Development of a new radar

Abstract

The present invention relates to the integration of various sensors used in vehicles into a single sensor. The invention specifically relates to the development of a system using a single piece of equipment to enhance safety during reverse maneuvers through a precise rear-view alarm and parking assistance, hands-free trunk-opening (Radar Kick Sensor), blind spot detection (Blind Spot Detection BSD) to prevent accidents by detecting vehicles in the driver's blind spots, and a rear pre-crash warning system (Rear Pre-crash Warning System RPCW) that alerts the driver of possible rear-end collisions.

Description

[0001] DESCRIPTION

[0002] DEVELOPMENT OF A NEW RADAR

[0003] Technical Field

[0004] The present invention relates to the integration of various sensors used in vehicles into a single sensor. In particular, the invention relates to the development of a system using a single piece of equipment to enhance safety during reverse maneuvers through a precise rear-view alarm and parking assistance, hands-free trunk-opening (Radar Kick Sensor), blind spot detection (Blind Spot Detection BSD) to prevent accidents by detecting vehicles in the driver's blind spots, and a rear pre-crash warning system (Rear Pre-crash Warning System RPCW) that alerts the driver of possible rear-end collisions.

[0005] State of the Art

[0006] Sensors used in vehicles provide various functions in order to enhance vehicle safety, performance, and driving comfort. These sensors are designed to offer drivers a better driving experience and prevent accidents. Modern vehicle sensors operate using different technologies to perform various tasks. These sensors not only facilitate the parking of the vehicles, but also ensure a safe driving experience by detecting dangers that the driver may not be aware of. These sensors form the cornerstone of autonomous driving systems in vehicles and are equipped with continuously evolving technologies to enhance driving safety. The widespread usage of such sensors in modern vehicles contributes to the reduction in traffic accidents and a more enjoyable driving experience by increasing road safety.

[0007] Currently, one of the most commonly used sensors in vehicles is the parking sensor. Parking sensors are ultrasonic or radar sensors installed on the front and rear bumpers of vehicles. These sensors detect nearby objects and alert the driver. Typically, they operate via an audible alert (beep sound) or a visual indicator (distance display on the screen) within the vehicle. These sensors are used to prevent collisions while parking in tight spaces.

[0008] Blind spot sensors, which help to reduce dangers during driving, are used to monitor the vehicle's blind spots. These sensors are usually placed at the corners of the rear bumper and use radar technology. They detect vehicles that are not visible in the driver’s side mirrors and inform the driver by lighting up an indicator in the side mirrors or providing an audible alert. Thus, the risk of collision during lane changes is reduced. Another sensor that helps to reduce dangers during driving is the rear collision warning sensors. These sensors monitor the traffic behind the vehicle and alert the driver when an approaching vehicle is detected. Typically, radar or camera systems are used. If an approaching vehicle is closing in quickly, the system informs the driver via a visual and / or audible alert. This system is especially useful in reducing the risk of collisions during stop- and-go situations at traffic lights or in heavy traffic.

[0009] The kick sensor, which allows hands-free opening of the trunk, is generally used for opening the trunk lid. This sensor is placed under the rear bumper of the vehicle and is activated by moving the foot under the bumper. This movement is detected by the sensor, and the trunk lid is opened automatically. It offers a practical solution for opening the trunk when the hands are full. Kick sensors typically use capacitive sensors or motion detection technology.

[0010] Various technologies are currently used in the development of sensors. The most commonly used sensor systems include ultrasonic sensors, radar sensors, camera systems, lidar sensors, and capacitive sensors.

[0011] The purpose of using sensors in vehicles is to enhance vehicle safety, provide a comfortable driving experience, and improve usability. The use of such sensors in modern vehicles significantly improves the driving experience. However, the usage of a separate equipment for each sensor increases both the cost and the number of components around the vehicle.

[0012] Parking assistance systems in passenger vehicles are widely implemented by using ultrasonic sensors placed on the rear of the vehicle. Since ultrasonic sensors are mounted on the bumper, they incur additional installation costs but have the capability to detect objects at short distances. For Blind Spot Detection and Rear Pre-crash Warning Systems, cameras are widely used. However, the performance of the camera is insufficient under poor lighting / overly bright light and adverse weather conditions, leading to high False Alarm Rates (FAR).

[0013] There is a need for a system that can eliminate the disadvantages of ultrasonic sensors and cameras used in vehicles and reduce the use of multiple equipment.

