Improved radar object detection via dynamic and static Doppler spectrum splitting

Abstract

A system for improving object detection comprises at least one radar device configured to detect signals reflected from objects and a control module.The control module is configured to generate a radar spectrum based on the signals detected by the at least one radar device, divide the radar spectrum into static reflections and dynamic reflections that differ from the static reflections, extract features from the static reflections using a first machine learning module, extract features from the dynamic reflections using a second machine learning module that differs from the first machine learning module, merge the extracted features from the static reflections and the extracted features from the dynamic reflections, and detect static objects and dynamic objects based on the merged features from the static reflections and the dynamic reflections. Other example systems and methods for enhancing object detection are also disclosed.

Description

IntroductionThe information provided in this section is for the purpose of generally presenting the context of the disclosure. The work of the present inventors, insofar as described in this section, as well as aspects of the description that cannot be qualified as prior art at the time of filing, are neither expressly nor silently accepted as prior art against the present disclosure.The present disclosure relates to improved radar object detection using dynamic and static Doppler spectrum partitioning.A vehicle may include a driver assistance system that relies on blind spot detection, adaptive cruise control, lane departure warning, etc. sensors. In some cases, the sensors may include a radar device (e.g., a long range or short range radar device) to detect objects in the vicinity of the vehicle. Specifically, the radar device detects the objects by emitting waves (e.g., electromagnetic waves, radio waves, etc.) and then detecting the waves reflected from the objects.SummaryA vehicle system for improving object detection in a vehicle includes at least one radar device configured to detect signals reflected from objects in proximity to the vehicle and a control module in communication with the at least one radar device. The control module is configured to generate a radar spectrum based on the signals detected by the at least one radar device, partition the radar spectrum into static reflections and dynamic reflections that deviate from the static reflections, extract features from the static reflections with a first machine learning module, extract features from the dynamic reflections with a second machine learning module that is different from the first machine learning module, merge the extracted features from the static reflections and the extracted features from the dynamic reflections, and detect static objects and dynamic objects near the vehicle based on the merged features from the static reflections and the dynamic reflections.In other features, the first machine learning module and the second machine learning module are separate deep neural networks.In other features, the control module is configured to determine Doppler bins for the static reflections based on a speed of the vehicle.In other features, the control module is configured to separate the determined static reflection Doppler bins from the radar spectrum to split the radar spectrum into static reflections and dynamic reflections.In other features, the radar spectrum is a radar tensor formed from range values (range values), angle values, and Doppler values, and associates each Doppler bin with one of the range values and one of the angle values at the speed of the vehicle.In other features, the control module is configured to estimate the speed of the vehicle based on the data from the radar tensor.In other features, the vehicle system further comprises a sensor configured to detect the speed of the vehicle.In other features, the control module is configured to receive one or more signals from the sensor indicative of the speed of the vehicle.In other features, the control module is configured to generate a static feature map based on the features extracted from the static reflections and generate a dynamic feature map based on the features extracted from the dynamic reflections.In other features, the static feature map includes static feature vectors. Each static feature vector is associated with a trainable feature identified by the first machine learning module.In other features, the dynamic feature map includes dynamic feature vectors. Each dynamic feature vector is associated with a trainable feature identified by the second machine learning module.In other features, the control module is configured to concatenate the static feature map and the dynamic feature map together to merge the extracted features from the static reflections and the extracted features from the dynamic reflections.In other features, the vehicle system further includes a vehicle control module in communication with the control module. The vehicle control module is configured to receive one or more signals from the control module indicative of the detected static objects and dynamic objects.In other features, the vehicle control module is configured to control at least one vehicle control system based on the one or more signals.In other features, the vehicle control module is configured to generate a map including the detected static objects and dynamic objects proximate the vehicle based on the one