JP2021516763A - Methods and equipment for object detection using beam steering radar and convolutional neural network systems - Google Patents

Abstract

The examples disclosed herein relate to radar systems in autonomous vehicles for object detection and classification. The radar system detects objects in the path and surrounding environment of an autonomous vehicle with a radar module having at least one beam steering antenna, a transmitter / receiver, and a scan parameter controller adapted to adjust the scan parameter set of the transmitter / receiver. It also has a machine learning module for classification and a recognition module with a classifier, which transmits object data and radar control information to the radar module.

Description

Cross-reference of related applications

This application claims priority to U.S. Provisional Patent Application No. 62 / 626,569 filed on February 5, 2018, and U.S. Provisional Patent Application No. 62, filed on January 4, 2018. It relates to US Patent Application No. 16 / 240,666 filed on January 4, 2019, claiming the priority of 62 / 613,675 (both incorporated herein by reference in their entirety). ).

Autonomous driving is rapidly shifting from the realm of science fiction novels to achievable reality. Advanced-Driver Assistance Systems (“ADAS”) that automate, adapt and enhance vehicles for safety and better driving are already on the market. The next step is driving functions such as maneuvering, acceleration, braking, and monitoring the surrounding environment and driving conditions, such as changing lanes or speeds when necessary to avoid traffic, crossing pedestrians, animals, etc. Become an increasingly vehicle that takes control of. Object and image detection requirements are important and specify the time required to capture the data, process it, and turn it into an action. During this time, ensure accuracy, consistency, and cost optimization.

A way to make this work is the ability to detect and classify objects in the surrounding environment at the same or even better levels as humans. Humans are proficient in recognizing and recognizing the world around them using a very complex human visual system that essentially has two main functional parts: the eye and the brain. In autonomous driving technology, the eye can include a combination of multiple sensors such as cameras, radar, and riders, while the brain can involve multiple artificial intelligence, machine learning, and deep learning systems. The goal is to have a human-like intelligence to fully understand the dynamic and fast-moving environment in real time and to act in response to changes in the environment.

The present application may be more fully understood in connection with the following detailed description in conjunction with the accompanying drawings, the accompanying drawings are not drawn to scale, and similar reference characters are used throughout in the attached drawings. Refers to a similar part.

In a vehicle showing an exemplary environment in which an autonomous beam steering radar system is used to detect and identify an object.

It is the schematic of the autonomous driving system for the autonomous vehicle by various examples.

It is the schematic of the beam steering radar system of FIG. 2 according to various examples.

It is a flowchart for operating the beam steering radar system implemented as shown in FIG. 3 according to various examples.

Radar signals and their associated scan parameters according to various embodiments are shown.

Graphs showing the distance resolution per bandwidth of a beam steering radar system according to various embodiments are shown.

Show an exemplary trade-off between scan parameters and design goals. Show an exemplary trade-off between scan parameters and design goals.

A flowchart for adjusting scan parameters according to various embodiments is shown.

Shown are exemplary scan parameters for the adjustment of FIG. 8A.

A flowchart for adjusting scan parameters according to various embodiments is shown.

A flowchart for adjusting scan parameters according to various embodiments is shown.

An exemplary scan parameter for the adjustment of FIG. 10 is shown.

Methods and devices for object detection using beam steering radar and convolutional neural network systems are disclosed. Methods and devices take raw data from a beam steering radar in an autonomous vehicle and process that data through a recognition module to provide information about multiple objects within the vehicle's field of view (“Field-of-View, FoV”). To extract and include. This information may be a parameter, measurement, or descriptor of the detected object such as position, size, velocity, object category. Objects include structural elements within the vehicle's FoV, such as roads, walls, buildings, and the central line of the road, as well as other vehicles, pedestrians, bystanders, cyclists, plants, trees, animals, and the like. obtain. The beam steering radar incorporates at least one beam steering antenna that is dynamically controlled to change its electrical or electromagnetic configuration to enable beam steering. Dynamic control is assisted by a recognition module, which, when identifying an object in the vehicle FoV, steers the beam and focuses on the area of interest by adjusting its radar scan parameters to the beam steering radar. Inform.