[0014] Technical Problems that the Invention Aims to Solve

[0015] Ultrasonic sensors, frequently used in vehicles, are not successful in detecting distant targets. However, long-range detection is necessary for rear collision warning and blind spot detection applications. Therefore, these applications commonly rely on camera data for generating warnings. However, the performance of the camera sensor is quite low under adverse weather conditions and insufficient lighting. The aim is to develop a single sensor instead of both the camera and ultrasonic sensors by eliminating their disadvantages.

[0016] Frequency Modulated Continuous Wave (FMCW) radar systems that operates in the millimeter-wave band, have the capability to detect stationary and moving targets at the desired range distances along with the relative velocity of the targets according to the sensor, depending on the form of the signal transmitted by the radar. Thanks to this advantage of the radar, the system can detect targets both at short and long distances.

[0017] Another advantage of the invention is that the radar system, being Multi-Input Multi-Output (Ml MO), is capable of generating high-accuracy 3D position information of targets, depending on the structural arrangement and number of antennas. Currently, generating accurate information with a wide Field of View (FOV) using a small number of antennas is a significant challenge. By utilizing the arrangement of receiver and transmitter antennas, high-resolution data in both azimuth and elevation directions are obtained via a small number of antennas.

[0018] The antennas used in FMCW MIMO radar systems need to have different characteristics, depending on the application. For example, to detect targets at high ranges, the antennas need to have high directivity and gain. By considering the conservation of energy radiated from the antenna, as the directivity of the antenna increases with a fixed transmitter gain, the FOV gets narrow.

[0019] By considering the parameters mentioned above, different FOVs and detection ranges are needed for different applications. With the present invention, the sensor is configured in different modes, combining multiple advanced driver assistance system features into a single sensor. In this way, instead of installing multiple sensors in different parts of the vehicle, four different applications are carried out via a single sensor.

[0020] The figures below will be used for better understanding of the system according to the invention.

[0021] Explanation of The Figures

[0022] Figure 1 is a view of the radar system according to the invention that is mounted on the vehicle bumper. Figure 2 is a schematic illustration of the system architecture showing the connection of the radar system according to the invention with the vehicle.

[0023] Figure 3 is a schematic illustration of the components inside the radar system according to the invention.

[0024] Figure 4 is a schematic illustration of the antenna arrangement inside the radar system according to the invention.

[0025] Figure 5 is a schematic illustration of the implementation of the hands-free trunk-opening function of the radar system according to the invention.

[0026] Figure 6 is a schematic illustration of the blind spot detection system application of the radar system according to the invention.

[0027] Figure 7 is a schematic illustration of the rear collision warning system application of the radar system according to the invention.

[0028] Figure 8 is a schematic diagram of the operating method of the radar system according to the invention.

[0029] Reference Numbers for Parts and Sections to Aid in Explaining the Invention

[0030] 1- Vehicle

[0031] 2- CAN BUS line

[0032] 3- Radar system

[0033] 4- Radar detection block

[0034] 5- Radar analog external unit

[0035] 6- Radar digital signal processor

[0036] 7- Radar application processor

[0037] 8- Power stage

[0038] 9- Communication unit

[0039] TX- Transmitting antenna

[0040] RX- Receiving antenna

[0041] Names of Process Stages and Reference Numbers to Aid in Explaining the Invention

[0042] 100: The controlling stage whether the vehicle is parked

[0043] 200: The selecting stage of of hands-free trunk-opening function 300: The checking stage whether the vehicle is in reverse gear

[0044] 400: The selecting stage of the rear-view alarm and parking assistance function 500: The checking stage whether the vehicle is in forward gear

[0045] 600: The selecting stage of the rear collision warning and blind spot detection function

[0046] 700: FMCW signal-sending stage

[0047] 800: The stage of obtaining the ADC data from the reflected signals

[0048] 900: Range and Doppler spectral estimation stage

[0049] 1000: Adaptive thresholding and Doppler compensation stages

[0050] 1100: The estimating stage of the angle of arrival

[0051] 1200: Selecting stage of the radar system function

[0052] 1300: Clustering stage

[0053] 1400: Obtaining stage of radar outputs

[0054] 1500: Feature-extraction stage

[0055] 1600: Machine-learning stage

[0056] DETAILED DESCRIPTION OF THE INVENTION

[0057] The radar system (3) according to the present invention primarily relates to the integration of a precise rear-view alarm and parking assistance, hands-free trunk-opening (Radar Kick Sensor), blind spot detection (Blind Spot Detection BSD) to prevent accidents by detecting vehicles in the driver's blind spots, and a rear pre-crash warning system (Rear Pre-crash Warning System RPCW), into a single piece of equipment, in order to enhance safety in vehicles (1) during reverse maneuvers.

