characteristics can be designed. However, Luneberg lenses are generally difficult to design and manufacture because they are three-dimensional structures and require materials with various refractive indices. Because the focal plane of a Luneberg lens is spherical, it is difficult to integrate it into a flat printed circuit board, requiring a complex design for connection to the radar motherboard.

[0062] In MIMO radar, the size of the radar enclosure is not significantly limited because the MIMO radar can synthesize a large-diameter virtual array using only a small number of transmitting and receiving antennas.

[0063] Figure 6 shows the normalized radiation pattern 620 centered at 0°, obtained using one TX antenna 605a and four RX antennas 610, and measured over their field of view (FOV). The resulting antenna beam pattern is the same as that of the four-element array shown in Figure 2.

[0064] However, as shown in Figure 7, by introducing a second TX antenna 705b and four identical RX antennas 710, an array of four virtual RX antennas 715 is added to the existing physical RX antenna 710. As a result, a virtual array of eight RX antennas is created, resulting in the eight-element beam pattern 720 shown in Figure 2. Therefore, the angular resolution of the antenna array doubles with the addition of the four virtual antennas 715, but at the cost of only one additional TX antenna 605b. Further examination of the antenna beam pattern in Figure 2 reveals that the angular accuracy doubles as the beam width is halved, thus demonstrating the effect on angular resolution.

[0065] Figure 8 shows a system according to the present disclosure. Preferably, the system is configured to use FMCW modulation. An FMCW waveform, also known as a chirp, is a complex sine wave whose frequency increases linearly with time. FMCW radar transmits the chirp at a period called the pulse repetition interval (PRI), and in most cases the frequency is described as a sawtooth wave, although other types of chirps also exist. The resulting echo of a target from the field will contain a delayed and attenuated copy of the transmitted chirp. The transmitted and received signals of an automotive radar are generally millimeter-wave signals, preferably in the range of 76–81 GHz. When the received signal is combined with the transmitted chirp, it becomes a beat signal. The beat signal is a complex sine wave with a frequency much lower than the transmitted and received frequencies. Generally, the beat signal is in the MHz frequency band for automotive radar. The frequency of the beat signal is proportional to the distance to the detected object. By collecting multiple chirps within a single "frame", it is possible to determine the Doppler frequency change in the "slow time" dimension.

[0066] Beat frequency estimation is typically performed in the digital domain after the beat signal has been digitally sampled. Since the beat frequency is much lower than the radar bandwidth, a slow analog-to-digital converter (ADC) can be used. Specifically, the beat signal is sampled, and each chirp sample is preferably placed in a separate column of a matrix. The row headers of the matrix correspond to the "fast" time taken over a single chirp, and the column headers correspond to the "slow" time taken over multiple chirps.

[0067] By applying an FFT to each column of the matrix shown in Figure 8, the beat frequency can be detected, thereby determining the distance of an object. Subsequently, by applying another FFT along the rows of this matrix, the Doppler frequency can be detected, thereby determining the velocity of the object. The use of these two FFTs is generally called a two-dimensional FFT, which enables tracking of objects in terms of both distance and velocity. Performing a two-dimensional FFT has the advantage of reducing noise levels through matched filtering of the object's beat frequency and Doppler frequency. The number of objects that fall within the same distance-velocity range is usually small, although this depends on the radar's distance and velocity resolution.

[0068] While distance-velocity plots provide much useful information, they lack detail regarding the angular position of an object. The angular position of an object is the angle between 1) a line connecting the object and the center of the physical RX antenna, and 2) a line running from the center of the physical RX antenna, perpendicular to the line of the physical RX antenna. If this angle is 0, the object is directly in front of the receiver. However, as shown in Figure 9, the angular position of an object can be obtained by taking instantaneous distance-velocity plots for each of the RX antennas (including both real and virtual RX antennas) and performing a third "angle" FFT, i.e., a 3D FFT.

[0069] For automotive safety radar systems, the current requirement for angular accuracy of less than 1° necessitates a uniform array of over 100 (virtual and real) antennas. Processing signals from such a large array is extremely costly, requiring a CPU with high floating-point operation performance (FLOPS). While it's possible to use a cheaper, slower CPU, this would result in longer detection times and consequently lower safety.

[0070] The above signal processing steps are typically performed after the Constant False Alarm Rate (CFAR) threshold setting step. The CFAR step filters out received signals below a predetermined threshold (the threshold is selected to be greater than the ambient noise) so that only signals with a signal-to-noise ratio exceeding this threshold are retained.

