Detection and classification of tunnels using light detection and ranging

Abstract

A driver assistance system includes a light detection and ranging (LiDAR) sensor. The detection module is configured to detect a tunnel in a path of the vehicle. The detection module includes a zoning module configured to bind echoes from the LiDAR sensor into a plurality of zones. The clustering and feature extraction module is configured to: identify clusters in a region; determining centers and variances of clusters in the x-axis direction, the y-axis direction and the z-axis direction in the plurality of regions; identifying a plurality of features based on the center and variance of the cluster; and concatenating the plurality of features into one or more concatenated features. The classification module is configured to receive the one or more concatenation features and declare at least one of a proximity state, an internal state, and an unobstructed state of the tunnel in response to the one or more concatenation features.

Description

Technical Field

[0001] The information provided in this section is for the purpose of generally presenting the background of this disclosure. The work of the currently named inventors is acknowledged, neither explicitly nor implicitly, as prior art to this disclosure, to the extent described in this section and in respect of aspects that may not otherwise qualify as prior art at the time of filing.

[0002] This disclosure relates to driver assistance systems, and more specifically to driver assistance systems including light detection and ranging (LiDAR) sensors. Background Technology

[0003] Vehicles with various levels of driver assistance (such as fully or partially autonomous vehicles) typically rely on radio detection and ranging (radar) systems to detect and avoid objects in the vehicle's path. Radar-based systems encounter problems detecting objects when the vehicle travels through tunnels (or other similar infrastructure, such as under overpasses). For example, when moving through a tunnel, radar-based systems detect phantom objects and / or erroneous trajectories caused by the tunnel infrastructure. Summary of the Invention

[0004] A driver assistance system for a vehicle includes a LiDAR (Light Detection and Ranging) sensor configured to emit light pulses and receive echoes. A detection module is configured to detect tunnels in the vehicle's path. The detection module includes a partitioning module configured to bin the echoes from the LiDAR sensor into multiple partitions. The detection module also includes a clustering and feature extraction module configured to: identify clusters within the partitions; determine the center and variance of the clusters in the x-axis, y-axis, and z-axis directions of the multiple partitions; identify multiple features based on the center and variance of the clusters; and concatenate the multiple features into one or more concatenated features. A classification module is configured to receive one or more concatenated features and, in response to the one or more concatenated features, declare at least one of a tunnel's proximity state, interior state, and unobstructed state.

[0005] Among other features, multiple zones are included, specifically a first zone and a second zone. The first zone corresponds to echoes with values ​​in the z-axis direction from the LiDAR sensor that are less than a predetermined height. The second zone corresponds to echoes with values ​​in the z-axis direction that are greater than a predetermined height. The driver assistance system includes a Global Positioning System (GPS). The LiDAR sensor converts the echoes into a Frenet frame in response to data from the GPS system.

[0006] Among other features, the inertial measurement system is configured to detect vehicle pitch, with LiDAR sensors responding to vehicle pitch to compensate for echoes. The classifier module includes a pre-trained model configured to detect proximity, interior, and obstacle-free states in response to one or more cascaded features. The filter module is configured to filter the output of the classifier module using a Hidden Markov Model (HMM). The HMM filters out infeasible state transitions.

[0007] Among other features, the pre-trained model is configured to: detect a near state in response to the variance of the cluster in the x-axis direction of the second region being less than the first variance and the variance of the cluster in the z-axis direction of the second region being greater than the second variance; and detect an internal state in response to the variance of the cluster in the x-axis direction of the second region being greater than the third variance and the variance of the cluster in the z-axis direction of the second region being less than the fourth variance, wherein the first variance is less than the third variance and the third variance is greater than the fourth variance.

[0008] Among other features, the classifier module is configured to detect walls in the vehicle's path in response to the variance of the clusters in the x-axis and z-axis directions.

[0009] A method for assisting a driver of a vehicle includes: emitting light pulses and receiving echoes using a light detection and ranging (LiDAR) sensor; binning the echoes from the LiDAR sensor into multiple zones; identifying clusters within the zones; determining the center and variance of clusters in the x-axis, y-axis, and z-axis directions of the multiple zones; identifying multiple features based on the center and variance; concatenating the multiple features into one or more concatenated features; and detecting at least one of an approach state, an interior state, and an unobstructed state of a tunnel in response to the one or more concatenated features.

