阅读说明：本技术 具有方位和高度之高角分辨率的紧凑型雷达切换/mimo阵列天线 (Compact radar switching/MIMO array antenna with high angular resolution of azimuth and elevation ) 是由 诺姆·阿金 阿莫斯·巴伦 约拉姆·什切青 于 2018-01-30 设计创作，主要内容包括：本发明公开了一种使用较少发射(Tx)和接收(Rx)阵列元件来增大雷达切换/MIMO天线阵列的有效孔径的方法,根据该方法,雷达物理接收(Rx)/发射(Tx)元件阵列被布置在至少两个相对的Rx行和至少两个相对的Tx列中,使得每行包括彼此均匀间隔的多个接收(Rx)元件,且每列包括彼此均匀间隔的多个发射(Tx)元件；该阵列形成矩形物理孔径。当用作切换阵列时,来自一个列的第一Tx元件被激活,以在预定时隙期间发射雷达脉冲；在所有Rx元件中接收第一发射的反射,从而围绕由矩形物理孔径内第一Tx元件的位置所确定的原点,虚拟地复制两个相对的Rx行。在不同时隙期间,对所有剩余的Tx元件重复该过程,从而围绕由矩形物理孔径内每个被激活Tx元件的位置所确定的原点,虚拟地复制两个相对的Rx行,同时每次在所有Rx元件中接收来自于每个Tx元件的发射的反射。这样,用复制的两个相对的Rx行,铺设具有两倍于矩形物理孔径尺寸的矩形虚拟孔径。该虚拟孔径确定雷达波束的宽度和旁瓣。(A method of increasing the effective aperture of a radar switching/MIMO antenna array using fewer transmit (Tx) and receive (Rx) array elements is disclosed, according to which an array of radar physical receive (Rx)/transmit (Tx) elements are arranged in at least two opposing Rx rows and at least two opposing Tx columns, such that each row comprises a plurality of receive (Rx) elements evenly spaced from each other and each column comprises a plurality of transmit (Tx) elements evenly spaced from each other; the array forms a rectangular physical aperture. When used as a switching array, a first Tx element from one column is activated to transmit a radar pulse during a predetermined time slot; the reflection of the first transmission is received in all Rx elements, virtually duplicating two opposite Rx rows around an origin determined by the position of the first Tx element within the rectangular physical aperture. This process is repeated for all remaining Tx elements during different time slots, virtually duplicating two opposite Rx rows around the origin determined by the position of each activated Tx element within the rectangular physical aperture, while receiving reflections of the transmissions from each Tx element in all Rx elements at a time. Thus, with two opposite Rx rows duplicated, a rectangular virtual aperture with twice the size of the rectangular physical aperture is laid out. The virtual aperture determines the width and side lobes of the radar beam.)

1. A method of increasing the effective aperture of a radar switching/MIMO antenna array using a small number of transmit (Tx) and receive (Rx) elements, comprising:

a) providing an array of radar physical receive (Rx)/transmit (Tx) elements arranged in at least two opposing Rx rows and at least two opposing (Tx) columns, and such that each row comprises a plurality of receive (Rx) elements evenly spaced from each other and each column comprises a plurality of transmit (Tx) elements evenly spaced from each other; the array forms a rectangular physical aperture;

b) activating a first Tx element in one column to transmit a radar pulse during a predetermined time slot and receive a reflection of a first transmission in all Rx elements to virtually duplicate the two opposing Rx rows around an origin determined by the location of the first Tx element within the rectangular physical aperture; and

c) repeating the previous step for all remaining Tx elements during different time slots, virtually duplicating the two opposing Rx rows around an origin point determined by the position of each activated Tx element within the rectangular physical aperture, wherein reflections from the transmission of each Tx element are received in all Rx elements at a time, thereby laying down a rectangular virtual aperture having twice the size of the rectangular physical aperture with the duplicated two opposing Rx rows.

2. The method of claim 1, wherein the spacing between Tx columns is λ/2.

3. The method of claim 1, wherein the spacing between Rx elements is λ/2 times the number of Tx columns per side.

4. The method of claim 1 wherein the Tx array elements use different pulse waveforms that are separable in the receiver, transmitted simultaneously rather than sequentially.

5. The method of claim 1, wherein the transmit array elements are activated at different times or simultaneously using orthogonal waveforms.

