Traffic light state estimation method and system for vehicle, computing device and vehicle
Through ultra-wideband positioning technology, and combined with machine learning methods, the accuracy and reliability of traffic light status information acquisition in the existing technology are solved, and high-precision traffic light status estimation under occlusion or harsh conditions is achieved, reducing the Computational resource consumption.
Patent Information
- Application Number
- CN202310848109.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-07-11
- Publication Date
- 2025-01-17
AI Technical Summary
When obtaining traffic light status information, the prior art is susceptible to factors such as occlusion, bad weather and communication network unreliability, resulting in high-precision computing resources consumption and low accuracy. The method based on high-precision maps requires the latest map data and updates. or the error cannot be recognized normally.
Ultra-wideband positioning technology is used to obtain vehicle trajectory data, and the status information of traffic lights is estimated by processing vehicle status data and relative position data, combined with maximum likelihood estimation, Bayesian reasoning or machine learning.
Improves the estimation accuracy of traffic light status information, reduces the demand for computing resources, enhances reliability and security in occluded or inclement weather conditions without relying on high-quality map data.
Smart Images
Figure CN119319840A_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to the field of intelligent vehicle technology, and in particular to a traffic light state estimation method, system, computing device and vehicle for a vehicle. Background Art
[0002] Traffic light status information is important information required in various road traffic scenarios. Accurate and fast acquisition of traffic light status information is of great significance to vehicle users.
[0003] There are several main ways for existing vehicles to obtain traffic light status information: 1) Image processing methods, for example, using convolutional neural networks (CNN) to classify images taken by cameras installed on vehicles. CNN can learn to recognize the characteristics of traffic lights (such as shape, color and position) by training on large data sets of labeled images, and then output the traffic light status information of any given image; 2) Methods based on vehicle self-organizing networks (VANETs) to obtain traffic light status information by communicating with roadside units; 3) Image processing methods based on high-precision maps, which combine image processing with accurate 3D maps and self-positioning technology. 3D maps can provide information about the location and direction of traffic lights in the scene, and then regions of interest (ROI) can be extracted from the captured images. ROIs can then be processed by CNN or other algorithms to identify and obtain traffic light status information.
[0004] However, the existing technologies have different defects. Image processing methods require a lot of computing resources and memory, and may not be able to handle scenes with occlusion or other complex scenes, such as a scene where a large bus in front of the vehicle blocks the traffic light, and scenes with low visibility in bad weather (such as rain, snow, fog), etc. VANET-based methods depend on the availability and reliability of the communication network and may not be able to cope with dynamic changes or malicious attacks. Image processing methods based on high-precision maps require a high-quality and up-to-date map database. When the map data is not updated or an error occurs, the traffic light status information cannot be normally identified. In addition, this method has the same defects as the image processing method. Summary of the invention
[0005] The object of the present invention is to overcome at least part of the above-mentioned defects, and provides a traffic light state estimation method, system, computing device and vehicle for a vehicle.
[0006] According to a first aspect of the present invention, a method for estimating the state of a traffic light for a vehicle is provided, comprising: acquiring vehicle trajectory data based on ultra-wideband positioning technology; processing the vehicle trajectory data to acquire target data corresponding to a target traffic light; and performing state estimation based on the target data to acquire state information of the target traffic light.
[0007] According to a preferred embodiment of the first aspect of the present invention, the vehicle trajectory data includes vehicle status data, and the method includes: acquiring the vehicle status data through a sensor of the vehicle; and processing the vehicle trajectory data based on the vehicle status data to obtain the target data.
[0008] According to a preferred embodiment of the first aspect of the present invention, the vehicle status data includes vehicle speed data and acceleration data, and the method includes: processing the vehicle trajectory data based on the vehicle speed data and acceleration data to obtain vehicle start data and vehicle stop data; performing state estimation based on the vehicle start data and vehicle stop data to obtain status information of the target traffic light.
[0009] According to a preferred embodiment of the first aspect of the present invention, the vehicle trajectory data includes relative position data between multiple vehicles obtained based on ultra-wideband positioning technology, and the method includes: performing state estimation based on the relative position data between the vehicles, the vehicle start data and the vehicle stop data to obtain the state information of the target traffic light.
[0010] According to a preferred embodiment of the first aspect of the present invention, the method for obtaining the vehicle start data and the vehicle stop data includes: segmenting the vehicle trajectory data by clustering, region growing and / or threshold setting; classifying the segmentation results by a k-nearest neighbor algorithm, a support vector machine and / or a decision tree to obtain the vehicle start data and the vehicle stop data.