[0010] Among other features, multiple zones include a first zone and a second zone. The first zone corresponds to echoes from the LiDAR sensor with values ​​in the z-axis direction that are less than a predetermined height, and the second zone corresponds to echoes with values ​​in the z-axis direction that are greater than a predetermined height.

[0011] Among other features, the method includes converting the echoes into a Frenet frame. The method includes: detecting the vehicle's pitch; and compensating for the echoes from the LiDAR sensor in response to the vehicle's pitch.

[0012] Among other features, the method includes using a pre-trained model to detect proximity states, internal states, and barrier-free states in response to one or more concatenated features. The method also includes filtering the output of the pre-trained model using a Hidden Markov Model (HMM). The HMM filters out infeasible state transitions.

[0013] Among other features, the pre-trained model is configured to: detect a near state in response to the variance of the cluster in the x-axis direction of the second region being less than the first variance and the variance of the cluster in the z-axis direction of the second region being greater than the second variance; and detect an internal state in response to the variance of the cluster in the x-axis direction of the second region being greater than the third variance and the variance of the cluster in the z-axis direction of the second region being less than the fourth variance, wherein the first variance is less than the third variance and the third variance is greater than the fourth variance.

[0014] Among other features, the pre-trained model is configured to detect walls in the vehicle's path in response to variances in the x-axis and z-axis directions.

[0015] Other applicable areas of this disclosure will become apparent from the detailed description, claims, and drawings. The detailed description and specific examples are intended for illustrative purposes only and are not intended to limit the scope of this disclosure.

[0016] This invention also includes the following technical solutions:

[0017] 1. A driver assistance system for a vehicle, comprising:

[0018] A light detection and ranging (LiDAR) sensor, the light detection and ranging (LiDAR) sensor being configured to emit light pulses and receive echoes; and

[0019] A detection module, configured to detect tunnels in the vehicle's path, includes:

[0020] A partitioning module, configured to divide the echoes from the LiDAR sensor into multiple zones;

[0021] Clustering and feature extraction module, wherein the clustering and feature extraction module is configured to:

[0022] Identify clusters in the region.

[0023] Determine the center and variance of the cluster in the x-axis, y-axis, and z-axis directions of the plurality of regions.

[0024] Multiple features are identified based on the cluster's centroid and variance, and

[0025] The plurality of features are concatenated into one or more concatenated features; and

[0026] A classification module is configured to receive one or more concatenated features and, in response to the one or more concatenated features, declare at least one of the following: proximity state, interior state, and unobstructed state of the tunnel.

[0027] 2. The driver assistance system according to technical solution 1, wherein the plurality of zones includes a first zone and a second zone.

[0028] 3. The driver assistance system according to technical solution 2, wherein the first region corresponds to an echo with a value in the z-axis direction from the LiDAR sensor that is less than a predetermined height, and the second region corresponds to an echo with a value in the z-axis direction that is greater than the predetermined height.

[0029] 4. The driver assistance system according to technical solution 1 further includes a global positioning system (GPS), wherein the LiDAR sensor converts the echo into a Frenet frame in response to data from the GPS system.

[0030] 5. The driver assistance system according to technical solution 4 further includes an inertial measurement system configured to detect the pitch of the vehicle, wherein the LiDAR sensor compensates for the echo in response to the pitch of the vehicle.

[0031] 6. The driver assistance system according to technical solution 3, wherein the classifier module includes a pre-trained model configured to detect the proximity state, the interior state, and the obstacle-free state in response to the one or more concatenated features.

[0032] 7. The driver assistance system according to technical solution 6 further includes a filter module, which is configured to filter the output of the classifier module using a hidden Markov model.

[0033] 8. The driver assistance system according to technical solution 7, wherein the hidden Markov model filters out infeasible state transitions.

[0034] 9. The driver assistance system according to technical solution 6, wherein the pre-trained model is configured as follows:

[0035] The proximity state is detected in response to the variance of the clusters in the x-axis direction of the second region being less than the first variance and the variance of the clusters in the z-axis direction of the second region being greater than the second variance; and

[0036] The internal state is detected in response to the fact that the variance of the cluster in the x-axis direction of the second region is greater than the third variance and the variance of the cluster in the z-axis direction of the second region is less than the fourth variance, wherein the first variance is less than the third variance and the third variance is greater than the fourth variance.

