System and method of precision localization algorithm of autonomous platform

Abstract

The present invention relates to a system composed of high-level and low-level observation modules for precise position estimation. A precise position estimation system of an autonomous driving platform according to one embodiment includes a high-level observation module and a low-level observation module for estimating a precise position by observing the movement of a vehicle. The high-level observation module collects GPS signals to estimate the position of the vehicle, the low-level observation module estimates a correction value for a dynamic difference greater than a reference value, and the high-level observation module can correct the position of the vehicle using the estimated correction value.

Description

Technology Field

[0001] The present invention uses a system composed of high-level and low-level observation modules for precise position estimation, and each observation module is used to determine the current driving situation of the vehicle and to fuse appropriate sensor information accordingly. Background Technology

[0002] Current autonomous driving systems use GPS to accurately estimate a vehicle's location, but in urban areas, performance cannot be guaranteed due to numerous external factors that degrade GPS performance, such as dense forests of buildings, overpasses, and underground parking lots.

[0003] Existing algorithms for accurate location estimation while reducing reliance on GPS systems mainly consist of perception sensor-based Visual Localization and advanced vehicle model-based Odometry Localization. However, Visual Localization is vulnerable to changes in the surrounding environment, such as lighting, weather, and viewpoint variations, and suffers from high computational loads. In contrast, Odometry Localization faces the problem of accumulating error increasing over time due to the high reliance on model-based predictions.

[0004] Although the Kalman filter methodology is widely used to estimate the state of a system from observations with errors, there is a need for a method that not only fuses various information for error compensation but also utilizes and fuses appropriate sensor information using the Kalman filter tailored to vehicle characteristics and driving conditions. Prior art literature

[0005] Korean Published Patent No. 2020-0021688 "System for position estimation of an autonomous vehicle and method for position estimation thereof" Korean Registered Patent No. 2454882 "Method, device, apparatus, storage medium and program for estimating vehicle driving position" Korean Registered Patent No. 2093743 "Lane-level vehicle position information positioning system using sensor fusion" The problem to be solved

[0006] The present invention aims to determine the position of a vehicle by fusing multiple sensor signals to reduce dependence on a GPS system in performing precise position estimation.

[0007] The present invention aims to develop a methodology that considers vehicle characteristics and driving conditions by designing a hierarchical Kalman filter structure, rather than simply fusing three pieces of information.

[0008] The present invention aims to accurately estimate vehicle movements, such as slippage, even in various driving situations, such as crab walking, by utilizing an advanced vehicle model and a hierarchical Kalman filter module. means of solving the problem

[0009] A precision position estimation system of an autonomous driving platform according to one embodiment includes a high-level observation module and a low-level observation module for estimating a precision position by observing the movement of a vehicle, wherein the high-level observation module collects GPS signals to estimate the position of the vehicle, the low-level observation module estimates a correction value for a dynamic difference greater than a reference value, and the high-level observation module can correct the position of the vehicle using the estimated correction value.

[0010] According to one embodiment, the low-level observation module, in the process of generating a correction value for a dynamic difference greater than the reference value, first predicts the speed of the vehicle based on a vehicle model, secondly predicts the speed of the vehicle based on IMU (Inertial Measurement Unit) sensor values ​​regarding the movement of the vehicle, and estimates the speed information of the vehicle by comparing the first and second predicted speeds of the vehicle using an Extended Kalman Filter, and the high-level observation module can correct the position of the vehicle based on the estimated speed information.

[0011] The low-level observation module according to one embodiment can secondarily predict the longitudinal speed or lateral speed of the vehicle based on acceleration and angular velocity information of the vehicle as IMU (Inertial Measurement Unit) sensor values.

[0012] The low-level observation module according to one embodiment can secondarily predict the longitudinal speed or lateral speed of the vehicle based on the odometry information of the vehicle.

[0013] According to one embodiment, the high-level observation module can estimate the position of the vehicle by using an Extended Kalman Filter on the collected GPS signal.