[0058] The illustrations and figures used to explain the invention are merely exemplary. Therefore, it should be noted that visual design flexibility is possible while maintaining the technical elements.

[0059] Figure 1 shows the view of the radar system (3) according to the invention mounted on the vehicle (1) bumper. Figure 2 represents the system architecture showing the connection of the radar system (3) with the vehicle (1). Figure 3 provides a schematic illustration of the components inside the radar system (3). Figure 4 illustrates the schematic illustration of the antennas’ arrangement within the radar system (3). Figure 5 shows a schematic illustration of the implementation of the hands-free trunk-opening function of the radar system (3). Figure 6 illustrates the schematic illustration of the blind spot detection system application of the radar system (3). Figure 7 provides a schematic illustration of the rear collision warning system application of the radar system (3). Figure 8 is the schematic diagram of the operating method of the radar system (3).

[0060] A sensor equipment according to the invention is a radar system (3), which is integrated with the CAN BUS line (2) of a vehicle (1) and receives speed, brake, gear, and signal information from the vehicle (1) and provides notifications based on sensor outputs, characterized by comprising: a radar detection block (4) configured to detect targets; at least three transmitter antennas (TX) that are arranged in the y-axis within the radar detection block (4) at intervals of half of the wavelength; at least four receiver antennas (RX) that are arranged in the x-axis within the radar detection block (4) at intervals of half the wavelength; at least one radar analog external unit (5) that transmits radar signals (FMCW signals) to the external environment using said transmitter antennas (TX), collects the signals reflected back from the targets, and digitizes them; at least one radar digital signal processor (6) that processes the digitized signals from said radar analog external unit (5) using digital signal processing methods to detect the targets with reduced noise and a radar application processor (7) that performs clustering, classification, and tracking of the targets detected by said radar digital signal processor (6) based on application-specific requirements; wherein said vehicle (1) comprises a power stage (8) that regulates the power received from a power supply to an appropriate voltage, filters out unwanted noise, and protects against disruptive effects such as reverse polarity and overvoltage, a communication unit (9) that establishes wired and wireless connections with a vehicle (1).

[0061] In the design of the series-fed microstrip antenna, the antenna dimensions increase to enhance directivity in both the azimuth (horizontal) and elevation (vertical) directions and to allow the radar signal to detect targets with low Radar Cross Section (RCS) at greater distances. In the present invention, the directional gain of the antennas designed in a MIMO structure with 3 transmitting antennas (TX) and 4 receiving antennas (RX) was selected to be low to keep the sensor size small. However, by ensuring that the antennas are fed and radiated simultaneously, the directivity of the antenna and, consequently, the signal-to- noise ratio of the radar signal for long-distance applications have been increased. When the radar system operates using Time Division Multiplexing (TDM), the structure formed by 3 transmitting antennas (TX) and 4 receiving antennas (RX) has 12 virtual antennas. With the present invention, the antennas are arranged in the MIMO structure with the transmitting antennas (TX) arranged along the y-axis and the receiving antennas (RX) arranged along the x-axis at intervals of half the wavelength, as shown in Figure 4, to achieve high resolution in both the azimuth and elevation axes. This arrangement allows for the utilization of high-resolution data obtained from both the azimuth and elevation axes for target detection.

[0062] The selection of the carrier frequency and bandwidth in radar signals varies depending on the application. In the present invention, the carrier frequency and bandwidth for the four different applications implemented with a single sensor have been configured appropriately. Transitions are made between the configurations that differ based on the speed, gear, etc., information received from the vehicle (1) for each application. In this way, the up-chirp signal designed for each application and the selected algorithm parameters produce results with a low false alarm rate that comply with regulatory requirements.