[0037] 10. The driver assistance system according to technical solution 1, wherein the classifier module is configured to detect walls in the path of the vehicle in response to the variance of clusters in the x-axis and z-axis directions.

[0038] 11. A method for assisting a driver of a vehicle, comprising:

[0039] Use a light detection and ranging (LiDAR) sensor to emit light pulses and receive the echoes;

[0040] The echoes from the LiDAR sensor are divided into multiple zones;

[0041] Identify clusters within the region;

[0042] Determine the center and variance of the cluster in the x-axis, y-axis and z-axis directions of the plurality of regions;

[0043] Multiple features are identified based on the center and variance;

[0044] The plurality of features are concatenated into one or more concatenated features; and

[0045] In response to the one or more concatenated features, at least one of the proximity state, interior state, and unobstructed state of the tunnel is detected.

[0046] 12. The method according to technical solution 11, wherein the plurality of zones includes a first zone and a second zone.

[0047] 13. The method according to technical solution 12, wherein the first region corresponds to an echo with a value in the z-axis direction from the LiDAR sensor that is less than a predetermined height, and the second region corresponds to an echo with a value in the z-axis direction that is greater than the predetermined height.

[0048] 14. The method according to technical solution 11 further includes converting the echo into a Frenet frame.

[0049] 15. The method according to technical solution 14 further includes:

[0050] Detect the vehicle's pitch; and

[0051] The system compensates for echoes from the LiDAR sensor in response to the vehicle's pitch.

[0052] 16. The method according to technical solution 13 further includes using a pre-trained model in response to the one or more concatenated features to detect the proximity state, the interior state, and the barrier-free state.

[0053] 17. The method according to technical solution 16 further includes filtering the output of the pre-trained model using a hidden Markov model.

[0054] 18. The method according to technical solution 17, wherein the hidden Markov model filters out infeasible state transitions.

[0055] 19. The method according to technical solution 16, wherein the pre-trained model is configured as follows:

[0056] The proximity state is detected in response to the variance of the clusters in the x-axis direction of the second region being less than the first variance and the variance of the clusters in the z-axis direction of the second region being greater than the second variance; and

[0057] The internal state is detected in response to the fact that the variance of the cluster in the x-axis direction of the second region is greater than the third variance and the variance of the cluster in the z-axis direction of the second region is less than the fourth variance, wherein the first variance is less than the third variance and the third variance is greater than the fourth variance.

[0058] 20. The method according to technical solution 16, wherein the pre-trained model is configured to detect walls in the path of the vehicle in response to the variance in the x-axis direction and the z-axis direction. Attached Figure Description

[0059] This disclosure will be more fully understood in light of the specific embodiments and accompanying drawings, wherein:

[0060] Figure 1 This is a functional block diagram of a vehicle according to the present disclosure, including a driver assistance controller, a tunnel detection system, an autonomous driving module, a global positioning system (GPS), and a light detection and ranging (LiDAR) sensor;

[0061] Figure 2 The diagram shows the orientation of the vehicle relative to the coordinate system;

[0062] Figures 3A to 3C The illustrations respectively show LIDAR echo data of a vehicle traveling toward a tunnel opening, a vehicle in the tunnel, and a vehicle traveling toward a wall, according to this disclosure.

[0063] Figure 4 The diagram illustrates the states and transitions of a Hidden Markov Model used to reduce noise and eliminate infeasible state transitions; and

[0064] Figure 5A and Figure 5B This is a flowchart of a method for identifying tunnels using a LIDAR sensor, according to the present disclosure.

[0065] In the accompanying drawings, reference numerals may be used repeatedly to identify similar and / or identical elements. Detailed Implementation

[0066] To mitigate these errors, some radar-based systems attempt to detect tunnels before entering them. However, due to sensor limitations, radar-based systems cannot accurately detect tunnels. Even without infrastructure / tunnels, radar-based systems experience numerous false detections and miss a significant proportion of targets while traveling through tunnels.