[0014] According to one embodiment, the high-level observation module can estimate the position of the vehicle by calculating the error covariance with the predicted value of the next step based on the initial value of the collected GPS signal during the process of estimating the position of the vehicle using the Extended Kalman Filter, calculating the Kalman gain which is the weight of the estimation value calculation formula to determine the ratio of the observed value to the predicted value, and then calculating the estimated value.

[0015] The low-level observation module according to one embodiment can estimate the speed of the vehicle by using an Extended Kalman Filter on the Inertial Measurement Unit (IMU) sensor values ​​and odometry information of the vehicle.

[0016] The low-level observation module according to one embodiment can estimate the speed of the vehicle by calculating the error covariance with the predicted value of the next step based on the initial values ​​of the vehicle's IMU (Inertial Measurement Unit) sensor values ​​and odometry information, calculating the Kalman gain which is the weight of the estimation value calculation formula to determine the ratio of the observed value to the predicted value, and then calculating the estimated value.

[0017] A precise position estimation method for an autonomous driving platform according to one embodiment may include the steps of collecting GPS signals to estimate the position of a vehicle, estimating a correction value for a dynamic difference greater than or equal to a reference value, and correcting the estimated position of the vehicle using the estimated correction value.

[0018] The step of estimating a correction value for a dynamic difference greater than the reference value according to one embodiment, and correcting the estimated position of the vehicle using the estimated correction value, may include, in the process of generating a correction value for a dynamic difference greater than the reference value, a step of first predicting the speed of the vehicle based on a vehicle model, a step of secondarily predicting the speed of the vehicle based on IMU (Inertial Measurement Unit) sensor values ​​regarding the movement of the vehicle, and a step of estimating the speed information of the vehicle by comparing the first and second predicted speeds of the vehicle using an Extended Kalman Filter.

[0019] The step of estimating a correction value for a dynamic difference greater than or equal to the reference value according to one embodiment, and correcting the estimated position of the vehicle using the estimated correction value, may include the step of correcting the position of the vehicle based on the estimated speed information.

[0020] The step of secondarily predicting the speed of the vehicle based on IMU (Inertial Measurement Unit) sensor values ​​regarding the movement of the vehicle according to one embodiment may include the step of secondarily predicting the longitudinal speed or lateral speed of the vehicle based on acceleration and angular velocity information of the vehicle as IMU (Inertial Measurement Unit) sensor values.

[0021] The step of secondarily predicting the speed of the vehicle according to one embodiment may include the step of secondarily predicting the longitudinal speed or lateral speed of the vehicle based on the odometry information of the vehicle.

[0022] The step of estimating the position of a vehicle by collecting GPS signals according to one embodiment may include estimating the position of the vehicle by using an Extended Kalman Filter on the collected GPS signals, calculating the error covariance with the predicted value of the next step based on the initial value of the collected GPS signals, calculating the Kalman gain which is the weight of the estimation value calculation formula to determine the ratio of the observed value to the predicted value, and then calculating the estimated value to estimate the position of the vehicle.

[0023] The step of estimating a correction value for a dynamic difference greater than a reference value according to one embodiment, and correcting the position of the estimated vehicle using the estimated correction value, may include the step of estimating the speed of the vehicle by using an Extended Kalman Filter on the vehicle's IMU (Inertial Measurement Unit) sensor value and odometry information, calculating the error covariance with the predicted value of the next step based on the initial value of the vehicle's IMU (Inertial Measurement Unit) sensor value and odometry information, calculating the Kalman gain which is the weight of the estimation value calculation formula to determine the ratio of the observed value to the predicted value, and then calculating the estimated value to estimate the speed of the vehicle. Effects of the invention

[0024] According to one embodiment, the position of a vehicle can be determined by fusing multiple sensor signals to reduce reliance on a GPS system when performing precise position estimation.

[0025] According to one embodiment, instead of simply fusing three pieces of information, a hierarchical Kalman filter structure can be designed to develop a methodology that considers the characteristics of the vehicle and the driving conditions of the vehicle.