[0011] According to a preferred embodiment of the first aspect of the present invention, the method for acquiring the vehicle trajectory data includes: a positioning algorithm based on received signal strength, a positioning algorithm based on flight time, a positioning algorithm based on arrival time, a positioning algorithm based on arrival time difference, and a positioning algorithm based on arrival angle.
[0012] According to a preferred embodiment of the first aspect of the present invention, the manner of performing state estimation based on the target data includes: maximum likelihood estimation, Bayesian reasoning or machine learning.
[0013] According to a second aspect of the present invention, a traffic light state estimation system for a vehicle is provided, comprising: a trajectory data acquisition module, used to acquire vehicle trajectory data based on ultra-wideband positioning technology; a data processing module, used to process the vehicle trajectory data to acquire target data corresponding to a target traffic light; and a state estimation module, used to perform state estimation based on the target data to acquire state information of the target traffic light.
[0014] According to a third aspect of the present invention, a computing device is provided, comprising: at least one processor; and a memory for storing machine-readable instructions, which, when executed, enables the at least one processor to execute the traffic light state estimation method for a vehicle as described in the first aspect above.
[0015] According to a fourth aspect of the present invention, a vehicle is provided, comprising: a traffic light state estimation system for a vehicle according to the second aspect or a computing device according to the third aspect.
[0016] The present invention collects vehicle trajectory data based on ultra-wide band (UWB) positioning technology with high positioning accuracy to estimate traffic light status information, and the estimation accuracy of the traffic light status information is high; the present invention does not require a large amount of computing resources and memory, and even if there is an obstruction between the vehicle and the traffic light or in severe weather conditions, it will not affect the vehicle's acquisition of the traffic light status, thereby improving driving safety; in addition, the invention adopts UWB communication, which has higher security and reliability. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] Other features and advantages of the present invention will be better understood from the following detailed description of preferred embodiments in conjunction with the accompanying drawings, in which the same reference numerals represent the same or similar components.
[0018] Figure 1 A schematic flow chart of an exemplary method for estimating a traffic light state of a vehicle according to an embodiment of the present disclosure is shown.
[0019] Figure 2 A block diagram of an exemplary traffic light state estimation system for a vehicle according to an embodiment of the present disclosure is shown.
[0020] Figure 3 A block diagram of an exemplary computing device according to an embodiment of the present disclosure is shown. DETAILED DESCRIPTION
[0021] The following detailed description is merely exemplary in nature and is not intended to limit the application and use. Furthermore, the present invention is not intended to be limited by any stated or implied theory presented in the foregoing technical field, background technology, summary of the invention or the following detailed description.
[0022] As mentioned above, current mainstream vehicles either obtain traffic light status information by communicating with traffic light control systems or by image recognition methods. These methods are easily affected by weather, scenes, and communication networks and fail to obtain accurate information, which in turn affects the user experience and even driving safety.
[0023] Therefore, some exemplary embodiments of the present disclosure provide a traffic light state estimation method, system, computing device and vehicle for a vehicle. Figure 1-Figure 3 This is described in more detail.
[0024] Reference Figure 1 , shows a flow chart of an exemplary method 100 for estimating the state of a traffic light for a vehicle according to an embodiment of the present disclosure. The method 100 is applicable to any type of mobile vehicle, such as a passenger car (including a driver's seat and not more than nine seats), a commercial vehicle (a coach, a truck, etc.), or other types of vehicles. Figure 1 As shown, the method 100 includes steps 110-130.
[0025] Step 110, obtaining vehicle trajectory data based on ultra-wideband positioning technology. UWB positioning technology uses nanosecond non-sinusoidal narrow pulses to transmit data, so it occupies a wide spectrum range and has the advantages of low system complexity, low power spectrum density of transmitted signals, insensitivity to channel fading, low interception capability, and high positioning accuracy. The vehicle trajectory data may include information such as the vehicle's location, time, speed, acceleration, and driving direction.
[0026] Preferably, in some examples, the vehicle's speed, acceleration, driving direction and other information can be obtained through the vehicle's internal sensors, without the need to obtain through UWB positioning technology, which reduces the amount of calculation and makes information acquisition faster and more accurate. In other examples, by fusing the data obtained by USB positioning technology with the data obtained by the vehicle's internal sensors or the vehicle's navigation system, the accuracy of the data used for state estimation is further improved, thereby improving the accuracy of traffic light state estimation.