[0067] The tunnel detection system according to this disclosure uses echo data from a Light Detection and Ranging (LiDAR) sensor to detect tunnels for the driver assistance controller. Compared to typical radar systems, LiDAR sensors offer higher resolution, higher precision, and greater accuracy. Utilizing LiDAR sensors to detect tunnels more accurately reduces erroneous braking caused by false radar echoes.

[0068] Now for reference Figures 1 to 3C This illustrates the operation of the tunnel detection module 112. Figure 1 In this system, vehicle 100 includes a driver assistance controller 110, which includes a tunnel detection module 112 and an autonomous driving module 190 supporting full or partial autonomous driving levels. A global positioning system 120 determines the vehicle's position and outputs the vehicle position and steering path to the driver assistance controller 110. A radar system 122 optionally generates radio frequency (RF) pulses and outputs radar echo signals to the driver assistance controller 110.

[0069] LiDAR sensor 124 generates light pulses and outputs echo signals to driver assistance controller 110. In some examples, LiDAR sensor 124 includes one or more lasers 130. In some examples, LiDAR sensor 124 includes one or more scanners 128 that scan the one or more lasers 130 in the vehicle's steering path or field of view.

[0070] Inertial Measurement Unit (IMU) 134 generates yaw and pitch data for the vehicle and outputs the yaw and pitch data to Driver Assistance Controller 110. Echo data from LiDAR sensor 124 is converted into a Frenet frame using data from GPS 120 (and / or vehicle pitch and / or yaw from IMU 134). The echoes are stored in a point cloud, which is converted into self-motion in the Frenet frame using the center of the road (represented by distance and vertical offset along the road center) to account for the vehicle's steering path and improve accuracy.

[0071] The tunnel detection module 112 is configured to detect when the edge or entrance of the tunnel is in the vehicle's path (corresponding to the tunnel approach state). The tunnel detection module 112 is also configured to detect when the vehicle is in the tunnel (corresponding to the interior state), the walls in the vehicle's path, and / or when the vehicle leaves the tunnel (corresponding to the unobstructed state). The tunnel detection module 112 includes a LiDAR echo data storage device 140 to store echo points (e.g., LiDAR point cloud data) from the LiDAR sensor 124.

[0072] The tunnel detection module 112 includes a partitioning module 142, configured to bin echoes from the LiDAR sensor into multiple spatial regions or zones. A clustering and feature extraction module 144 is configured to identify clusters (or echo groups) within the multiple zones. The clustering and feature extraction module 144 is configured to determine the center and variance of clusters in the x-axis, y-axis, and z-axis directions within the multiple zones. The clustering and feature extraction module 144 is configured to identify multiple features based on the center and variance. The clustering and feature extraction module 144 is configured to concatenate multiple features into one or more concatenated features.

[0073] In some examples, multiple zones are included, comprising a first zone and a second zone. The first zone corresponds to echoes with values ​​in the z-axis direction less than a predetermined height. The second zone corresponds to echoes with values ​​in the z-axis direction greater than a predetermined height. In some examples, the IMU 134 output is used to correct for vehicle pitch and / or yaw variations and improve accuracy. Using data from the IMU 134 allows for the detection of tunnels with sloping entrances.

[0074] In some examples, the clustering and feature extraction module 144 outputs one or more concatenated features to the classifier module 156. The classifier module 156 uses a pre-trained model to determine whether one or more concatenated features correspond to the edge of the tunnel (e.g., representing an approach state), the vehicle being inside the tunnel (inside state), the walls of the tunnel (and in the vehicle's path), or leaving the tunnel (or an unobstructed state).

[0075] Filter module 158 receives the classifier's output and uses a Hidden Markov Model (HMM) with predefined states and transitions. The HMM reduces noise by eliminating infeasible state transitions. For example, each state can only have some transitions to other states. For instance, an unobstructed state cannot be followed by an inner state. In some examples, an unobstructed state can be followed by another unobstructed state or a nearby state, a nearby state can be followed by another nearby state or an inner state, and an inner state can be followed by another inner state or an unobstructed state.

[0076] In some examples, the driver assistance controller 110 includes an autonomous driving module 190, which is configured to control one or more vehicle control inputs 192, such as steering wheel or steering input, accelerator pedal or propulsion input, vehicle speed, brake pedal or brake input, steering signals, and so on.