6. The method of claim 1, wherein data is collected simultaneously from groups of receive array elements.

7. A radar sensor, comprising:

a) a radar physical receive (Rx)/transmit (Tx) element array disposed in at least two opposing Rx rows and at least two opposing Tx columns such that each row includes a plurality of receive (Rx) elements evenly spaced from each other and each column includes a plurality of transmit (Tx) elements evenly spaced from each other; the array forms a rectangular physical aperture;

b) a processor adapted to:

b.1) activating a first Tx element in a column to transmit a radar pulse during a predetermined time slot and to receive a reflection of a first transmission in all Rx elements, virtually replicating the two opposing Rx rows around an origin determined by the position of the first Tx element within the rectangular physical aperture; and

b.2) repeating the previous step for all remaining Tx elements during different time slots, virtually copying the two opposite Rx rows around an origin determined by the position of each activated Tx element within the rectangular physical aperture, wherein reflections from each Tx element's transmission are received in all Rx elements at a time, laying down a rectangular virtual aperture with twice the size of the rectangular physical aperture with the copied two opposite Rx rows.

8. The array of claim 7, wherein the radar frequency is at least 77 GHz.

9. The array of claim 7, comprising two parallel rows and two parallel columns, the rows being "full" in the azimuth direction and the columns being "sparse" in the azimuth direction.

10. The array of claim 7, comprising parallel rows and parallel columns, the rows being azimuthally "full" and the columns being azimuthally "sparse".

11. The array of claim 7, comprising parallel rows and parallel columns, the rows being "sparse" in height and the columns being "full" in height.

12. The array of claim 7, wherein all Tx elements are activated in such a way that one Tx element is activated per slot, thereby increasing the effective size of the aperture up and to the right.

Technical Field

The invention relates to the field of imaging radars. More particularly, the present invention relates to a compact radar switching array antenna with high angular resolution of azimuth and elevation, and with an increased effective aperture.

Background

In recent years, many industries are turning to autonomous solutions (autonomous solutions), such as the automotive industry, transportation, etc. These autonomous platforms should operate in an environment that interacts with both fixed and moving objects. For this purpose, these systems require a sensor suite in order to be able to reliably and efficiently sense their surroundings. For example, in order for an autonomous car to plan its route on a road with other vehicles, a trajectory planner (trajectory planner) must have a 3D map of the environment with the identity of the moving objects. Visual sensors are also degraded by inclement weather and poor visibility (e.g., fog, smoke, sand, storms, etc.), which can also limit the assessment of radial velocity. Light detection and ranging devices (lidar, for measuring the distance to an object by illuminating it with a laser) are expensive, most with moving parts and very limited range. Radar is an enhanced (non-alternative) technology.

Due to the natural limitations of visual sensors of optical (laser) technology in distance accuracy and reliability issues, the best solution to generate the above-mentioned 3D maps is through radar technology. It presents a series of new requirements that current radars do not meet.

In general, a larger aperture of the receive antenna allows more radiation to be received and provides greater sensitivity, or equivalently, a narrower main lobe. Thus, the receive antenna may receive weaker signals and provide a relatively accurate indication of the direction of the signal.

On the other hand, vehicle-mounted radars (including automotive imaging radars) require only a low sensitivity (because the distance is relatively short and the signal reflected from the target is relatively strong). However, vehicle-mounted radars do not require detection of point targets (e.g., aircraft or missiles), but require high accuracy in order to provide an image of the environmental information, which is used as an input to a simultaneous localization and mapping (SLAM) algorithm. The simultaneous localization and mapping (SLAM) algorithm should detect the location of obstacles, such as other nearby cars or pedestrians. A high precision narrow lobe will provide a clearer contour of the target image. The lobe width is determined only by the equivalent aperture and normalized to the wavelength of the radar transmitted signal (rather than by the number of receive antenna elements within the aperture which would affect the sensitivity, i.e., the ability to detect weakly reflected signals and the ambiguity resolution and side lobe levels).

Another key performance parameter of imaging radars is the side lobe level(s) of the antenna array. In the case of a large object (e.g., a wall) in the side lobe direction, the attenuated form of the reflection from the object appears to appear in the direction of the main lobe and may mask reflections originating from obstacles (e.g., pedestrians), or create phantom obstacles that may cause the vehicle to stop. Therefore, in automotive imaging radars, it is important to reduce side lobes as much as possible.