[0027] In some examples, the vehicle trajectory information is obtained by communicating with multiple USB positioning base stations (anchor points) deployed around traffic lights through a USB positioning module installed on the vehicle body. USB positioning technology can achieve centimeter-level positioning of the vehicle, with good positioning accuracy and high signal security. Optionally, the present invention can obtain the required vehicle trajectory data through any one of a positioning algorithm based on received signal strength (RSS, Received Signal Strength), a positioning algorithm based on time of flight (TOF, Time of Flight), a positioning algorithm based on time of arrival (TOA, Time of Arrival), a positioning algorithm based on time difference of arrival (TDOA, Time Difference Of Arrival) and a positioning algorithm based on angle of arrival (AOA, Angle of Arrival).
[0028] Step 120, the vehicle trajectory data is processed to obtain target data corresponding to the target traffic light. The target data corresponding to the target traffic light may be data within a spatial range corresponding to the target traffic light, such as within a radius of 200 meters with the target traffic light as the center; or data within a time range corresponding to the target traffic light, such as when the target traffic light is a multi-cycle traffic light, the vehicle trajectory data may be processed based on the time periods corresponding to each cycle of the traffic light to obtain data for each time period.
[0029] In some examples, before the main processing of the vehicle trajectory data, the data is first pre-processed to improve the quality of the data and make it suitable for further analysis. The optional processing methods include noise reduction, outlier detection, data fusion, data aggregation, etc.
[0030] Further, in some examples, the pre-processed data is segmented and classified. For example, the vehicle trajectory data is segmented into segments corresponding to different stages or different movements at a traffic light intersection. Segmentation can be performed using methods such as clustering, region growing, threshold setting, etc. In some examples, the vehicle trajectory data can be classified according to the start and stop mode of the vehicle. Classification can be performed using methods such as k-nearest neighbor algorithm, support vector machine, decision tree, etc.
[0031] In some examples, the vehicle trajectory data includes vehicle state data, and the method 100 obtains the vehicle state data through the vehicle sensor, and processes the vehicle trajectory data based on the vehicle state data to obtain the target data. The vehicle generally obtains the vehicle's speed, acceleration, angle and other information through the speed sensor, three-axis acceleration sensor, tilt sensor and the like in its internal electronic stability control module. Specifically, the vehicle state data includes the vehicle's speed data and acceleration data, and the method 100 processes the vehicle trajectory data based on the vehicle's speed data and acceleration data to obtain the vehicle start data and vehicle stop data; and performs state estimation based on the vehicle start data and the vehicle stop data to obtain the state information of the target traffic light. The vehicle start data and the vehicle stop data are highly correlated with the traffic light state, so by screening the vehicle trajectory data to extract the data corresponding to the vehicle start and stop states, the accuracy of the traffic light state estimation can be guaranteed while reducing the amount of data in the state estimation process.
[0032] In some examples, the vehicle trajectory data includes relative position data between multiple vehicles obtained based on ultra-wideband positioning technology, and the method 100 includes: performing state estimation based on the relative position data between the vehicles, the vehicle start data, and the vehicle stop data to obtain the state information of the target traffic light. Specifically, the segmentation and classification accuracy of the vehicle trajectory data can be further improved based on the relative position data between the vehicles. For example, the vehicle start data can be divided into the start data of the vehicle closest to the traffic light and the start data of other vehicles.
[0033] Step 130, perform state estimation based on the target data to obtain state information of the target traffic light. In some examples, the state of the traffic light, such as color, remaining time, etc., is inferred based on the vehicle start data and vehicle stop data corresponding to the vehicle start phase and vehicle stop phase.
[0034] The options for state estimation based on target data include: maximum likelihood estimation, Bayesian reasoning, machine learning, etc. Maximum likelihood estimation is a method for finding the signal timing that maximizes the probability of observing the vehicle trajectory given a signal model. Optional techniques include expectation maximization algorithm and gradient descent algorithm. Bayesian reasoning is a method for finding the signal timing with the highest posterior probability given some prior knowledge of the vehicle trajectory and signal model. Optional techniques include Markov chain Monte Carlo algorithm, particle filter algorithm, etc. Machine learning is a method for learning the function that maps the vehicle trajectory to the signal timing from training data. Optional techniques include neural networks, fuzzy logic systems, genetic algorithms, etc.
[0035] In some examples, the estimated result of the traffic light status information may be verified, for example, by comparing the estimated result of the present method with the actual traffic light status received through real-time communication or traffic light status information obtained by other existing methods (such as image processing methods, image recognition methods based on high-precision maps).