[0063] The radar system (3) of the present invention includes a method based on changing functions for carrying out different applications and generating radar outputs. When the process flow of this method is generally considered, it primarily comprisis:

[0064] The stage of checking whether the vehicle is parked (100),

[0065] The stage of selecting the hands-free trunk-opening function (200), The stage of checking whether the vehicle is in reverse gear (300), The stage of selecting the rear-view alarm and parking assistance function (400), The stage of checking whether the vehicle is in forward gear (500),

[0066] The stage of selecting the rear collision warning and blind spot detection function (600),

[0067] The stage of sending FMCW signals (700),

[0068] The stage of obtaining ADC data from the reflected signals (800),

[0069] The stage of range and Doppler spectral estimation (900),

[0070] The stage of adaptive thresholding and Doppler compensation (1000),

[0071] The stage of estimating the angle of arrival (1100),

[0072] The stage of selecting the radar system function (1200),

[0073] The clustering stage (1300),

[0074] The stage of obtaining radar outputs (1400), The feature extraction stage (1500),

[0075] The machine learning stage (1600).

[0076] The Stage of Checking Whether the Vehicle is Parked (100): In the stage of checking whether the vehicle is parked, the radar system (3) receives information from the vehicle's CAN BUS line (2) to determine whether the vehicle is parked. If the vehicle (1) is parked, the process continues with the stage of selecting the hands-free trunk-opening function (200); otherwise, it continues with the stage of checking whether the vehicle is in reverse gear (300).

[0077] The Stage of Selecting the Hands-Free trunk-opening Function (200): In the stage of selecting the hands-free trunk-opening function, if the vehicle (1) is parked and the user cannot open the trunk due to their hands being full, the function that allows the trunk to be opened with a specific foot movement is activated. The process then continues with the stage of sending FMCW signals (700).

[0078] The Stage of Checking Whether the Vehicle is in Reverse Gear (300): In the stage of checking whether the vehicle is in reverse gear, the radar system (3) receives information from the vehicle's CAN BUS line (2) to determine whether the vehicle is in reverse gear. If the vehicle (1) is in reverse gear, the process continues with the stage of selecting the rearview alarm and parking assistance function (400); otherwise, it continues with the stage of checking whether the vehicle is in forward gear (500).

[0079] The Stage of Selecting the Rear-View Alarm and Parking Assistance Function (400): In the stage of selecting the rear-view alarm and parking assistance function, two transmitting antennas (TX) in the radar operate in frequency-modulated continuous wave form with a 79 GHz carrier frequency and a 4 GHz bandwidth, via a time division multiplexing technique in three different modes simultaneously. The transmitting antennas (TX) send up-chirp signals within a 100 ms window. The transmitting antennas (TX) simultaneously radiate. With this function, targets with different sizes and electrical characteristics, and a relative speed of ±2.5 m / s, located at a minimum distance of 30 cm and a maximum distance of 2 m from the vehicle (1), are detected. As a result, targets that were difficult or problematic to detect (especially dry grass, bushes, etc.) are successfully detected at the required resolution and distance with a low false alarm rate compared to the state of the art. The process then continues with the stage of sending FMCW signals (700). The Stage of Checking Whether the Vehicle is in Forward Gear (500): In the stage of checking whether the vehicle is in forward gear, the radar system (3) receives information from the vehicle's CAN BUS line (2) to determine whether the vehicle is in forward gear. If the vehicle (1) is in forward gear, the process continues with the stage of selecting the rear collision warning and blind spot detection function (600).

[0080] The Stage of Selecting the Rear Collision Warning and Blind Spot Detection Function (600): In the stage of selecting the rear collision warning and blind spot detection function, three transmitting antennas (TX) in the radar are simultaneously activated and operate in frequency-modulated continuous wave form with a 76.5 GHz carrier frequency band and a 1 GHz bandwidth. The transmitting antennas (TX) send up-chirp signals within a 100 ms frame window. To successfully detect targets at different angles with a high signal-to-noise ratio at long ranges, the three transmitting antennas (TX) arranged along the y-axis are simultaneously activated. With this function, moving targets approaching the vehicle (1) at a maximum distance of 50 m and a maximum relative speed of 30 m / s (~2.5m2RCS) are successfully detected with the radar system (3) placed only under the bumper without installing systems at different points of the vehicle.

[0081] The Stage of Sending FMCW Signals (700): In the stage of sending FMCW signals, the radar system (3) sends FMCW signals from the transmitting radar antennas according to the function of the radar system (3) in a sawtooth waveform.