It is understood that the following description provides a number of specific details to provide a complete understanding of the examples. However, it is understood that the examples can be carried out without limiting these specific details. In other examples, well-known methods and structures may not be described in detail to avoid unnecessarily obscuring the description of the examples. Moreover, the examples may be used in combination with each other.

FIG. 1 shows an exemplary environment in which a beam steering radar system in an autonomous vehicle is used to detect and identify an object. Own vehicle 100 is an autonomous vehicle having a beam steering radar system 106 for transmitting radar signals to scan FoVs or specific areas. As described in more detail below, radar signals are transmitted according to a scan parameter set that can be tuned to result in multiple transmission beams 118. Scan parameters can include, among other things, the total angle of the scan area from the radar transmission point, the power of the transmitted radar signal, the scan angle of each incremental transmission beam, and the angle between each beam or overlaps between them. The entire FoV or a portion thereof can be scanned by compiling such a transmission beam 118, which may be in continuous adjacent scan positions or in a specific or random order. It should be noted that the term FoV is used herein with reference to radar transmission and does not mean an optical FoV having an unobstructed view. The scan parameters may also indicate the time interval between these incrementally transmitted beams, as well as the start and stop angle positions for a full or partial scan.

In various examples, the own vehicle 100 may also have other recognition sensors such as the camera 102 and the rider 104. These recognition sensors are not essential for the own vehicle 100, but may be useful for increasing the object detection capability of the beam steering radar system 106. The camera sensor 102 can be used to detect visible objects and conditions and to support the performance of various functions. The rider sensor 104 can also be used to detect an object and provide this information to adjust vehicle control. This information may include information such as highway congestion, road conditions, and other conditions that affect vehicle sensors, actions, or movements. Camera sensors are currently used in advanced driver assistance systems (“ADAS”) to assist drivers in driving functions such as parking (eg, rear cameras). Cameras are capable of capturing texture, color, and contrast information, similar to the human eye, albeit at a high level of detail, but they are susceptible to harmful weather conditions and lighting fluctuations. The camera 102 may have high resolution but cannot resolve objects larger than 50 meters.

A lidar sensor typically measures the distance to an object by calculating the time taken by the pulse of light moving to and back to the sensor. When positioned on top of the vehicle, the rider sensor can provide a 360 ° 3D view of the surrounding environment. Other approaches can use several riders in different locations around the vehicle to provide a full 360 ° view. However, rider sensors such as the Rider 104 are still very expensive, bulky in size, sensitive to weather conditions, and limited to short distances (typically <150-200 meters). Radar, on the other hand, has been used in vehicles for many years and operates in all weather conditions. Radar also has the advantage of using much less processing than other types of sensors, detecting objects behind obstacles, and determining the speed of moving objects. At resolution, the rider's laser beam is focused in a small area, has a wavelength smaller than the RF signal, and can achieve a resolution of about 0.25 degrees.

In various embodiments, the beam steering radar system 106 can provide a 360 ° true 3D vision of the vehicle's path and surroundings and a human interpretation, as described in more detail below. The radar system 106 uses a beam steering antenna module (having at least one beam steering antenna) to form and steer an RF beam in all directions within 360 ° FoV to quickly and more than about 300 meters of an object. Can be recognized with high accuracy over long distances. The long-range capabilities of the radar 106 as well as the short-range capabilities of the camera 102 and the rider 104 allow the sensor fusion module 108 within the own vehicle 100 to enhance its object detection and identification.

Here, attention is paid to FIG. 2, which shows a schematic diagram of an autonomous driving system for the own vehicle according to various embodiments. The autonomous driving system 200 is a system for use in own vehicle that provides partial or full automation of driving functions. Driving functions include maneuvering, accelerating, braking, and monitoring the surrounding environment and driving conditions, such as changing diagonal lines or speeds when necessary to avoid traffic, crossing pedestrians, animals, etc. It may be included. The autonomous driving system 200 includes a beam steering radar system 202 and other sensor systems such as camera 204, rider 206, infrastructure sensor 208, environment sensor 210, motion sensor 212, user preference sensor 214, and other sensors 216. including. The autonomous driving system 200 also includes a communication module 218, a sensor fusion module 220, a system controller 222, a system memory 224, and a V2V communication module 226. It is understood that this configuration of the autonomous driving system 200 is an exemplary configuration and is not intended to be limited to the particular structure exemplified in FIG. Additional systems and modules not shown in FIG. 2 may be included in the autonomous driving system 200.