or more signals.In other features, a vehicle includes the vehicle system.A method system for improving object detection includes generating a radar spectrum based on signals detected by at least one radar device, partitioning the radar spectrum into static reflections and dynamic reflections that deviate from the static reflections, extracting features from the static reflections with a first machine learning module, extracting features from the dynamic reflections with a second machine learning module that differs from the first machine learning module, merging the extracted features from the static reflections and the extracted features from the dynamic reflections, and detecting static objects and dynamic objects based on the merged features from the static reflections and the dynamic reflections.In other features, splitting the radar spectrum includes determining Doppler bins for the static reflections.In other features, dividing the radar spectrum includes separating the determined Doppler bins for the static reflections from the radar spectrum.In other features, the method further includes generating a static feature map based on the extracted features from the static reflections and generating a dynamic feature map based on the extracted features from the dynamic reflections.In other features, the static feature map includes static feature vectors. Each static feature vector is associated with a trainable feature identified by the first machine learning module.In other features, the dynamic feature map includes dynamic feature vectors. Each dynamic feature vector is associated with a trainable feature identified by the second machine learning module.In other features, merging the extracted features from the static reflections and the extracted features from the dynamic reflections includes concatenating the static feature map and the dynamic feature map.Further areas of applicability of the present disclosure will become apparent from the detailed description, claims and drawings. The detailed description and specific examples are for illustrative purposes only and are not intended to limit the scope of the disclosure.Brief Description of the DrawingsThe present disclosure will be more fully understood from the detailed description and the accompanying drawings, wherein: FIG. 1 is a functional block diagram of an example vehicle system for improving object detection according to the present disclosure; FIG. 2 is a vehicle including portions of the vehicle system of FIG. 1 in accordance with the present disclosure; FIG. 3 is a functional block diagram of an example control process for improving object detection according to the present disclosure; FIG. 4 is a flowchart of an example control process for improving object detection according to the present disclosure; and FIG. 5 is a graph showing improvements in object detection for a system utilizing features for improving object detection as compared to a conventional system.In the drawings, reference numerals may be reused to identify similar and / or identical elements.Detailed DescriptionA radar device emits waves and then detects the waves reflected from both static and dynamic objects. The radar device or a control module connected to the device then processes data from the reflected waves to obtain a radar spectrum for the objects. The radar spectrum may include data in the form of a three-dimensional radar tensor, where two dimensions include range values and angle values (e.g., range-azimuth space) and the third dimension includes Doppler values with velocity information (e.g., velocity of the objects relative to the radar device). In such examples, the radar tensor combines data for both static and dynamic objects. However, detecting objects by processing a radar tensor of both types often results in limited performance. For example, interference between static and dynamic objects may be induced due to a wide spread function of the radar device, and due to the different shapes and Doppler spreads of the static and dynamic objects, different processing approaches (e.g., filtering) may be required.The systems and methods and in accordance with the present disclosure provide technical solutions for improving radar object detection by partitioning or dividing a radar spectrum into various static and dynamic components, independently extracting features from each component with differently trained machine learning modules (e.g., deep neural networks, etc.), and then merging the extracted features from the static and dynamic components. By doing so, static and dynamic objects can be more accurately detected compared to conventional radar systems. For example, the systems and methods herein improve the accuracy of bounding boxes and velocity estimates for objects in radar applications, thereby providing improved results compared to conventional radar systems that process the entire radar tensor. In short, the systems and methods herein provide improved object detection at long distances and in adverse high angular resolution weather conditions, which in turn increases the coverage and travel speed of autonomous driving.Referring now to FIG. 1, a block diagram of an example vehicle system 100 for improving object detection is presented. As shown in FIG. 1, the vehicle system 100 generally includes a control module 102 and a plurality of radar devices 104, 106, 108 that communicate with the control module 102. Additionally, the vehicle system 100 may optionally include a speed sensor 110 in communication with the control