[0071] This disclosure eliminates the need to use the third angle FFT shown in Figure 9. Instead, a beamforming network is added between the physical RX antenna and each receiver, as shown in Figure 10. Preferably, the beamforming network includes lenses, more preferably Rotman lenses. The beamforming network 1030 is added between the physical RX antenna 1010 and each receiver (not shown). Although four RX antennas 1010 are shown in Figure 10, those skilled in the art will recognize that more or fewer antennas may be used. Adding an additional TX antenna 1005, as previously described with reference to Figure 7, results in a receiver array consisting of virtual elements 1015.

[0072] Preferably, the beamforming network includes a Rottman lens. Figure 11 shows the prior art geometric arrangement of a conventional Rottman lens. This geometric arrangement includes a plurality of input ports, i.e., beam ports, a lens cavity (cavity resonator) region, and a plurality of output ports, i.e., array ports. For illustrative purposes, the lens shown in Figure 11 has five beam ports and four array ports, but it should be understood that any number of ports can be used. The beam ports are located along the focal arc at one end of the parallel plate region of the lens, and the array ports are located at the opposite end. The array ports are connected to an array of antenna elements via phase correction lines. When each receiving array element is excited by the wavefront, the signal propagates through the phase correction lines to the parallel plate region before being sampled at the beam ports. The path lengths within the Rottman lens are designed so that the signal is coherently coupled at one or two beam ports corresponding to the relative angle at which the signal arrives from the object. b Conventional Rottman lenses, which have beam ports and are designed to have maximum focal angles F1 and F2 in ±α°, B The number of beams n b Fixed angle θ for the second beam port n It is formed by.

number

[0073] Thus, the Rotman lens acts as a circuit element that can be used to replace a digital FFT with an analog DFT for a finite number of fixed beam angles.

[0074] To apply MIMO technology to a Rotman lens, signals from separate transmitters are coherently combined. Preferably, this is performed following a 2D FFT step and a CFAR step, and additional phase compensation may be required to correct for phase errors due to the movement of the target during the time the chirp is received at a given lens beam port from separate transmitters. However, ignoring these steps for the sake of simplicity, one sequence of chirps from M transmitters is used to create n b The signals received at the second beamport are mathematically combined, and additional phase terms are applied to each to form beams at the correct angles.

number

number

[0075] By applying this technology, for example, Figure 12 plots the radar reflected wave from a car at an angular position of -6 degrees as the normalized received power. Here, the beam at this angle is formed from the output of one of the Rotman lens designs with 10 receiving antenna elements. As can be clearly seen from Figure 12, there is a peak in the received power at -6 degrees.

[0076] Therefore, although this system uses MIMO detection, it can be classified as a Multiple Input Multiple Steered Output (MIMSO) system because the output is directionally controlled. Furthermore, since the Rottman lens is designed with time delay rather than phase delay, this configuration allows for a simple way to ensure broadband operation.

[0077] Next, with reference to Figure 10, the operation of the MIMSO system will be described. As previously mentioned in the MIMO system, multiple transmitters and multiple receivers are used, with each physical receiver located at one of the Rottman lens output (beam) ports. As a result, the virtual Rottman lens is also connected to the virtual array, thus implementing the virtual array in the MIMSO system.

[0078] Preferably, a modified Rotman lens design is used instead of the design used in typical beam steering applications. In conventional Rotman lenses with a uniform linear array, the beam port of the Rotman lens is typically designed to detect the angle in proportion to the half-power beam width (HPBW), in other words, the beam is designed to overlap at its -3dB point (relative to the beam peak). However, in the case of the Rotman lens according to this disclosure, the HPBW of the lens beam is designed to increase by a factor of about M relative to the desired HPBW of the MIMSO system, where M is the number of transmitters in the MIMSO system. As a result, the HPBW of the beam is favorably narrowed by a factor of about 1 / M. As will be apparent to those skilled in the art, the example of two TX antennas is not limiting, and the Rotman lens beam port may be designed to appropriately detect angles from M × (M) TX antennas.

[0079] As mentioned above, certain types of electrical circuits can behave as if they were performing a mathematical function on the received signal. The simplest example of this is a low-pass filter, which can be considered a mathematical integrator because it allows DC and low-frequency signals to pass through while rejecting higher-frequency signals. In the MIMSO system, the Rottman lens behaves similarly, in this case performing the function of an angular FFT.