[0026] According to one embodiment, by using an advanced vehicle model and a hierarchical Kalman filter module, movements such as vehicle slippage can be accurately estimated even in various driving situations, such as crab walking. Brief explanation of the drawing

[0027] FIG. 1 is a diagram illustrating a precision position estimation system of an autonomous driving platform according to one embodiment. FIG. 2 is a configuration diagram of a precision position estimation system according to the present invention. Figure 3a is a diagram illustrating a correction technique based on a front-wheel drive vehicle model. Figure 3b is a diagram illustrating a correction technique based on a four-wheeled vehicle model. Figure 4 is a flowchart of an extended Kalman filter. Fig. 5 is a four-wheel steering vehicle model. Figure 6 shows the result of estimating the longitudinal speed of a vehicle using a low-level observation module according to the present invention. Figure 7 is the result of estimating the lateral velocity of a vehicle using a low-level observation module according to the present invention. Figure 8 is the result of precise position estimation of a vehicle using a high-level observation module according to the present invention. FIG. 9 is a diagram illustrating a precise position estimation method of an autonomous driving platform according to one embodiment. FIG. 10 is a diagram illustrating a method for estimating a correction value for a dynamic difference greater than a reference value according to one embodiment. Specific details for implementing the invention

[0028] Specific structural or functional descriptions of embodiments according to the concept of the present invention disclosed herein are provided merely for the purpose of explaining embodiments according to the concept of the present invention, and embodiments according to the concept of the present invention may be implemented in various forms and are not limited to the embodiments described herein.

[0029] Embodiments according to the concept of the present invention may be subject to various modifications and may take various forms; therefore, embodiments are illustrated in the drawings and described in detail in this specification. However, this is not intended to limit the embodiments according to the concept of the present invention to specific disclosed forms, and includes modifications, equivalents, or substitutions that fall within the spirit and scope of the present invention.

[0030] Terms such as "first" or "second" may be used to describe various components, but said components should not be limited by said terms. For the sole purpose of distinguishing one component from another, for example, without departing from the scope of rights according to the concept of the present invention, the first component may be named the second component, and similarly, the second component may be named the first component.

[0031] When it is stated that one component is "connected" or "connected" to another component, it should be understood that while it may be directly connected or connected to that other component, there may also be other components in between. Conversely, when it is stated that one component is "directly connected" or "directly connected" to another component, it should be understood that there are no other components in between. Expressions describing the relationships between components, such as "between," "exactly between," or "directly adjacent to," should be interpreted in the same way.

[0032] The terms used herein are used merely to describe specific embodiments and are not intended to limit the invention. Singular expressions include plural expressions unless the context clearly indicates otherwise. In this specification, terms such as “comprising” or “having” are intended to specify the existence of the described features, numbers, steps, actions, components, parts, or combinations thereof, and should be understood as not precluding the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof.

[0033] Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as generally understood by those skilled in the art to which the present invention pertains. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an ideal or overly formal sense unless explicitly defined in this specification.

[0035] Hereinafter, embodiments will be described in detail with reference to the attached drawings. However, the scope of the patent application is not limited or restricted by these embodiments. Identical reference numerals in each drawing indicate identical components.

[0036] FIG. 1 is a drawing illustrating a precision position estimation system (100) of an autonomous driving platform according to one embodiment.

[0037] The present invention aims to determine the position of a vehicle by fusing multiple sensor signals to reduce dependence on a GPS system in performing precise position estimation.

[0038] Furthermore, the present invention aims to develop a methodology that considers vehicle characteristics and driving conditions by designing a hierarchical Kalman filter structure, rather than simply fusing three pieces of information.

[0039] Furthermore, by utilizing an advanced vehicle model and a hierarchical Kalman filter module, the aim is to accurately estimate vehicle movements, such as slippage, even in various driving situations like crab walking.