[0036] Reference Figure 2 , an embodiment of the present disclosure provides a block diagram of a traffic light state estimation system 200 for a vehicle, and each module of the system 200 can be implemented by software, hardware (for example, integrated circuit, FPGA, etc.) or a combination of software and hardware. Figure 2 As shown, the system 200 includes: a trajectory data acquisition module 210, which is used to acquire vehicle trajectory data based on ultra-wideband positioning technology; a data processing module 220, which is used to process the vehicle trajectory data to acquire target data corresponding to the target traffic light; and a state estimation module 230, which is used to perform state estimation based on the target data to acquire state information of the target traffic light. This embodiment is similar to the implementation of the method embodiment described above, so it will not be repeated here.
[0037] Reference Figure 3 , a block diagram of an exemplary computing device 300 according to an embodiment of the present disclosure is shown. Figure 3 As shown, the computing device 300 includes at least one processor 310 and a memory 320 coupled to the at least one processor 310. The memory 320 is used to store machine-readable instructions, and when the machine-readable instructions are executed by the at least one processor 310, the processor 310 executes the method in the above embodiment (for example, any one or more steps of the aforementioned method 100).
[0038] An embodiment of the present disclosure provides an exemplary vehicle, which includes the traffic light state estimation system 200 or the computing device 300 for a vehicle in the above embodiments. This embodiment is similar to the implementation of the above embodiments, so it will not be repeated here.
[0039] The foregoing discussion is merely the disclosure and description of the exemplary embodiments of the present invention. Through these discussions and the accompanying drawings and claims, those skilled in the art will readily appreciate that various changes, modifications and variations may be made to the present invention without departing from the spirit and scope of the present invention as defined by the appended claims.
Claims
1. A method for estimating a traffic light state of a vehicle, comprising: Obtain vehicle trajectory data based on ultra-wideband positioning technology; Processing the vehicle trajectory data to obtain target data corresponding to a target traffic light; State estimation is performed based on the target data to obtain state information of the target traffic light.
2. The method for estimating the traffic light state of a vehicle according to claim 1, wherein: The vehicle trajectory data includes vehicle status data, and the method includes: Acquiring the vehicle status data through a sensor of the vehicle; The vehicle trajectory data is processed based on the vehicle state data to obtain the target data.
3. The method for estimating the traffic light state of a vehicle according to claim 2, wherein: The vehicle state data includes speed data and acceleration data of the vehicle, and the method includes: Processing the vehicle trajectory data based on the speed data and acceleration data of the vehicle to obtain vehicle start data and vehicle stop data; The state estimation is performed based on the vehicle start data and the vehicle stop data to acquire the state information of the target traffic light.
4. The method for estimating the traffic light state of a vehicle according to claim 3, wherein: The vehicle trajectory data includes relative position data between multiple vehicles obtained based on ultra-wideband positioning technology, and the method includes: The state estimation is performed based on the relative position data between the vehicles, the vehicle start data and the vehicle stop data to obtain the state information of the target traffic light.
5. The method for estimating the traffic light state of a vehicle according to claim 3, wherein: The method for obtaining the vehicle start data and the vehicle stop data includes: Segmenting the vehicle trajectory data by clustering, region growing and / or threshold setting; The segmentation results are classified by a k-nearest neighbor algorithm, a support vector machine and / or a decision tree to obtain the vehicle start data and the vehicle stop data.
6. The method for estimating the traffic light state of a vehicle according to claim 1, wherein: The vehicle trajectory data is obtained by at least one of the following methods: a positioning algorithm based on received signal strength, a positioning algorithm based on flight time, a positioning algorithm based on arrival time, a positioning algorithm based on arrival time difference, and a positioning algorithm based on arrival angle.
7. The method for estimating a traffic light state for a vehicle according to claim 1, wherein: The method of performing state estimation based on the target data includes: maximum likelihood estimation, Bayesian reasoning or machine learning.
8. A traffic light state estimation system for a vehicle, comprising: A trajectory data acquisition module, used to acquire vehicle trajectory data based on ultra-wideband positioning technology; A data processing module, used for processing the vehicle trajectory data to obtain target data corresponding to the target traffic light; A state estimation module is used to perform state estimation based on the target data to obtain state information of the target traffic light.
9. A computing device comprising: at least one processor; as well as A memory storing machine-readable instructions, which, when executed by the at least one processor, enable the at least one processor to perform the traffic light state estimation method for a vehicle according to any one of claims 1-7. 10 . A vehicle, comprising the traffic light state estimation system for a vehicle according to claim 8 or the computing device according to claim 9 .