module 102, a vehicle control module 112 in communication with the control module 102, and a display module 114 in communication with the vehicle control module 112 (and / or the control module 102). Although FIG. 1 illustrates the vehicle system 100 as including multiple separate modules, it should be understood that any combination of the modules (e.g., the control module 102, the vehicle control module 112, the display module 114, etc.) and / or functionality thereof may be incorporated into one or more modules.Although not shown in FIG. 1, the modules and sensors of the vehicle system 100 may share parameters via a network such as a controller area network (CAN). In such examples, the parameters may be shared across one or more data buses of the network. In fact, various parameters can be made available from a specific module and / or sensor to other modules and / or sensors via the network.In various embodiments, the vehicle system 100 of FIG. 1 may be deployed in any suitable vehicle, such as an electric vehicle (e.g., a pure electric vehicle, a plug-in hybrid electric vehicle, etc.), an internal combustion engine vehicle, etc. Moreover, the vehicle system 100 may be used in an autonomous vehicle, a semi-autonomous vehicle, etc. FIG. 2, for example, illustrates a vehicle 200 having the control module 102 and the vehicle control module 112 of FIG. 1, as well as one or more radar devices 204 (e.g., one or more of the radar devices 104, 106, 108 of FIG. 1, etc.) in communication with the control module 102. In such examples, each radar device 204 detects reflected signals from both static and dynamic objects in the vicinity of the vehicle 200. In other examples, the systems disclosed herein may be implemented in other suitable non-vehicular radar-incorporating applications that detect reflected signals from static and dynamic objects.With continued reference to FIG. 1, each radar device 104, 106, 108 may be any suitable radar device, such as a long range radar device, a short range radar device, etc. While the vehicle system 100 of FIG. 1 is illustrated and described as having three radar devices 104, 106, 108, it should be understood that the vehicle system 100 and / or other example systems may include more or fewer radar devices (e.g., one radar device, four radar devices, six radar devices, etc.), if desired. Regardless of the number of radar devices used, each radar device emits waves or signals and then detects waves or signals reflected from both static and dynamic objects, as is conventional. The radar devices 104, 106, 108 may then send radar data indicative of the reflected waves to the control module 102.In various embodiments, a radar spectrum is then generated on the basis of the signals detected by at least one of the radar devices 104, 106, 108. For example, the control module 102 may receive the radar data, process the data, and form the radar spectrum based on the radar data. In other examples, a control module associated with one or more of the radar devices 104, 106, 108 may process the radar data, form the radar spectrum based on the radar data, and send the radar spectrum to the control module 102. In such examples, the radar spectrum generated by the control module 102 or received by the control module 102 may be a radar tensor formed from range values, angle values, and Doppler values, as is conventional.FIG. 3, for example, illustrates an example control process 300 that may be used by the vehicle system 100 of FIG. 1 to improve object detection. In the example of FIG. 3, the process 300 may be implemented with one or more modules in the control module 102, as discussed further below. Although the example control process 300 is described with respect to the vehicle system 100 of FIG. 1 including the control module 102, the control process 300 may be used in another suitable system. In process 300, a three-dimensional radar tensor 302 is generated or received. In this example, the three-dimensional radar tensor 302 includes angle values (e.g., in the y-direction), range values (e.g., in the x-direction), and Doppler values (e.g., in the z-direction) for both static and dynamic objects.The control module 102 of FIG. 1 then partitions the radar spectrum into static reflections and dynamic reflections that deviate from the static reflections. In such examples, the control module 102 may first identify the reflection points for the static and dynamic reflections. In this way, the control module 102 avoids processing noise that may interfere with the analysis of the static and dynamic reflections. This may be accomplished by comparing detection points to a threshold (e.g., a value above a noise level to remove noise).In various embodiments, the control module 102 may split the radar spectrum based on a speed of the vehicle (e.g., the vehicle 200 in FIG. 2 ). In such examples, the speed of the vehicle may be received as input (e.g., a speed input 304 of FIG. 3 ) or otherwise obtained in various ways. For example, the control module 102 may receive one or more signals from the speed sensor 110 indicative of the speed of the vehicle. In such examples, the speed sensor 110 may be an inertial measurement unit (IMU) or other suitable sensor for generally detecting the speed of the vehicle.In other examples, the control module 102 may estimate the speed of the vehicle if, for example, the speed sensor 110 is not available. In such examples, the speed may be estimated based on data from the radar tensor. For example, the Doppler values may be used based on proximity to