[0077] exist Figure 2 The vehicle is shown relative to the x-axis (e.g., the direction of forward movement), y-axis (e.g., the direction of lateral movement), and z-axis (e.g., the height direction).

[0078] exist Figure 3A In this context, two or more zones include a first zone (zone 1) and a second zone (zone 2), where the first zone (zone 1) includes z-axis data within a first predetermined range (e.g., Figure 2 The second zone (zone 2) includes z-axis data within a second predetermined range. In some examples, the first zone includes z-axis values ​​within a range of 0 to 10 feet, and the second zone includes z-axis values ​​within a range of 10 to 18 feet. In other examples, the first zone includes z-axis values ​​within a range of 0 to 14 feet, and the second zone includes z-axis values ​​within a range of 14 to 18 feet.

[0079] The tunnel detection module 112 includes a clustering and feature extraction module 144, which analyzes the data in the region. The clustering and feature extraction module 144 identifies clusters of LiDAR echoes within each region. For each echo in the region, the clustering and feature extraction module 144 identifies the center of the cluster and calculates the variance within the cluster in the x, y, and z directions. Figure 2 ).

[0080] The tunnel detection module 112 includes a concatenation module 152, which concatenates features and outputs the concatenated features to a classifier module 156. In some examples, the classifier module 156 includes a pre-trained model that analyzes and classifies the concatenated features output by the concatenation module 152.

[0081] For example, tunnel detection module 112 uses partitions or bins corresponding to regions where tunnel features are typically found to extract features. Binning reduces the amount of computation required. For example, the data is divided into two distinct spatial regions (e.g., the z-axis is divided into two zones (e.g., corresponding to 0 to 10 feet and 10 to 18 feet). After identifying clusters within the regions, the variances in the x, y, and z directions within each cluster are classified.

[0082] exist Figure 3AIn this context, the edge or entrance 210 of tunnel 211 corresponds to the first category of classifier module 156. Classifier module 156 searches for statistical features corresponding to the edge or entrance 210 of tunnel 211. In some examples, classifier module 156 searches for a second zone (zone 2) corresponding to the range between 10 and 18 feet. Classifier module 156 identifies an edge or entrance 210 when the variance of echo points 212 corresponds to a low distribution along the x-axis and a high distribution along the z-axis. In some examples, when an edge or entrance is identified, tunnel detection module 112 sets a first flag.

[0083] exist Figure 3B In this context, the second classification corresponds to the vehicle being located inside tunnel 211. The second classification of classifier module 156 corresponds to echo points 212 having a low distribution on the z-axis and a high distribution on the x-axis (corresponding to echoes from the tunnel ceiling). In some examples, tunnel detection module 112 sets a second flag when the vehicle detects a tunnel within its turning path.

[0084] exist Figure 3C In this context, the third category of classifier module 156 corresponds to the situation where the vehicle encounters wall 230 in its path or field of view. After the first category identifies the edge or entrance of the tunnel, an attempt to identify the third category can begin. The third category corresponds to a low distribution on the x-axis and a high distribution on the z-axis. In some examples, the third category includes two zones (e.g., a distribution range from 0 to 18 feet).

[0085] The fourth category of classifier module 156 corresponds to the unobstructed state when there are significantly fewer or no echoes in the vehicle's path or field of view.

[0086] In some examples, the tunnel detection module 112 uses a Hidden Markov Model (HMM). The HMM includes an accessible state, an approach state, and an interior state. In some examples, the state transitions of the HMM are limited to accessible → approach → interior → accessible.

[0087] exist Figure 4 In the Hidden Markov Model, the accessibility state 410 indicates the absence of a tunnel (tunnel detection flag not yet set). When the tunnel detection module 112 detects the edge or entrance 210 of tunnel 211, the proximity state 414 begins. After declaring proximity state 414, the tunnel detection module 112 determines whether a vehicle is inside tunnel 211, and if true, transitions to interior state 418. After transitioning to interior state 418, if there is no echo within the field of view, the tunnel detection module 112 transitions to accessibility state 410.

[0088] Now for reference Figure 5AThe diagram illustrates a method for detecting tunnels. At 310, a LiDAR sensor generates light pulses and receives echo points. The LiDAR sensor also receives GPS data and locates echo points in the vehicle's path or field of view. At 314, the echo points or echoes are collected into a box or zone. At 318, features are extracted from the zone. At 322, the features are concatenated. At 326, a classifier uses a model (such as a pre-trained machine learning model) to identify when a vehicle is approaching, in, or leaving the tunnel.