It is therefore an object of the present invention to provide a compact radar switching array antenna with high angular accuracy (resolution) of azimuth and height and with an increased effective aperture while using fewer transmit (Tx) and receive (Rx) elements to meet cost, space, power and reliability requirements.

It is a further object of the invention to provide a compact radar antenna array having high angular accuracy (resolution) of azimuth and elevation, and having an increased effective aperture, while reducing unwanted side lobes.

Other objects and advantages of the present invention will become apparent as the description proceeds.

Disclosure of Invention

The present invention relates to a method of increasing the effective aperture of a radar switching/MIMO antenna array using a smaller number of transmit (Tx) and receive (Rx) array elements, according to which an array of radar physical receive (Rx)/transmit (Tx) elements is arranged in at least two opposing Rx rows and at least two opposing Tx columns such that each row comprises a plurality of receive (Rx) elements evenly spaced from each other and each column comprises a plurality of transmit (Tx) elements evenly spaced from each other; the array forms a rectangular physical aperture. When used as a switching array, a first Tx element from one column is activated to transmit a radar pulse during a predetermined time slot; the reflection of the first transmission is received by all Rx elements, virtually copying (virtual replication) the two opposite Rx rows around the origin determined by the position of the first Tx element within the rectangular physical aperture. This process is repeated for all remaining Tx elements during different time slots, virtually duplicating two opposing Rx rows around the origin determined by the position of each activated Tx element within the rectangular physical aperture, while receiving reflections of the transmissions from each Tx element in all Rx elements at a time. Thus, a rectangular virtual aperture having twice the size of the rectangular physical aperture is tiled with two opposing Rx rows that are duplicated. The virtual aperture determines the width and side lobes of the radar beam.

The above-described replication method works equally well in MIMO or hybrid switched/MIMO designs, where some signals are transmitted simultaneously by multiple Tx array elements using orthogonal waveforms that are then separated in the receiver.

There are 2 sets of Tx columns, one on each side of the physical aperture. Each group may contain one or more columns nominally spaced at a distance of lambda/2.

The spacing between the leftmost column of the left group and the leftmost column of the right group is exactly the pitch between the Rx elements multiplied by the number of Rx elements.

Likewise, the spacing between the bottom and top rows is exactly the spacing between the Tx elements times the number of Tx elements.

These two design constraints are crucial to maintaining low sidelobes.

In one aspect, orthogonal waveforms may be used to activate transmit array elements at different times or simultaneously.

Data may then be collected simultaneously from the groups of receiving array elements.

The invention also relates to a radar sensor comprising:

a) a radar physical receive (Rx)/transmit (Tx) element array disposed in at least two opposing Rx rows and at least two opposing Tx columns such that each row includes a plurality of receive (Rx) elements evenly spaced from each other and each column includes a plurality of transmit (Tx) elements evenly spaced from each other; the array forming a rectangular physical aperture;

b) a processor adapted to:

b.1) activating a first Tx element in one column to transmit a radar pulse during a predetermined time slot and receive a reflection of the first transmission in all Rx elements, virtually duplicating two opposing Rx rows around an origin determined by the position of the first Tx element within the rectangular physical aperture; and

b.2) repeating the previous step for all remaining Tx elements during different time slots, virtually duplicating two opposite Rx rows around an origin determined by the position of each activated Tx element within the rectangular physical aperture, wherein reflections from each Tx element transmission are received in all Rx elements at a time, thus laying down a rectangular virtual aperture with twice the size of the rectangular physical aperture with the duplicated two opposite Rx rows.

The radar frequency may be at least 77 GHz.

The array may include parallel rows that are azimuthally "full" and parallel columns that are azimuthally "thin".

The array may also include parallel rows that are "sparse" in height and parallel columns that are "full" in height.

Activating all Tx elements in such a way that one Tx element is activated per slot increases the effective size of the aperture in both the vertical (e.g., up) and horizontal (e.g., right) directions.

Drawings

In the figure:

FIGS. 1a-1e illustrate a sparse/full frame array design for high azimuthal and elevation angular coordinate resolution in accordance with the present invention; and

fig. 2 shows typical outputs of the DSP from the top down view (a) and from the camera viewpoint.