the speed of a vehicle. In some examples, the static reflection points of the radar tensor may be detected by clustering Doppler values of the detection points. In such examples, the largest cluster may be static reflection points. For example, the speed of a vehicle may be estimated from a static reflection detection point index (e.g., a cluster) and the estimated speed may be broken down into vectors (e.g., v x, v y) which are then used to split the radar spectrum into static reflections and dynamic reflections.As an example, Equations 1-3 below represent an approach to estimating the speed of a vehicle based on a cluster of Doppler values of detection points. In Equations 1-3, f represents an index of Doppler values f 0, f 1, f N, H represents a matrix of angle values θ, θ 01, θ N corresponding to the Doppler values f 0, f 1, f N, T represents a transposition operator on the matrix H, and v represents the estimated velocity based on the matrix H and a corresponding Doppler value.The control module 102 may then split the radar spectrum by determining or calculating Doppler bins for the static reflections based on the speed of the vehicle. In various embodiments, the control module 102 may include a Doppler bin calculation module 306 of FIG. 3 for such determinations. For example, the three-dimensional radar tensor 302 of FIG. 3 includes multiple cells (e.g., 100 range cells, 50 angle cells, 60 Doppler cells, etc.). In such examples, the Doppler bins represent an index of cells, for example, a first Doppler cell, a second Doppler cell, etc., for a particular range value and angle value. The control module 102 (or the Doppler bin calculation module 306 of FIG. 3 ) may calculate or determine each Doppler bin for one of the range values and one of the angle values at the speed (as provided by the speed input 304) of the vehicle.As an example, the Doppler bin calculation module 306 may implement Equation 4 below to determine a Doppler bin. In Equation 4, f represents a Doppler value, λ represents a wavelength of the transmitted radar signal, v x, v v represent vectors associated with the speed of the vehicle, and θ represents an angle value. In such examples, the angle value may be scanned for each region. For example, the angle may be sampled for one range value of -50 degrees to 50 degrees, for another range value of -50 degrees to 50 degrees, etc. Thereafter, the Doppler value may be determined based on the velocity vectors.With continued reference to FIGS. 1 and 3, the control module 102 may separate the determined static reflection Doppler bins from the radar spectrum to split the radar spectrum into static reflections and dynamic reflections. For example, in the process 300 of FIG. 3, a three-dimensional radar tensor 308 is shown that includes Doppler bins 310 for the static reflections. In such examples, the control module 102 may separate the determined static reflection Doppler bins from other Doppler bins based on the vehicle speed. For example, the control module 102 may determine a relative speed between the vehicle and a detected object. If the relative velocity suggests that the object is not moving, the control module 102 may determine that the object is static and classify the corresponding Doppler bins as bins for the static reflections.Once the static reflection Doppler bins are known, the control module 102 may then separate the determined static reflection Doppler bins to split the radar spectrum into static reflections and dynamic reflections, as shown in FIG. 3. For example, the process 300 of FIG. 3 is illustrated as splitting the radar tensor 302 into radar tensors 312, 314 having static and dynamic Doppler bins along different static and dynamic paths. In such examples, the Doppler bins for the static reflections (e.g., bins 310 of radar tensor 308) are identified as discussed above. Thereafter, all remaining bins not identified as static reflections are determined as dynamic bins for dynamic reflections. In FIG. 3, radar tensor 312 (e.g., the same as tensor 308) includes static Doppler bins 313 and radar tensor 314 includes dynamic Doppler bins 315. In this example, the dynamic Doppler bins 315 are the remaining bins in the radar tensor 314 that are not identified as the static Doppler bins (represented by the filled black bins).With continued reference to FIG. 1, the control module 102 may then independently extract features from the static reflections and features from the dynamic reflections with various machine learning modules 116, 118. In such examples, the machine learning modules 116, 18 may be separate deep neural networks (e.g., convolutional neutral networks, etc.) or other suitable machine learning modules trained on different data specific to dynamic and static characteristics of detected objects. For example, the machine learning modules 116, 118 may be trained (or retrained if necessary) based on one or more labeled data sets for known detected objects. Once the detection errors of the machine learning modules 116, 118 are satisfactory (e.g., above a threshold), the machine learning module 116 may be implemented by a static feature extraction module 316 of FIG. 3 to extract features from the static reflections, and the machine learning module 118 may be implemented by a dynamic feature extraction module 318 of FIG. 3 to extract features from the dynamic reflections.In various embodiments, different and / or independent filters may be used on the static and dynamic paths, if desired. Additionally, if desired, a fixed maximum of inputs may be