[0089] Now for reference Figure 5B The diagram illustrates a method for extracting features from regions. At 360, clusters are identified within each region. At 364, the center and variance of the clusters are determined in the x, y, and z directions.

[0090] The foregoing description is illustrative in nature and is not intended to limit this disclosure, its application, or use. The broad teachings of this disclosure can be implemented in many forms. Therefore, while this disclosure includes specific examples, its true scope should not be so limited, as other modifications will become apparent from a study of the accompanying drawings, specification, and the following claims. It should be understood that one or more steps within the method may be performed in a different order (or simultaneously) without altering the principles of this disclosure. Furthermore, while each embodiment has been described above as having certain features, any one or more of those features described with respect to any embodiment of this disclosure may be implemented in and / or combined with features of any other embodiment, even if such combination is not explicitly described. In other words, the described embodiments are not mutually exclusive, and the arrangement of one or more embodiments with respect to each other remains within the scope of this disclosure.

[0091] Spatial and functional relationships between components (e.g., between modules, between circuit elements, between semiconductor layers, etc.) are described using various terms, including “connected,” “joined,” “linked,” “adjacent,” “next to,” “on top of,” “above,” “below,” and “set.” Unless explicitly described as “direct,” when the relationship between the first and second components is described in the foregoing disclosure, the relationship can be a direct relationship in which no other intermediate components exist between the first and second components, or an indirect relationship in which one or more intermediate components (spatial or functional) exist between the first and second components. As used herein, at least one of the phrases A, B, and C should be interpreted as representing a logic (A or B or C) using the non-exclusive logic “OR,” and should not be interpreted as representing “at least one A, at least one B, and at least one C.”

[0092] In a diagram, as indicated by the arrows, the direction of the arrows typically shows the flow of information (such as data or instructions) of interest to the diagram. For example, when components A and B exchange various types of information, but the information passed from component A to component B is relevant to the diagram, the arrow might point from component A to component B. This unidirectional arrow does not mean that no other information is passed from component B to component A. Furthermore, for information sent from component A to component B, component B may send a request for or confirmation of receipt of that information to component A.

[0093] In this application, the terms "module" or "controller" are used in accordance with the following definitions and may be replaced by the term "circuit". The term "module" may refer to, belong to, or include: application-specific integrated circuit (ASIC); digital, analog, or mixed-signal analog / digital discrete circuit; digital, analog, or mixed-signal analog / digital integrated circuit; combinational logic circuit; field-programmable gate array (FPGA); processor circuitry (shared, dedicated, or grouped) that executes code; memory circuitry (shared, dedicated, or grouped) that stores code executed by the processor circuitry; other suitable hardware components that provide the described functionality; or combinations of some or all of the above, such as in a system-on-a-chip.

[0094] A module may include one or more interface circuits. In some examples, the interface circuit may include a wired or wireless interface connected to a local area network (LAN), the Internet, a wide area network (WAN), or a combination thereof. The functionality of any given module disclosed herein may be distributed across multiple modules connected via the interface circuit. For example, multiple modules may allow for load balancing. In another example, a server (also known as a remote or cloud) module may perform certain functions on behalf of a client module.

[0095] The term "code" as used above can include software, firmware, and / or microcode, and can 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 conjunction with additional processor circuitry, executes some or all of the code from one or more modules. References to multiple processor circuits include multiple processor circuits on a discrete chip, multiple processor circuits on a single chip, multiple cores of a single processor circuit, multiple threads of a single processor circuit, or combinations thereof. 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 conjunction with additional memory, stores some or all of the code from one or more modules.

[0096] The term memory circuit is a subset of the term computer-readable medium. As used herein, the term computer-readable medium does not include transient electrical or electromagnetic signals propagating through a medium (such as on a carrier wave); therefore, the term computer-readable medium can be considered tangible and non-transient. Non-limiting examples of non-transient tangible computer-readable media are non-volatile memory circuits (such as flash memory circuits, erasable programmable read-only memory circuits, or mask read-only memory circuits), volatile memory circuits (such as static random access memory circuits or dynamic random access memory circuits), magnetic storage media (such as analog or digital magnetic tape or hard disk drives), and optical storage media (such as CDs, DVDs, or Blu-ray discs).