[0082] The Stage of Obtaining ADC Data from the Reflected Signals (800): In the stage of obtaining ADC data from the reflected signals, FMCW signals (6) emitted from the transmitting antennas (TX) hit one or more targets, reflect from the target, and return to the receiving antennas (RX) of the radar. Raw ADC data is obtained from each receiving antenna (RX) of the radar.

[0083] The Stage of Range and Doppler Spectral Estimation (900): In the stage of range and Doppler spectral estimation, the raw ADC data obtained is passed through a spectral imaging algorithm with high resolution and low computational load in both the range and Doppler directions. In the spectral imaging methods known in the state of the art, the correct estimation of frequency is directly related to the sampling stage used in the spatial transformation. Selecting a low sampling stage increases the accuracy of frequency estimation but also requires increasing the memory allocated for this process on the processor side. It is necessary to obtain high-accuracy information with minimal data to use the memory unit effectively. With the mentioned spectral imaging algorithm, high-frequency estimation with small data is achieved.

[0084] The Stage of Adaptive Thresholding and Doppler Compensation (1000): The up-chirp signals sent from the transmitting antennas (TX) are transmitted using the TDM method. Therefore, there is a time delay between the radar signals corresponding to each transmitter-receiver antenna (TX) (RX) pair. In the stage of adaptive thresholding and Doppler compensation, the time delay is added as a phase term to the radar signal read from the receiving antenna (RX) (Doppler compensation), and target detection is performed using a thresholding method with a False Alarm Rate that varies depending on the range. This allows the radar to store information about both small nearby targets and large distant targets in its memory unit without increasing the memory unit's capacity.

[0085] The Stage of Estimating the Angle of Arrival (1100): The number of antennas in the azimuth and elevation directions is one of the most important parameters that determine the resolution in direction finding. The points processed in the stage of adaptive thresholding and Doppler compensation (1000) are passed through the angle-of-arrival estimation algorithm to determine the direction. This algorithm allows for accurate determination of the angle at which the signal reflected from the target arrives. With this algorithm, high angular accuracy target detection points are obtained with a small number of antennas.

[0086] The Stage of Selecting the Radar System Function (1200): In the stage of selecting the radar system function, if the vehicle (1) is parked, the algorithm continues from the feature extraction stage (1500); if the vehicle is not parked, the algorithm continues from the clustering stage (1300).

[0087] The Clustering Stage (1300): Successfully clustering high-resolution points belonging to each target is crucial for producing outputs with a low false alarm rate. The clustering algorithms known in the literature fall short in clustering targets with different RCS at different ranges. If the vehicle (1) is not parked, the process continues with the clustering stage (1300). In this step, a density-based clustering algorithm (DBSCAN-Density Based Spatial Clustering of Applications with Noise) is designed to successfully cluster points belonging to both large targets at close range and small targets at long range. In the mentioned clustering algorithm, adaptive parameter selection is made based on the target range and radar cross-section to reduce the radar's false alarm rate. This way, the most meaningful combinations of imaging spaces are determined, and the information obtained from these spaces is combined. The Stage of Obtaining Radar Outputs (1400): In the stage of obtaining radar outputs, high-resolution position and speed information related to stationary and moving targets is obtained from the points obtained in the clustering stage (1300).

[0088] Feature Extraction Stage (1500): If the vehicle (1) is parked, the process continues with the feature extraction step. Meaningful features are obtained with the outputs produced by the angle-of-arrival estimation algorithm resulting from previous data.

[0089] Machine Learning Stage (1600): In the machine learning step, the features obtained in the feature extraction stage (1500) are used in classifying the movement to be performed in the region defined by the machine learning method.

[0090] INDUSTRIAL APPLICABILITY OF THE INVENTION The radar system according to the invention reduces the number of various sensors commonly used in vehicles to increase driving safety and assist the driver by integrating them into a single system.