In various embodiments, the beam steering radar system 202 is at least one beam for providing a dynamically controllable and maneuverable beam that can focus on one or more parts of the 360 ° FoV of the vehicle. Includes steering antenna. The beam emitted from the beam steering antenna is reflected from the vehicle's path and objects in the surrounding environment and is received and processed by the radar system 202 to detect and identify the object. The radar system 202 includes a recognition module that is trained to detect, identify, and control the radar module, if desired. Camera sensors 204 and rider 206 may also be used to identify objects in the vehicle's path and surroundings at very low distances.

Infrastructure sensor 208 can provide information from the infrastructure in operation, such as traffic warnings and indicators, including smart road configurations, billboard information, traffic lights, stop signs, traffic warnings, and the like. This is a growing area and the uses and capabilities extracted from this information are endless. The environmental sensor 210 detects various conditions such as temperature, humidity, fog, visibility, and rainfall, among others. The motion sensor 212 provides information about the functional motion of the vehicle. This can be tire pressure, fuel level, brake wear, etc. The user preference sensor 214 can be configured to detect conditions that are part of the user preference. This may be temperature control, smart window shading, etc. Other sensors 216 may include additional sensors for monitoring conditions in and around the vehicle.

In various embodiments, the sensor fusion module 220 optimizes these various features to provide a more or less comprehensive view of the vehicle and environment. Many types of sensors may be controlled by the sensor fusion module 220. These sensors may coordinate with each other to share information and consider the impact of one control action on another system. In one embodiment, under congested driving conditions, the noise detection module (not shown) may identify the presence of multiple radar signals that may interfere with the vehicle. This information can be used by a recognition module within the radar 202 to adjust the radar scan parameters to avoid these other signals and minimize interference.

In another embodiment, the environmental sensor 210 can detect that the weather is changing, reducing visibility. In this situation, the sensor fusion module 220 can be configured to improve the ability of the vehicle to navigate in these new conditions. This configuration may include turning off cameras or lidar sensors 204-206, or reducing the sampling rate of these visibility-based sensors. This depends on the sensor (s) adapted to the current situation. In response, the recognition module configures radar 202 for these conditions. For example, radar 202 can reduce the beamwidth to provide a more focused beam and thus can provide finer sensing capabilities.

In various embodiments, the sensor fusion module 220 can transmit control directly to the metastructured antenna based on historical conditions and controls. The sensor fusion module 220 may also use some of the sensors in the system 200 to serve as feedback or calibration for other sensors. In this way, the motion sensor 212 may provide feedback to the recognition module and / or the sensor fusion module 220 to create templates, patterns, and control scenarios. These can be based on successful actions or based on inadequate results that the sensor fusion module 220 learns from past actions.

The data from sensors 202-216 can be combined within the sensor fusion module 220 to improve the target detection and identification performance of the autonomous driving system 200. The sensor fusion module 220 may itself be controlled by a system controller 222 capable of interacting with and controlling other modules and systems in the vehicle. For example, the system controller 222 turns different sensors 202-216 on and off, as desired, or a driving hazard (a deer, a pedestrian, a cyclist, or another vehicle that suddenly appears in the vehicle's path). , Flying debris, etc.) may be identified and the vehicle may be instructed to stop.

All modules and systems in the autonomous driving system 200 communicate with each other via the communication module 218. The autonomous driving system 200 also includes a system memory 224 capable of storing information and data (eg, static and dynamic data) used in the operation of the system 200 and its own vehicle using the system 200. The V2V communication module 226 is used for communication with other vehicles. V2V communication may also include information from other vehicles that are not visible to the user, driver, or rider of the vehicle and can assist in vehicle coordinates to avoid accidents.