[0040] To this end, a precision position estimation system (100) of an autonomous driving platform according to one embodiment may include a high-level observation module (110) and a low-level observation module (120) to estimate a precision position by observing the movement of a vehicle.

[0041] A high-level observation module (110) according to one embodiment can estimate the location of the vehicle by collecting GPS signals. GPS is a satellite-based positioning system that can calculate the coordinates of a specific location by receiving signals transmitted from various satellites on Earth. Therefore, this high-level observation module enables accurate determination of where the vehicle is located.

[0042] A low-level observation module (120) according to one embodiment estimates a correction value for a dynamic difference greater than a reference value, and a high-level observation module (110) can correct the position of the vehicle using the estimated correction value.

[0043] In particular, the low-level observation module (120) can first predict the speed of the vehicle based on the vehicle model and secondarily predict the speed of the vehicle based on IMU (Inertial Measurement Unit) sensor values ​​regarding the vehicle's movement in the process of generating a correction value for a dynamic difference greater than a reference value. The IMU can measure movement information such as the acceleration and rotational speed of the vehicle using sensors such as an accelerometer and a gyroscope.

[0044] In addition, the low-level observation module (120) can further utilize odometry information regarding the movement of the vehicle during the second prediction process.

[0045] Odometry is a technology that estimates distance and direction by measuring changes that occur as a vehicle moves.

[0046] The low-level observation module (120) can use an Extended Kalman Filter to compare the speeds of the first and second predicted vehicles to estimate the speed information of the vehicle and provide it to the high-level observation module.

[0047] Accordingly, the high-level observation module (110) can correct the position of the vehicle based on the estimated speed information.

[0048] For example, the low-level observation module (120) can predict the longitudinal speed or lateral speed of the vehicle based on acceleration and angular velocity information of the vehicle as IMU (Inertial Measurement Unit) sensor values.

[0049] In addition, the low-level observation module (120) can predict the longitudinal speed or lateral speed of the vehicle based on the vehicle's odometry information.

[0050] In particular, the high-level observation module (110) can estimate the position of the vehicle by using an Extended Kalman Filter on the collected GPS signal.

[0051] In addition, the high-level observation module (110) can estimate the position of the vehicle by calculating the error covariance and the predicted value of the next step based on the initial value of the collected GPS signal during the process of estimating the position of the vehicle using the Extended Kalman Filter, and by calculating the Kalman gain, which is the weight of the estimation value calculation formula, determining the ratio of the observed value to the predicted value, and then calculating the estimated value.

[0052] The low-level observation module (120) can estimate the speed of the vehicle by using an Extended Kalman Filter on the predicted value of the vehicle model, the sensor value of the vehicle's IMU (Inertial Measurement Unit) and odometry information.

[0053] In this process, the low-level observation module (120) can estimate the speed of the vehicle by using an Extended Kalman Filter to calculate the next step's predicted value and error covariance based on the initial values ​​of the vehicle model's predicted value, the vehicle's IMU (Inertial Measurement Unit) sensor value and odometry information, and the Kalman gain, which is the weight of the estimated value calculation formula, to determine the ratio of the observed value to the predicted value, and then calculating the estimated value.

[0054] FIG. 2 is a configuration diagram (200) of a precise position estimation system according to the present invention.

[0055] The present invention uses a system composed of high-level and low-level observation modules for precise position estimation, and each observation module is used to determine the current driving situation of the vehicle and to fuse appropriate sensor information accordingly.

[0056] In particular, the high-level observation module receives GPS signals and uses an Extended Kalman Filter for error compensation and state observation of the received GPS signals.

[0057] In the low-level observation module, the IMU sensor value and the initial value of odometry information as an in-vehicle can be received as input values ​​for the extended Kalman filter.

[0058] The processing in the Extended Kalman Filter is divided into the prediction phase and the observation and correction phase. In the prediction phase, the predicted value for the next step and the error covariance are calculated based on the initial value. In the observation and correction phase, the Kalman gain, which serves as the weight for the estimation formula, is calculated; this gain is then used to determine the ratio between the observed and predicted values, and the estimated value is calculated.