Claims

CLAIMS1. A radar system (3) which comprises a power stage (8) that regulates the power received from a power supply to an appropriate voltage, filters out unwanted noise, and protects against disruptive effects such as reverse polarity and overvoltage, a communication unit (9) that establishes wired and wireless connections with a vehicle (1) and is integrated with the CAN BUS line (2) of the vehicle (1) and receives speed, brake, gear, and signal information from the vehicle (1) and provides notifications based on sensor outputs, characterized by comprising: a radar detection block (4) configured to detect targets; at least three transmitter antennas (TX) that are arranged in the y-axis within the radar detection block (4) at intervals of half of the wavelength; at least four receiver antennas (RX) that are arranged in the x-axis within the radar detection block (4) at intervals of half the wavelength; at least one radar analog external unit (5) that transmits radar signals (FMCW signals) to the external environment using said transmitter antennas (TX), collects the signals reflected back from the targets, and digitizes them; at least one radar digital signal processor (6) that processes the digitized signals from said radar analog external unit (5) using digital signal processing methods to detect the targets with reduced noise and a radar application processor (7) that performs clustering, classification, and tracking of the targets detected by said radar digital signal processor (6) based on application-specific requirements.

2. The radar system (3) according to claim 1 , characterized by comprising at least 3 transmitting antennas (TX) and at most 4 receiving antennas (RX) that are fed and radiate simultaneously, depending on the operating modes.

3. A radar system (3) according to claim 1, characterized by comprising at least 3 transmitting antennas (TX) and at most 4 receiving antennas (RX) that are fed and radiate on a time-divided basis, depending on the operating modes.

4. A method that is developed for integrating sensors used in vehicles (1) into a single sensor equipment, characterized by comprising the following stages: the stage of checking whether the vehicle is parked (100),the stage of selecting the hands-free trunk-opening function (200), the stage of checking whether the vehicle is in reverse gear (300), the stage of selecting the rear-view alarm and parking assistance function (400), the stage of checking whether the vehicle is in forward gear (500), the stage of selecting the rear collision warning and blind spot detection function (600), the stage of sending FMCW signals (700), the stage of obtaining ADC data from the reflected signals (800), the stage of range and Doppler spectral estimation (900), the stage of adaptive thresholding and Doppler compensation (1000), the stage of estimating the angle of arrival (1100), the stage of selecting the radar system function (1200), the clustering stage (1300), the stage of obtaining radar outputs (1400), the feature extraction stage (1500), the machine learning stage (1600).

5. A radar system (3) according to claim 3, characterized by comprising the process stages of activating the function that allows the trunk to be opened with a specific foot movement in the stage of selecting the hands-free trunk-opening function (200).

6. A radar system (3) according to claim 3, characterized by comprising the process stages of operating in frequency-modulated continuous wave form with a 79 GHz carrier frequency and a 4 GHz bandwidth via a time division multiplexing technique with two simultaneously operating transmitting antennas (TX) in the stage of selecting the rear-view alarm and parking assistance function (400), and sending up-chirp signals within a 100 ms frame window.

7. A radar system (3) according to claim 3, characterized by comprising the process stages of operating in frequency- modulated continuous wave form with a 76.5 GHz carrier frequency band and a 1 GHz bandwidth via three simultaneously-activated transmitting antennas (TX) in the stage of selecting the rear collision warning and blind spot detection function (600), and sending up-chirp signals within a 100 ms window.

8. A radar system (3) according to claim 3, characterized by comprising the process stages of sending Frequency Modulated Continuous Wave signals (6) from the transmittingantennas (TX) of the radar system (3) in a sawtooth waveform depending on the function of the radar system (3) in the stage of sending FMCW signals (700).

9. A radar system (3) according to claim 3, characterized by comprising the process stages of passing the raw ADC data which are obtained in the stage of range and Doppler spectral estimation (900), through a spectral imaging algorithm in the range and Doppler directions,10. A radar system (3) according to claim 3, characterized by comprising the process stages of thresholding the radar signal which is read from the receiving antenna (RX) with a Constant False Alarm Rate detection algorithm and adding Doppler compensation terms in the stage of adaptive thresholding and Doppler compensation (1000).

11. A radar system (3) according to claim 3, characterized by comprising the process stages of detecting the direction by passing the points which are obtained in the stage of estimating the angle of arrival (1100), from an angle-of-arrival estimation algorithm.

12. A radar system (3) according to claim 3, characterized by comprising the process stages of operating a density-based clustering algorithm (DBSCAN-Density Based Spatial Clustering of Applications with Noise) to successfully cluster points belonging to both large targets at close range and small targets at long range in the clustering stage (1300).

13. A radar system (3) according to claim 3, characterized by comprising the process stages of obtaining meaningful features with the outputs which are produced by the angle-of-arrival estimation algorithm, resulting from previous data in the feature extraction stage (1500).

Images