FIG. 3 shows a schematic diagram of the beam steering radar system of FIG. 2 according to various embodiments. The beam steering radar system 300 is a "digital eye" with a true 3D vision and can be interpreted by humans in the world. The "digital eye" and human-like interpretive power are provided by two main modules, the radar module 302 and the recognition module 304. Radar module 302 includes at least one beam steering antenna 306 for providing a dynamically controllable and maneuverable beam capable of focusing on one or more parts of the 360 ° FoV of the autonomous vehicle. It should be noted that current beam steering antenna implementations can direct beams from 120 to 180 ° FoV. Multiple beam steering antennas may be required to provide maneuverability to reach the entire 360 ° FoV.

In various embodiments, the beam steering antenna 306 is integrated with the RFIC to provide RF signals at multiple steering angles. The antenna may be a metastructured antenna, a phase array antenna, or any other antenna capable of emitting RF signals at millimeter wave frequencies. A metastructure generally defined herein is an engineering structure that can control and manipulate incident radiation in a desired direction based on its geometry. Metastructured antennas include, for example, feed or power split layers for splitting power and providing impedance matching, RF circuit layers with RFICs for steering angle control and other functions, and multiple microstrips, gaps, etc. It may include various structures and layers, including metastructured antenna layers with patches, vias, etc. The metastructure layer may include a metamaterial layer. Various configurations, shapes, designs, and dimensions of the beam steering antenna 306 can be used to implement specific designs and meet specific constraints.

Radar control is partially provided by the recognition module 304. The radar data generated by the radar module 302 is provided to the recognition module 304 for object detection and identification. Radar data is acquired by transmitter / receiver 308, which has a radar chipset capable of transmitting RF signals radiated by the metastructured antenna 306 and receiving reflections of these RF signals. Object detection and identification within the recognition module 304 is performed within the machine learning module (“MLM”) 312 and the classifier 314. Upon identifying an object in the vehicle's FoV, the recognition module 304 provides object data and control to the antenna controller 310 and scan parameter controller 316 in the radar module 302 to adjust beam steering and beam characteristics as needed. Provide instructions. The antenna control device 310 controls antenna parameters such as the steering angle, while the scan parameter control device 316 adjusts the scan parameters of the radar signal in the transmitter / receiver 308. For example, the recognition module 304 detects the cyclist on the path of the vehicle and directs the radar module 302 to add at a given steering angle and within the portion of the FoV corresponding to the cyclist's location. RF beam can be focused.

MLM312 in various embodiments implements CNN, which in various embodiments is a fully convolutional neural network (“FCN”) with three stacked convolutional layers from input to output (additional layers are also on CNNs). May be included). Each of these layers also performs a rectified linear activation function and batch normalization as an alternative to conventional L2 normalization, and each layer has 64 filters. Unlike many FCNs, the data is relatively small in size and runs uncompressed, so the data is not compressed as it propagates through the network. In various embodiments, the CNN may be trained with raw radar data, synthetic radar data, rider data, etc., and may be retrained with radar data. Multiple training options can be implemented to train CNNs to achieve good object detection and identification performance.

The classifier 314 also includes a CNN or other object classifier to improve the object identification capability of the recognition module 304 using the velocity information and microdropler signature in the radar data acquired by the radar module 302. But it may be. When an object is moving slowly or off the road lane, it is not a motorized vehicle, but rather a person, animal, cyclist, etc. Similarly, if one object is moving at high speed, but is slower than the average speed of another vehicle on the highway, the classifier 314 will use this speed information to move that vehicle more slowly. Determine if it is a prone track or another object. Similarly, the location of objects, such as highways in some countries (eg, the United States), indicates slow-moving vehicles. If the movement of the object does not follow the path of the road, the object can be an animal such as a deer running across the road. All of this information comes from the various sensors and information available to the vehicle, including information provided by weather and traffic services, as well as information provided by other vehicles such as smart roads and smart traffic signs or the environment itself. Can be judged.

Note that the velocity information is specific to the radar sensor. Radar data is a multidimensional format with data taples of the form _{(ri i} , θ _i , φ _i , I _i , vi _{i ), where r i} , θ _i , φ _i are the location coordinates of the object. the stands, r _i denotes the distance or spacing between the radar system 300 and the object along its line of sight, theta _i is the azimuthal angle, phi _i are elevation, I _i is the transceiver the intensity or reflectance indicates the amount of transmit power is returned to 308, v _i is the velocity between the radar system 300 and the object along its line of sight. The location and speed information provided by the recognition module 304 to the radar module 302 allows the antenna controller 310 and the scan parameter controller 316 to adjust their parameters as appropriate.