Description of The Preferred Embodiment

The present invention provides a high resolution compact radar switch (switch) array antenna design with high azimuth and elevation angular accuracy, increased effective aperture, and reduced unwanted side lobes, using fewer transmit (Tx) and receive (Rx) elements.

In order to obtain high resolution, phased arrays have been designed based on the MIMO-SAR sparse/full (thin/full) method, using Switched or non-Switched Antenna Arrays (SAA). In this type of radar, there are some transmit array elements and a set of receive array elements, and orthogonal waveforms are used to activate the transmit array elements at different times or simultaneously and to collect data from the set of receive array elements simultaneously. To achieve high resolution, the radar should have a large aperture relative to the carrier signal wavelength (0.4 cm for 79GHz λ). Furthermore, for full 3D sensing, a large aperture is required in both azimuth and elevation. To achieve maximum aperture in both directions while minimizing the number of elements and the physical size of the array, a special framework design is proposed that uses a tau x-Rx duality in a sparse/full array design. According to this design, the Rx arrays are azimuthally full and azimuthally sparse, and the Tx arrays are azimuthally full and azimuthally sparse.

Furthermore, the resulting equivalent array aperture (which determines the beamwidth and sidelobe attenuation) is approximately twice the physical size of the array. With appropriate signal processing, the resulting aperture is twice the size of the actual frame.

Conventional solutions increase the number of receive (Rx) and transmit (Tx) elements within the array physical aperture and lack the multiplication effect described above. For example, a typical array may include columns of 12Tx elements and rows of 144Rx elements, resulting in an array of 144 x 12 — 1728 elements. This significantly increases the cost and reduces the reliability of such an array.

On the other hand, as will be shown later, the novel design proposed by the present invention enables the same accuracy and side lobe attenuation with only 48 Rx elements and 36Tx elements, which results in a total of only 84 elements in the array (saving about 95%).

The equivalent aperture is normalized to the wavelength λ at which the radar pulse is transmitted. Therefore, in order to obtain high resolution, it is preferable to emit at high frequency. In this example, the radar frequency is about 79 GHz.

Although the proposed design requires the transmission of 36 subsequent or orthogonal radar pulses to obtain a single frame, the savings in the number of components is significant.

FIG. 1a shows a first layout of a high resolution physical array with a reduced number of elements, according to an embodiment of the invention. The first array proposed by the present invention has two parallel rows 101 and 102 (in the azimuth direction) each having 72 receive (Rx) elements (their nominal pitch is about 0.5 λ) and two parallel columns 110 and 111 each having 6 transmit (Tx) elements (their pitch is exactly 72 times the Rx element pitch-36 λ apart from each other). It can be seen that this is a "sparse"/"full" rectangular arrangement (consisting of a full transmit array and a sparse receive array, or vice versa); where the rows are azimuthally "full" (i.e., containing a plurality of elements Rx having a pitch of about 0.5 λ) and the columns are azimuthally "sparse" (i.e., containing a plurality of elements Tx having a pitch 72 times the "full" pitch). For height, the roles are reversed, columns are "full" and rows are "sparse".

In the particular case shown in FIG. 1a, the "sparse" dimension for both azimuth and elevation is 2, and the equivalent array dimension is about twice the physical dimension.

The size of the physical aperture is AxB, which is defined by the rectangular dimensions, regardless of the total number of elements. In the switching mode, the basic concept is to transmit from only one Tx element in one time slot and receive reflections in all 48 Rx elements. In this example, transmission begins with the transmit elements 100 in the column 110 at a lower left position. The Rx elements in rows 101 and 102 receive the reflected signal. It should be noted that the order of the transmitting elements in the switching mode may be arbitrary.

Figure 1b shows how the effective size of the aperture can be increased upwards without adding more elements. The next transmission is activated from the Tx element 103, the Tx element 103 being one position above the lower left position in the column 110. This is equivalent to virtually copying the physical rows 101 and 102 to one position above, as virtual rows 101a and 101 b. The Rx elements in rows 101a and 102a receive the reflected signal. This process is repeated for all Tx elements in column 110, while adding additional copies of physical rows 101 and 102 (which receive the reselected signal) each time until the area above rows 101 and 102 is tiled by the copied virtual rows. This increases the effective vertical dimension of the aperture rectangle to 2B (rather than the physical vertical dimension B).