[0097] The apparatus and methods described in this application can be implemented, in part or in whole, by a special-purpose computer, which is created by configuring a general-purpose computer to perform one or more specific functions contained in a computer program. The function blocks, flowchart components, and other elements described above serve as software specifications that can be translated into computer programs through the daily work of a skilled technician or programmer.

[0098] A computer program includes processor-executable instructions stored on at least one non-transitory tangible computer-readable medium. A computer program may also include or depend on stored data. A computer program may include a basic input / output system (BIOS) for interacting with the hardware of a special-purpose computer, device drivers for interacting with specific devices of the special-purpose computer, one or more operating systems, user applications, background services, background applications, etc.

[0099] Computer programs may include: (i) descriptive text to be parsed, such as HTML (Hypertext Markup Language), XML (Extensible Markup Language), or JSON (JavaScript Object Notation); (ii) assembly code; (iii) object code generated from source code by a compiler; (iv) source code executed by an interpreter; and (v) source code compiled and executed by a just-in-time (JIT) compiler, etc. As an example only, source code may be written using the syntax of languages ​​including C, C++, C#, Objective C, Swift, Haskell, Go, SQL, R, and Lisp. Fortran, Perl, Pascal, Curl, OCaml, HTML5 (Hypertext Markup Language 5th Edition), Ada, ASP (Active Server Pages), PHP (PHP: Hypertext Preprocessor), Scala, Eiffel, Smalltalk, Erlang, Ruby, Visual Lua, MATLAB, SIMULINK and

Claims

1. A driver assistance system for a vehicle, comprising: A light detection and ranging (LiDAR) sensor, which is configured to emit light pulses and receive echoes; as well as A detection module, configured to detect tunnels in the vehicle's path, includes: A partitioning module, configured to divide the echoes from the LiDAR sensor into multiple zones; Clustering and feature extraction module, wherein the clustering and feature extraction module is configured to: Identify clusters in the region. Determine the center and variance of the cluster in the x-axis, y-axis, and z-axis directions of the plurality of regions. Multiple features are identified based on the cluster's centroid and variance, and The plurality of features are concatenated into one or more concatenated features; and A classification module is configured to receive one or more concatenated features and, in response to the one or more concatenated features, declare at least one of the following: proximity state, interior state, and unobstructed state of the tunnel.

2. The driver assistance system according to claim 1, wherein, The multiple zones include Zone 1 and Zone 2.

3. The driver assistance system according to claim 2, wherein, The first region corresponds to echoes from the LiDAR sensor whose z-axis value is less than a predetermined height, and the second region corresponds to echoes whose z-axis value is greater than the predetermined height.

4. The driver assistance system according to claim 1, further comprising a Global Positioning System (GPS), wherein, The LiDAR sensor converts the echo into a Frenet frame in response to data from the GPS system.

5. The driver assistance system of claim 4, further comprising an inertial measurement system configured to detect the pitch of the vehicle, wherein, The LiDAR sensor compensates for the echo in response to the vehicle's tilt.

6. The driver assistance system according to claim 3, wherein, The classifier module includes a pre-trained model configured to detect the proximity state, the interior state, and the accessibility state in response to one or more concatenated features.

7. The driver assistance system of claim 6, further comprising a filter module configured to filter the output of the classifier module using a hidden Markov model.

8. The driver assistance system according to claim 7, wherein, The hidden Markov model filters out infeasible state transitions.

9. The driver assistance system according to claim 6, wherein, The pre-trained model is configured as follows: The proximity state is detected in response to the fact that the variance of the cluster in the x-axis direction in the second region is less than the first variance and the variance of the cluster in the z-axis direction in the second region is greater than the second variance; as well as The internal state is detected in response to the fact that the variance of the cluster in the x-axis direction of the second region is greater than the third variance and the variance of the cluster in the z-axis direction of the second region is less than the fourth variance, wherein the first variance is less than the third variance and the third variance is greater than the fourth variance.

10. The driver assistance system according to claim 1, wherein, The classifier module is configured to detect walls in the vehicle's path in response to the variance of the clusters in the x-axis and z-axis directions.

Images