Focusing on FIG. 4, a flowchart for operating the beam steering radar system implemented as shown in FIG. 3 is shown. First, the beam steering radar system initiates transmission of a beam steering radar scan using the scan parameter set (400). The radar scan may be a frequency-modulated continuous wave (“FMWC”) signal in various embodiments. The FMCW signal allows the radar system to measure the distance to an object by measuring the phase or frequency difference between the transmitted signal and the received / reflected signal or echo. To do. Within the FMCW format, there are various modulation patterns, including triangles, saws, rectangles, etc., each with its own advantages and purposes. Within the FMCW signal, there may be multiple waveforms or chirps, each corresponding to the transmitted beam. When the transmitting beam encounters an object, the beam reflects off the object and the return signal or echo is received by the radar system (402).

The echo is analyzed by the MLM in the recognition module 304 to detect the object (404). If no object is detected, the beam steering radar system 300 continues to wait for an echo with further scanning. Note that the beam steering radar system does not stop the transmitting beam. The scan is accomplished as long as the vehicle is in operation. When the recognition module 304 indicates that the object has been detected in the echo, the object information such as the location of the object and its velocity is extracted and sent to the radar module 302 (406). The recognition module 304 may also transmit information to focus the radar beam on the next scan. The object information notifies the scan parameter controller 316 that the scan parameters are adjusted (408) in various ways as described below with reference to FIGS. 5-11. The recognition module 304 then classifies the object (410), sends the object classification result to the sensor fusion module in the vehicle, combines it with object detection / classification from other sensors, and if necessary, which control action in the vehicle. Determine if (eg, speed, lane change, break, etc.) is taken.

FIG. 5 shows the radar signal and its associated scan parameters in more detail. The radar signal 500 is an FMCW signal including a series of chirps such as chirps 502 to 506. The signal 500 is defined by a set of parameters that affect the location of the object, its resolution, and the way it determines its velocity. Associated with the signal 500, parameters shown in Figure 5, (1) _{f max} and _f min to the minimum frequency and maximum frequency of the chirp signal, _{T total} to the total time of (2) one chirp sequence, (3) Radar _{T delay} , which represents the settling time of the phase locked loop (“PLL”) in the system, (4) T _{MEAS of the} actual measurement time (for example, a chirp sequence for detecting an object within 300 meters> 2 μs), (5) T _{chip for the total time of one chirp, (6) Repeat} of the repetition time between chirps (7) B for the total bandwidth of the chirp, (8) B _eff of the effective bandwidth of the chirp, (9) .DELTA.B _eff for bandwidth between successive measurements, (10) the number of measurements made for each chirp, _{N r} (i.e., for each chirp, or make measurements at times echo), and (11 ) Includes the number of chirps per sequence, N _c.

物体の速度及び速度解像度は、チャープシーケンスパラメータによっても完全に決定される。達成される最小速度又は解像度は、以下のように決定される（ｃは光の速度を表す）：

より高いレーダー周波数は、同じシーケンスパラメータに対してより良好な速度解像度を与えることに留意されたい。最大速度は、以下によって与えられる。

The object distance and the distance resolution, is completely determined by the chirp parameter N _r and B _eff as follows, as shown in FIG. 6, with reference to the graph 600 indicating the distance resolution per bandwidth B, completely It is determined.

The velocity and velocity resolution of the object are also completely determined by the chirp sequence parameters. The minimum speed or resolution achieved is determined as follows (where c represents the speed of light):

Note that higher radar frequencies give better velocity resolution for the same sequence parameters. The maximum speed is given by:

Further relationships between scan parameters are given by the following equations, in which Equation 5 represents the sample velocity, Equation 6 represents the time Δt between continuous measurements, and Equation 7 represents the distance resolution. , Equation 8 represents the charp time and Equation 9 represents the maximum speed:

Note that once the parameters R _max are fixed, v _max and R _min are no longer independent. Also, the sample velocity f _sample of Equation 5 also determines how the definition of distance resolution can be achieved for the selected maximum velocity and distance. It should be noted that the following equations 1-9 can be used to establish scan parameters for a given design goal. For example, in order to detect an object at a long distance and with high resolution, the radar system 300 needs to take a large number of measurements per chirp. If the goal is to detect objects over long distances and at high speeds, the chirp time must be low, limiting the chirp time. In the first case, high resolution detection over long distances is limited by the bandwidth of the signal processing unit in the radar system. In the second case, the maximum speed over a long distance is limited by the data acquisition speed (which also limits the resolution) of the radar chipset.