set on the Doppler domain to reduce an input dimension.In some examples, the control module 102 of FIG. 1 may generate a static feature map based on the extracted features from the static reflections and a dynamic feature map based on the extracted features from the dynamic reflections. For example, and as shown in FIG. 3, a static feature map 320 may be generated based on the extracted features output by the static feature extraction module 316 (e.g., in the static path) and a dynamic feature map 322 may be generated based on the extracted features output by the dynamic feature extraction module 318 (e.g., in the dynamic path). In such examples, the static feature map 320 includes static feature vectors, while the dynamic feature map 322 includes dynamic feature vectors. Each static feature vector is associated with a range value of the radar tensor 312, an angle value of the radar tensor 312, and a trainable feature identified by the machine learning module 116. Similarly, each dynamic feature vector is associated with a range value of the radar tensor 314, an angle value of the radar tensor 314, and a trainable feature identified by the machine learning module 118.In various embodiments, the control module 102 merges the extracted features from the static reflections and the extracted features from the dynamic reflections. For example, a merging module 324 of FIG. 3 may receive the static feature map 320 and the dynamic feature map 322 and merge the feature maps. In such examples, the merging module 324 (e.g., a module in the control module 102) may merge the extracted features in any suitable manner. For example, the merging module 324 may concatenate the static feature map 320 and the dynamic feature map 322 to merge the extracted features from the static reflections and the extracted features from the dynamic reflections. In other examples, the merging module 324 may use another suitable technique to merge extracted features, such as summing the extracted features, etc.The control module 102 may then generate a new map of vectors based on the merged features. For example, the control module 102 may include a multi-layer perceptron module 326 of FIG. 3 that receives the merged feature maps 320, 322 and applies multiple layers of 1x1 folds to generate an output of the combined features. In such examples, the multi-layer perceptron module 326 may generate new vectors with dynamic and static vectors, respectively. In other examples, the multi-layer perceptron module 326 may use a linear layer to generate an output of the combined features.The control module 102 then detects static and dynamic objects in the vicinity of the vehicle based on the combined features from the static and dynamic reflections. For example, the control module 102 may process the output of the combined features to detect static and dynamic objects. In various embodiments, the control module 102 may use conventional techniques to detect the objects. For example, the control module 102 may include an object detection module 328 of FIG. 3 that receives the output of the multi-layer perceptron module 326, processes data in the outputs to detect objects, and then generates and outputs a bounding box for each detected object along with parameters (e.g., a position, a shape, an orientation, a speed, etc.) associated with each bounding box. In such examples, the object detection module 328 may implement an algorithm that scans the combined features of the multi-layer perceptron module 326 to detect the object.In various embodiments, the control module 102 of FIG. 1 may then generate and send one or more signals indicative of the detected static objects and / or dynamic objects to the vehicle control module 112. In such examples, the control module 102 may generate the signals based on the bounding box for each detected object. Upon receipt, the vehicle control module 112 may generate a control signal 120 for controlling one or more vehicle control systems based on the detected static objects and / or dynamic objects. For example, the vehicle control module 112 may use the signals to control driver assistance systems in an autonomous vehicle, a semi-autonomous vehicle, etc. In such examples, the driver assistance systems may include, for example, assisted acceleration / deceleration, assisted steering operation (e.g., assisted avoidance steering operation, etc.), adaptive cruise control, and / or any other driver assistance system that relies on the detection of static and / or dynamic objects in the environment of the vehicle.In some examples, the detected static objects and dynamic objects may be presented on a map in the vehicle. For example, the vehicle control module 112 may generate a map that includes the detected static and / or dynamic objects proximate the vehicle based on signals received from the control module 102. The vehicle control module 112 may then provide signals indicative of the map to the display module 114 (e.g., a display in the vehicle), which may then display the map to the driver and / or other passengers in the vehicle. In other examples, the display module 114 may receive the signals indicative of the detected static objects and / or dynamic objects from the control module 102, generate the map including the detected static and / or dynamic objects proximate the vehicle based on the received signals, and display the map.FIG. 4 illustrates another example control process 400 that may be used by the vehicle system 100 of FIG. 1 to improve object detection associated with a vehicle (e.g., the vehicle 200 of FIG. 2 ). Although the example control process 400 is described with respect to the vehicle system 100 of FIG. 1 including the control module 102, the control process 300 may be used by another suitable system.As shown, the control process 400 begins at 402, where the control module 102 determines whether any radar signals reflected non-objects are detected or otherwise received. If no, control may return to 402, as shown in FIG. 4. If so, control continues to 404.At 404, the control module 102 generates or receives a radar spectrum based on the reflected signals detected by one or more radar devices (e.g., the radar devices 104, 106, 108 of FIG. 1 ). For example, and as discussed above, the control module 102 may receive radar data, process the data, and generate the radar spectrum based on the radar data. In other examples, the radar device(s) may process the radar data, form the radar spectrum based on the radar data, and send the radar spectrum to the control module 102, as discussed above. In any case, the radar spectrum may be a radar tensor formed from range values, angle values and Doppler values. Control then continues to 406.At 406, the control module 102 determines Doppler bins for the static reflections based on a speed of the vehicle. In such examples, the speed may be a value sensed by a suitable sensor in the vehicle or an estimated value. Additionally, the control module 102 may implement equation 4 when determining the Doppler bins for the static reflections as discussed above. Control then continues to 408.At 408, the control module 102 partitions the radar spectrum into Doppler bins for the static reflections and Doppler bins for the dynamic reflections. For example, the control module 102 may separate the determined static reflection Doppler bins from the radar spectrum. The remaining Doppler bins not identified as static reflections may then be Doppler bins for the dynamic reflections. Control then continues to 410, 412.At 410, 412, the control module 102 independently extracts features from the static reflections and the dynamic reflections with different machine learning modules. More specifically, the control module 102 uses one trained machine learning module (e.g., a deep neural network or other suitable machine learning module) to extract features from the static reflections at 410 and another trained machine learning module (e.g., a deep neural network or other suitable machine learning module) to extract features from the dynamic reflections at 412. In such examples, each machine learning module may be trained on various tagged data sets for known detected static and dynamic objects, as discussed above. Control then continues to 414.At 414, the control module 102 merges the extracted features from the static reflections and the extracted features from the dynamic reflections. In such examples, the control module 102 may merge (e.g., concatenate, etc.) a static feature map and a dynamic feature map representing at least the extracted features, as discussed above. Control then continues to 416.At 416, the control module 102 determines whether any static and / or dynamic objects are detected based on the merged features. For example, and as discussed above, the control module 102 may process the combined features to detect static and dynamic objects. In such examples, a conventional algorithm may be used to scan the combined features to detect the object. If no object is detected, control may return to 402, as shown in FIG. 4. However, if an object is detected, control continues to 418.At 418, the control module 102 generates one or more signals indicative of the detected static and / or dynamic objects, as discussed above. In such examples, the control module 102 may send the signal(s) to a vehicle control module (e.g., the vehicle control module 112 of FIG. 1 ) as discussed above. Control continues to 420.At 420, a vehicle action is initiated based on the signal(s). For example, the vehicle control module may generate a control signal for controlling one or more vehicle control systems based on the detected static objects and / or dynamic objects, as discussed above. Additionally and / or alternatively, a display module (e.g., display module 114 of FIG. 1 ) may display a map that includes the detected static and / or dynamic objects relative to the vehicle, as discussed above. In such examples, the control module 102, the display module, and / or the vehicle control module may generate the map based on the signal(s). Control may then end as shown in FIG. 4 or return to another suitable step (e.g., 420) if desired.The systems and methods described herein improve radar object detection over conventional systems and methods. For example, FIG. 5 shows a graph 500 illustrating improvements in object detection for a system utilizing features to improve object detection as described herein over a conventional system. Specifically, in FIG. 5, the x-axis represents a rectal (RECTAL) of object detection, and represents the y-axis of an accuracy of object detection. The graph 500 includes a line 502 representing an object detection accuracy of a system herein that processes a radar tensor in different stacks in which static and dynamic objects are separated, and a line 504 representing an object detection accuracy of a conventional system that processes a radar tensor in a single stack in which static and dynamic objects are combined. In the example of FIG. 5, the area under the curve (AUC) for line 502 is 61.85% and the AUC for line 504 is