Figure 1c shows how the effective size of the aperture can be increased to the right without adding more elements. The next transmission is activated from the Tx element 104 in column 111 at the lower right position, which is equivalent to virtually copying the complete positions of the physical rows 101 and 102 to the right as virtual rows 101b and 102 b. The Rx elements in the virtual copy rows 101b and 102b receive the reflected signal.

Fig. 1d shows how the effective size of the aperture is increased to the right without adding more elements. The next transmission is activated from Tx element 105, Tx element 105 being one position above the lower right position in column 111. This is equivalent to virtually copying the complete positions of the physical rows 101 and 102 to one position on the right and above to become virtual rows 101c and 102 c. The Rx elements in rows 101c and 102c receive the reflected signal. This process is repeated for all Tx elements in column 111, while adding additional copies of physical rows 101 and 102 (which receive the reselected signal) each time until the area above virtual copy row 101b and row 102b is tiled by the copied virtual rows. This increases the effective horizontal dimension of the aperture rectangle to 2A (rather than the physical horizontal dimension a).

Fig. 1e shows how the effective size of the aperture is increased up and to the right without adding more elements, which is obtained by activating all Tx elements in such a way that one Tx element is activated per slot. The resulting equivalent aperture size is 2Ax2B, which is twice the aperture defined by the physical element (which is AxB). This multiplication of the physical aperture also occurs for antennas that transmit simultaneously using orthogonal waveforms.

In a second variation of the above array, each of the 2 columns is replicated (N-1) times by a series of shifts of λ/2 each (e.g., to the right), thereby generating two triplet columns. The total number of elements in all 6 columns is now 6N. For N-3, the total number of Tx elements is 36.

Each of the two 72 element rows now decimates by 3:1, so there are now 24 elements per row, the nominal spacing of the elements is 3/2 λ, and the total number of elements in the two rows is 48.

Using only 36TX and 48 RX array elements, whether transmitted sequentially or simultaneously from all 36 elements, yields an equivalent array of 36x 48-1728 elements.

The two alternative designs above illustrate how to transform the number of Tx and Rx elements, for example, from 12Tx and 144Rx to 36Tx and 49 Rx. This is of great importance when applying the design to the available radar transceiver chips to save on the number of chips. The radar front end is based on an ST Microelectronics radar transceiver chip STRADA770, which has 4Rx channels and 3 Tx channels; several such chips can be optionally connected to create a larger array. In the proposed design, 12 such chips are used, which provide 48 Rx channels and 36Tx channels and allow to obtain a resolution of about 0.9 ° in azimuth and a resolution of about 2.4 ° in elevation (the @ boresight before the window); the FOV is 90 in azimuth, 22.5 in elevation, and the array size is about 13x10 cm.

Without detracting from the above, it should be noted that "antenna" in this submission is sometimes used as a synonym for an array element. Each array element may itself be a sub-array of antennas. For example, each of the Tx or Rx array elements may itself be a vertical sub-array of antennas.

Signal processing

The processing unit is based on Xilinx Zynq-7000 series combined microprocessor FPGA SoC. Which controls the radar front-end and simultaneously processes the IF data. In each radar frame, we transmit 100 μ s chirps of up to 1GHz bandwidth from each of the 36Tx channels, with some chirps increasing in frequency (up-chirps) and some chirps decreasing in frequency (down-chirps). The processing unit acquires the sampled IF data over the parallel bus using an integrated FIFO buffer. A512-point FFT is then calculated for each Rx-Tx pair. Then we use another 3D FFT on the results (corresponding to the velocity, azimuth and altitude dimensions), but in a hierarchical fashion so that only regions where objects with reasonable probability exist will be processed in the next step. This approach produces sparse outputs from each FFT step to the next, significantly reducing the amount of processing on average. Finally, a list of 4-D voxels (distance, velocity, orientation, height) is generated and indicates the probability of the object appearing in that voxel. The probability is calculated from a noise distribution (null hypothesis) calculated in a negative spectrum always consisting only of noise components and not of real objects.

Although some embodiments of the invention have been described by way of illustration, it will be apparent that the invention may be carried out with many modifications, variations and adaptations, and with the use of numerous equivalents or alternative solutions within the scope of the claims, without departing from such scope.