7A-B show exemplary trade-offs between scan parameters and design goals. In FIG. 7A, Table 700 shows the values of scan parameters for a single chirp sequence whose target is to detect an object over long distances. It should be noted that increasing the velocity and distance of the object results in lower distance resolution. It should also be noted that velocity resolution can be increased by adding more chirps to the sequence at the cost of increasing total sequence time. In FIG. 7B, Table 702 shows the values of scan parameters for a single chirp sequence where the target is to detect an object at close range. The distance resolution can be improved to a fine 0.2 meter when the maximum speed of the detected object is less than 20 m / s. In addition, it should be noted that better velocity resolution increases sequence time. For example, a velocity resolution of 0.3 m / s requires a sequence time of 7.1 ms.

The trade-off between scan parameters and design goals in terms of object detection capabilities allows multiple ways to adjust scan parameters during the operation of the beam steering radar system. 8A, 9 and 10 show flowcharts for adjusting scan parameters according to various embodiments, if desired. In FIG. 8A, the beam steering radar system has fewer chirps in the sequence, for example, by reducing the number of measurement points and / or chirps per chirp to reduce calculation time, or to reduce measurement time. Perform a low resolution scan first by using (800). This scan is followed by a higher resolution scan (802). Higher resolution scans, for example, within areas of expected traffic (eg, by using map tools to predict where the road is 300 meters away from the front of the vehicle), lower resolution. It takes place within an area flagged as containing one or more objects during the scan, or within an area flagged by another sensor, such as the vehicle's camera and / or rider sensor. FIG. 8B shows exemplary scan parameters for low resolution scans in row 806 and high resolution scans in row 808. The high resolution scan is performed twice as long as the measurement time and four times as many data points for processing as a low resolution scan.

In FIG. 9, the beam steering radar system first performs a scan with a wider beam width (900) and then rescans the FoV with a narrower beam (902). Similar to the higher resolution scans in FIG. 8, narrower scans are, for example, within areas with expected traffic, within areas flagged as containing one or more objects during low resolution scans. Or within the area flagged by other sensors, such as the camera and / or rider sensor in the vehicle. When the beam steering radar system is implemented with a plurality of antennas, both operations can be executed in parallel. Crosstalk can be omitted by operating each set of antennas in different frequency subbands within the permissible frequency band within the range of 76-81 GHz. It should also be noted that antennas with double beamwidth allow fixed FoVs to be scanned in half using the same scan parameters.

In FIG. 10, the beam steering radar system performs a first scan with a high range resolution, but performs a lower absolute velocity capture (1000). This is followed by a second scan (1002) with a lower range resolution, but with a high absolute velocity capture. Corresponding exemplary scan parameters are shown in FIG. 11 of Table 1100, where row 1102 shows a high range resolution scan with a maximum speed of 30 m / s and line 1104 shows a low range resolution scan with a maximum speed of 85 m / s. Regardless of the method used to adjust the scan parameters, the beam steering radar system can achieve multiple design goals and detect objective locations and velocities at both short and long distances. Objects are further classified into MLMs in a recognition module connected to a radar module, as shown in FIG.

These various examples support improved sensor performance, all-weather / all-state detection, advanced decision-making algorithms, and interaction with other sensors through sensor fusion. These configurations optimize the use of radar sensors because the radar is not hampered by weather conditions in many applications, such as self-driving a car. The radars described herein are effectively "digital eyes", have a true 3D vision, and can be interpreted by humans in the world.

It is understood that previous description of the disclosed examples will be provided to allow one of ordinary skill in the art to make or use the present disclosure. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein can be applied to other embodiments without departing from the spirit or scope of the present disclosure. Accordingly, this disclosure is not intended to be limited to the embodiments presented herein, but is given the broadest scope consistent with the principles and novel features disclosed herein. ..