[0080] Figure 13 shows the completion of MIMSO processing in a system using two transmitters. Similar to the MIMO system shown in Figure 8, distance and velocity are calculated using a two-dimensional FFT. However, in the MIMSO system according to this disclosure, the angular position of an object is determined by using the beamports of a beamforming network, such as a Rottman lens. Since the detection signals received at each beamport correspond to different detection angles, there is no need to perform a further three-dimensional FFT to determine the detection angle. As described above, the Rottman lens is optionally modified by narrowing the HPBW of the beam by a coefficient of approximately 1 / M, where M is the number of transmitting elements.

[0081] Furthermore, the Rottman lens can be optionally modified to generate beams at non-uniform angles. Figure 14 shows one embodiment in which the beam density directed towards the center of the radar FOV (where angular accuracy is most critical) is greater than the beam density directed towards the edges of the radar FOV (where angular accuracy is less critical).

[0082] The size of the virtual array is the number of transmitting elements multiplied by the number of receiving elements. Furthermore, by concentrating the beam toward the center of the radar field of view, it is possible to use signals from all transmitters for the narrow inner beam and discard signals from one or more of the transmitters for the outer beam, resulting in a smaller virtual array and a wider beam pattern.

[0083] Beamforming networks such as Rottman lenses, while slightly larger in size compared to MIMO radar systems, eliminate the need for a third FFT calculation. Instead, the third FFT is performed instantaneously at the beamport of the beamforming network. By eliminating the processing requirements and time required to determine the angular position of an object, the MIMSO systems disclosed herein represent a significant improvement over existing radars, such as those used in automotive applications. The faster detection achieved by MIMSO systems results in quicker reaction times, which in turn leads to safer outcomes in traffic accidents.

[0084] Rotman lenses are particularly well-suited for use in beamforming networks. Because Rotman lenses are planar, they can be mounted on a single PCB along with other radar components. This offers further cost-effective advantages compared to incorporating larger lenses, especially Luneberg lenses, into radar systems.

[0085] Furthermore, the MIMSO system described herein can also be modified to apply phase-comparison monopulse technology. The phase-comparison monopulse output is calculated based on results from a real RX antenna and an equivalent virtual RX beamport. The primary direction of an object is represented by the beamport where the detected object was found, corresponding to the angle of the Lotman lens beamport, and the enhanced direction is represented by the sum of the beamport direction and the resulting phase-comparison monopulse calculation. This composite MIMSO monopulse processing may improve the determination of angular position.

[0086] Alternatively, the phase-compared monopulse output can also be calculated based on the results between each pair of consecutive beamports of the Rotman lens (i.e., one beamport and the next consecutive beamport, as shown in Figure 11). The principal direction of the object is represented by the midpoint between the beamports where the detected object was found, and the enhanced direction is represented by the sum of this midpoint beamport direction and the resulting phase-compared monopulse calculation. It is also possible to use these two methods in combination.

[0087] Conventionally, the beam angle of a Rotman lens is fixed at the time of design, and it is not possible to continuously scan within the field of view like conventional phased array or MIMO systems. However, in the MIMSO system according to this disclosure using a Rotman lens, when combining received signals from separate transmitters, the φ applied m By adjusting an additional phase term, the beam angle of the combined MIMSO beam can be slightly altered. Furthermore, by combining beams formed by adjacent ports (with the appropriate phase), the resulting signal corresponds to a beam located at the average angle of those two individual ports, similar to the sum beam in a monopulse system. Combining these two techniques with lens design can make it possible to continuously scan the lens digitally across its field of view (FOV).

[0088] Furthermore, by using three or more transmitters in a MIMSO system, amplitude taper can be applied to the synthesized MIMSO signal. This advantageously enables digital reduction of the beam pattern's sidelobe levels, reducing clutter and decreasing the likelihood of false target detection. Alternatively, the sidelobe level from one beam can be reduced by subtracting the signals received by one or more other beams with a conversion factor applied. As a further alternative, these two sidelobe reduction techniques can be used in combination.

[0089] The MIMSO systems disclosed herein may consist of discrete elements, but are preferably fabricated on a printed circuit board. More preferably, the system is configured to determine the position and trajectory of an object with improved accuracy and speed using millimeter wavelength electromagnetic radiation in the E-band frequencies specified for automotive radar (i.e., electromagnetic radiation having frequencies in the range of 76 to 81 GHz).

[0090] Furthermore, the above-mentioned MIMSO system can be used in combination with one or more of the following: LiDAR and stereo cameras. Additionally, machine learning can be used to enhance the system's performance.