about 52.94%, indicating that the system processing the radar tensor in different stacks with separate static and dynamic objects is more accurate than the conventional system.The foregoing description is merely illustrative in nature and is intended to limit the disclosure, its application, or uses. The broad teachings of the disclosure may be embodied in a variety of forms. Therefore, while this disclosure includes particular examples, the true scope of the disclosure should not be so limited since other modifications will become apparent upon a study of the drawings, the specification, and the following claims. It should be understood that one or more steps within a method may be performed in different orders (or concurrently) without altering the principles of the present disclosure. Further, although each of the embodiments is described above as having particular features, one or more of those features described with respect to any embodiment of the disclosure may be implemented in one of the other embodiments and / or combined with features of any of the other embodiments, even if that combination is not expressly described. In other words, the described embodiments are not mutually exclusive, and permutations of one or more embodiments with each other remain within the scope of this disclosure.Spatial and functional relationships between elements (e.g., between modules, circuit elements, semiconductor layers, etc.) are described using various terms, including "connected," "engaged," "coupled," "adjacent," "near," "on," "above," "below," and "disposed.". Unless expressly described as "direct", when a relationship between first and second elements is described in the above disclosure, this relationship may be a direct relationship in which no other intervening elements are present between the first and second elements, but may also be an indirect relationship in which one or more intervening elements (either spatially or functionally) are present between the first and second elements. As used herein, the phrase "at least one of A, B, and C" is intended to mean a logical (A OR B OR C) using a non-exclusive logical OR, and should not be understood to mean "at least one of A, at least one of B, and at least one of C.".In the figures, the direction of an arrow as indicated by the arrow head generally illustrates the flow of information (e.g., data or instructions) of interest for the illustration. For example, if element A and element B exchange a variety of information, but information transmitted from element A to element B is relevant for illustration, the arrow may point from element A to element B. This unidirectional arrow does not imply that no other information is transmitted from element B to element A. Moreover, for information transmitted from element A to element B, element B may transmit requests for, or acknowledgments of, the information to element A.In this application, including the definitions below, the term "module" or the term "controller" may be replaced by the term "circuit". "Module" may refer to a circuit, part of which, or may include an application specific integrated circuit (ASIC); a digital, analog, or mixed analog / digital discrete circuit; a digital, analog, or mixed analog / digital integrated circuit; a combinational logic circuit; a field programmable gate array (FPGA); a processor circuit (shared, dedicated, or group) that executes code; a memory circuit (shared, dedicated, or group) that stores code executed by the processor circuit; other suitable hardware components that provide the described functionality; or a combination of some or all of the above components, such as in a system-on-chip.The module may include one or more interface circuits. In some examples, the interface circuits may include wired or wireless interfaces connected to a local area network (LAN), the Internet, a wide area network (WAN), or combinations thereof. The functionality of any given module of the present disclosure may be distributed among multiple modules connected via interface circuits. For example, multiple modules may allow load balancing. In another example, a server module (also known as a remote or cloud module) may perform some functions for a client module.The term code as used above may include software, firmware, and / or microcode, and may refer to programs, routines, functions, classes, data structures, and / or objects. The term shared processor circuit includes a single processor circuit that executes some or all of the code from multiple modules. The term group processor circuit includes a processor circuit that, in combination with additional processor circuits, executes some or all of the code from one or more modules. References to multiple processor circuits include multiple processor circuits on single chips, multiple processor circuits on a single chip, multiple cores of a single processor circuit, multiple threads of a single processor circuit, or a combination of the foregoing. The term shared memory circuit includes a single memory circuit that stores some or all of the code from multiple modules. The term group memory circuit includes a memory circuit that, in combination with additional memories, stores some or all of the code from one or more modules.The term memory circuitry is a subset of the term computer readable medium. The term computer-readable medium as used herein does not include transitory electrical or electromagnetic signals propagating through a medium (such as on a carrier wave); the term computer-readable medium may