Claims (20)

A radar system in an autonomous vehicle for object detection and classification.
It ’s a radar module,
With at least one beam steering antenna
With the transmitter / receiver
A radar module comprising a scan parameter controller adapted to adjust the scan parameter set of the transmitter / receiver.
A recognition module including a machine learning module and a classifier for detecting and classifying objects in the path of the autonomous vehicle and the surrounding environment, and a recognition module that transmits object data and radar control information to the radar module. A radar system equipped with. The first aspect of claim 1, wherein the radar module further includes an antenna control device for controlling the at least one beam steering antenna according to the object data and the radar control information transmitted by the recognition module to the radar module. Radar system. The radar system according to claim 1, wherein the machine learning module includes a convolutional neural network. The radar system of claim 1, wherein the transmitter / receiver initiates transmission of a beam steering radar scan using the scan parameter set. The radar system according to claim 4, wherein the scan parameter set is adjusted according to the object data and the radar control information transmitted by the recognition module to the radar module. The radar system according to claim 4, wherein the radar scan includes an FMCW signal. The scan parameters are the minimum and maximum frequencies of the FMCW signal, the total time of the charp sequence in the FMCW signal, the settling time of the phase lock loop, the measurement time, the total time of one chapter in the FMCW signal, and the FMCW signal. Repetition time between chapters, total chap bandwidth in said FMCW signal, effective bandwidth of chap in said FMCW signal, bandwidth between successive measurements, number of measurements taken per chap in said FMCW signal, and per sequence. The radar system of claim 6, comprising a parameter selected from the group consisting of the number of charps of. Object detection and classification method
Starting transmission of beam steering radar scan using a transmitter / receiver,
Receiving echo and
Detecting an object in the received echo and
Extracting the object information about the detected object and
A method for detecting and classifying objects, including adjusting a set of scan parameters in said transmitter / receiver for the next beam steering radar scan. The object detection and classification method according to claim 8, wherein the beam steering radar scan is an FMCW signal. The object detection and classification method according to claim 8, further comprising classifying the detected object by a convolutional neural network and a classifier. The object detection and classification method according to claim 8, wherein the object information includes an object location and a velocity. The object detection and classification method according to claim 8, further comprising transmitting object classification information to the sensor fusion module. The scan parameters are the minimum and maximum frequencies of the FMCW signal, the total time of the chirp sequence in the FMCW signal, the settling time of the phase lock loop, the measurement time, the total time of one chirp in the FMCW signal, and the FMCW signal. Repetition time between chapters, total chirp bandwidth in the FMCW signal, effective chirp bandwidth in the FMCW signal, bandwidth between successive measurements, number of measurements made per chirp in the FMCW signal, and per sequence. The object detection and classification method according to claim 9, which comprises a parameter selected from the group consisting of the number of chirps of. Object detection and classification method
Performing the first scan with the first scan parameter set,
Adjusting the first scan parameter set based on the object detection information from the received echo,
An object detection and classification method comprising performing a second scan with a second parameter set adjusted from the first scan parameter set. The object detection and classification method according to claim 14, wherein the first scan has a lower resolution than the second scan. The object detection and classification method according to claim 14, wherein the first scan is performed on a beam steering antenna having a first beam width. The object detection and classification method according to claim 14, wherein the second scan is performed on a beam steering antenna having a second beam width, and the second beam width is narrower than the first beam width. .. The object detection and classification method according to claim 14, wherein the second scan has a lower resolution than the first scan and a faster capture than the first scan. The first scan parameter set includes the minimum and maximum frequencies of the FMCW signal, the total time of the chirp sequence in the FMCW signal, the settling time of the phase lock loop, the measurement time, the total time of one chirp in the FMCW signal. The repetition time between chirps in the FMCW signal, the total bandwidth of the chirps in the FMCW signal, the effective bandwidth of the chirps in the FMCW signal, the bandwidth between successive measurements, the number of measurements taken per chirp in the FMCW signal. The object detection and classification method according to claim 14, further comprising a parameter selected from the group consisting of, and the number of chirps per sequence. The object detection and classification method according to claim 14, wherein the first and second scans are performed on a metastructured antenna.

Images