[0091] The above examples are provided for illustrative purposes only and are not intended to limit the scope of this disclosure. Rather, the scope of this disclosure is defined in the attached claims.

Claims

1. This is a radar for vehicles on the road. A beamforming network comprising a Rotman lens having a first number of input ports and a second number of beamforming ports, wherein the first and second numbers are greater than 1, A first number of antennas for receiving signals, each of the first number of antennas being connected to the respective input ports of the beamforming network, A second number of receivers, each of which at least two receivers are capable of operating simultaneously, wherein each receiver of the second number of receivers is connected to each beamforming port of the second number of beamforming ports, A third number of antennas for transmitting a signal, wherein the third number is greater than 1, A processing means configured to determine the position of an object relative to the radar based on the amplitude and phase of the signal received by the receiver, Equipped with, The beamforming network is configured to combine coherent signals received by at least two of the first number of antennas from at least two of the third number of antennas. The processing means is further configured to mathematically apply an additional phase term to each signal being synthesized in order to form a beam at a desired angle. The Rottman lenses are configured such that the individual beams formed by the Rottman lenses of the virtual array overlap, through the use of the third number of antennas. The beamwidth at half maximum (HPBW) of the Rottman lens is increased by M times relative to the desired HPBW, where M is the number of transmitters in the system, for a radar.

2. The first number of antennas are arranged on the first line, The third number of antennas are arranged on the second line, The first line is substantially parallel to the second line. The radar according to claim 1.

3. The first number of antennas are arranged in a line with a predetermined distance d between each antenna. The third number of antennas are arranged in a line with a distance of N × d between each antenna, where N is the first number. The radar according to claim 1.

4. The radar according to claim 1, wherein the individual beams formed by the virtual array of Rottman lenses overlap at an angle such that the gain is half the peak value of each beam.

5. The radar according to any one of claims 1 to 4, wherein the Rotman lens is configured such that the beam port is concentrated around the boresight of the radar.

6. The radar according to any one of claims 1 to 5, wherein the antenna is configured to operate with millimeter-wave radar signals.

7. A method for determining the position of an object relative to a radar for road vehicles, Transmitting signals from multiple transmitting antennas, Receiving the signal with multiple receiving antennas, The beamforming network, which includes a Rottman lens, provides the received signals from at least two receiving antennas to an input port, wherein the beamforming network has a plurality of beamforming ports, and the beamforming network is configured to coherently combine the signals received at its input ports and provide the combined signal to its beamforming ports. This involves processing each signal being synthesized to mathematically apply an additional phase term so that it forms a beam at a desired angle, Multiple receivers connected to each beamforming port of the beamforming network acquire the output from the beamforming network, The acquired output is processed, and the position of the object is determined at each receiver based on the amplitude and relative phase of the composite signal provided to the receiver from the beamforming port of the beamforming network. The Rottman lens is configured such that a virtual array of Rottman lenses is formed by using a third number of antennas, and the Rottman lens is configured such that the individual beams formed by the virtual array of Rottman lenses overlap, and the half-power beamwidth (HPBW) of the Rottman lens is increased by M times the desired HPBW, where M is the number of transmitters in the system. Methods that include...

8. The receiving antenna is configured as a linear array consisting of at least four antennas, wherein the distance between at least the central antennas of the array is d, and the distance between the outermost antenna and each adjacent antenna is greater than d, and the receiving antenna is configured with predetermined unequal intervals between each receiving antenna so as to reduce the number of receiving antennas used without affecting the characteristics of the main beam. The transmitting antenna is configured such that the transmitting antennas are arranged at predetermined unequal intervals between them, without affecting the characteristics of the main beam and reducing the number of transmitting antennas used. The method according to claim 7, including the method described in claim 7.

9. The receiving antennas are configured to be arranged on a first line having a predetermined distance d between each receiving antenna, When the number of receiving antennas is N, the transmitting antennas are configured to be positioned on a second line having a distance of N × d between each transmitting antenna. The method according to claim 7, including the method described in claim 7.

10. The method according to claim 8 or 9, comprising configuring the receiving antenna on a first line and the transmitting antenna on a second line, wherein the first line is not parallel to the second line.

11. The method according to any one of claims 7 to 10, wherein the Rottman lens is configured such that the beam port is concentrated around the boresight of the radar.

12. The method according to any one of claims 7 to 11, wherein the first signal is a millimeter-wave radar signal.

Images