therefore be considered tangible and non-transitory. Non-limiting examples of a non-transitory, tangible computer readable medium are nonvolatile memory circuits (such as a flash memory circuit, an erasable programmable read only memory circuit, or a mask read only memory circuit), volatile memory circuits (such as a static random access memory circuit or a dynamic random access memory circuit), magnetic storage media (such as an analog or digital magnetic tape or a hard disk drive), and optical storage media (such as a CD, a DVD, or a Blu-ray disk).The apparatuses and methods described in this application may be partially or fully implemented by a special purpose computer created by configuring a general purpose computer to perform one or more special functions embodied in computer programs. The above-described functional blocks, components of flowcharts, and other elements serve as software specifications that can be translated into the computer programs by the routine work of a person skilled in the art or programmer.The computer programs include processor-executable instructions stored on at least one non-transitory, tangible computer-readable medium. The computer programs may also contain or rely on stored data. The computer programs may include a basic input / output system (BIOS) that interacts with the special purpose computer hardware, device drivers that interact with particular special purpose computer devices, one or more operating systems, user applications, background services, background applications, etc.The computer programs may include: (i) a descriptive text to be analyzed, such as hypertext markup language (HTML), extensible markup language (XML), or javascript object notation (JSON), (ii) assembler code, (iii) object code generated from the source code by a compiler, (iv) source code for execution by an interpreter, (v) source code for compilation and execution by a just-in-time compiler, etc. As examples only, source code may be written using syntax of languages that include C, C++, C#, ObjectiveC-, Swift, Haskell, Go, SQL, R, Lisp, Java® Fortran, Perl, Pascal, Curl, OCaml, JavaScript® HTML5 (Hypertext Markup Language 5th Revision), Ada, ASP (Active Server Pages), PHP (PHP: Hypertext Preprocessor), Scala, Eiffel, Smalltalk, Erlang, Ruby, Flash® Visual Basic® Lua, MATLAB, SIMULINK, and Python® may be used.

Claims

A vehicle system for improving object detection in a vehicle, the vehicle system comprising: at least one radar device configured to detect signals reflected from objects in the vicinity of the vehicle; and a control module in communication with the at least one radar device, the control module configured to: generate a radar spectrum based on the signals detected by the at least one radar device; split the radar spectrum into static reflections and dynamic reflections that deviate from the static reflections; extract features from the static reflections with a first machine learning module; extract features from the dynamic reflections with a second machine learning module that differs from the first machine learning module; merge the extracted features from the static reflections and the extracted features from the dynamic reflections; and detecting static objects and dynamic objects in the vicinity of the vehicle based on the merged features from the static and dynamic reflections.The vehicle system of claim 1, wherein the control module is configured to determine Doppler bins for the static reflections based on a speed of the vehicle.The vehicle system of claim 2, wherein the control module is configured to separate the determined static reflection Doppler bins from the radar spectrum to split the radar spectrum into static reflections and dynamic reflections.The vehicle system of claim 2, wherein: the radar spectrum is a radar tensor formed from range values, angle values, and Doppler values; and each Doppler bin associates one of the range values and one of the angle values at the speed of the vehicle.The vehicle system of claim 1, wherein the control module is configured to: generate a static feature map based on the extracted features from the static reflections; and generate a dynamic feature map based on the features extracted from the dynamic reflections.The vehicle system of claim 5, wherein: the static feature map includes static feature vectors, each static feature vector being associated with a trainable feature identified by the first machine learning module; and the dynamic feature map includes dynamic feature vectors, each dynamic feature vector being associated with a trainable feature identified by the second machine learning module.The vehicle system of claim 5, wherein the control module is configured to concatenate the static feature map and the dynamic feature map to merge the extracted features from the static reflections and the extracted features from the dynamic reflections.The vehicle system of claim 1, further comprising a vehicle control module in communication with the control module, the vehicle control module configured to receive one or more signals from the control module indicative of the detected static objects and dynamic objects.The vehicle system of claim 8, wherein the vehicle control module is configured to control at least one vehicle control system based on the one or more signals.The vehicle system of claim 8, wherein the vehicle control module is configured to generate a map including the detected static objects and dynamic objects in the vicinity of the vehicle based on the one or more signals.

Images