[0059] The Extended Kalman Filter (EKF) is a type of filter used to estimate state in nonlinear systems. While the basic Kalman filter is effective for linear dynamic systems, many real-world problems are nonlinear. The Extended Kalman Filter is an extension of the filter through linear approximation to enable it to operate effectively in such nonlinear systems.

[0060] In order for the extended Kalman filter to compensate for errors in each piece of information and observe the system state, it can first use a dynamic model to predict the system's current state. This prediction process is based on previous state estimates and current control inputs.

[0061] Meanwhile, a plant model is a mathematical model of the dynamic characteristics of a process and can take the form of a difference equation or a transfer function. In particular, the plant model in Fig. 2 can be used to predict and analyze the operation of a system through simulation before implementing an actual system or process.

[0062] Next, the extended Kalman filter can calculate the difference from the actual measurement based on the predicted state.

[0063] The extended Kalman filter can correct the system state using residuals. In this process, linear approximation techniques, such as the Jacobian matrix, can be used to handle nonlinearity.

[0064] The Extended Kalman Filter can update the error covariance to track system uncertainty. Through this, the Extended Kalman Filter can effectively manage measurement and prediction errors.

[0065] By repeating these steps, the Extended Kalman Filter can perform state estimation even in nonlinear systems with uncertainty and compensate for errors in measured information. This can be utilized to accurately estimate state variables such as vehicle position and speed.

[0066] FIG. 3a is a drawing showing a front-wheel vehicle model-based correction technique (310), and FIG. 3b is a drawing showing a four-wheel vehicle model-based correction technique (320).

[0067] As shown in Fig. 3a, the front wheel steering system can be seen to change the direction angle of the vehicle while driving along the same path.

[0068] On the other hand, in the case of a four-wheel steering system, the vehicle can drive along a given path without changing its direction angle.

[0069] In situations where the vehicle’s direction angle does not change, such as when the vehicle is sliding, a low-level observation module can be used to predict whether the vehicle is intentionally moving.

[0070] Figure 4 is a flowchart of an extended Kalman filter.

[0071] The processing in the Extended Kalman Filter is divided into the Initial Estimation, Prediction, Observation, and Update stages. In the Prediction stage, the predicted value for the next step and the error covariance can be calculated based on the initial value.

[0072] The observation and calibration step calculates the Kalman gain, which is the weight of the estimate formula, and uses it to determine the ratio of the observed value to the predicted value, after which the estimate can be calculated.

[0073] The first step of the filtering process is the Initial Estimate, during which the initial state and error covariance can be established. This represents the initial prediction for the system to be estimated.

[0074] Meanwhile, in the prediction stage, the next step from the current state can be predicted based on initial values. In this case, if the system is nonlinear, a new predicted value can be calculated using the predicted value of the previous state and the nonlinear model, and the error covariance can be updated.

[0075] In the observation phase, the actual state of the system is measured and used to calculate the difference from the predicted value. This difference can be used to correct the prediction and improve the estimate.

[0076] In the Update step, Kalman gain is calculated, and Kalman gain corresponds to a weight that determines how much to adjust the predicted value by considering the discrepancy between the current predicted value and the observed value.

[0077] Subsequently, the estimate is updated using Kalman gains and the error covariance is recalculated; by repeating this process, the optimal estimate of the system state can be performed.

[0078] In the case of the high-level observation module, odometry information is applied to the vehicle model to predict the position of the next step, and this can be fused with GPS measurements to calculate the position.

[0079] Low-level observation modules are used to handle excessive dynamic differences in situations that do not correspond to normal driving conditions, such as when a vehicle is sliding.

[0080] In the case of the vehicle model, a rear wheel axle was added to the existing Bicycle Model to allow for rear-wheel steering, and the model was designed by considering not only lateral vehicle dynamics but also longitudinal vehicle dynamics.

[0081] Low-level observation modules are used to handle excessive dynamic differences in situations that do not correspond to normal driving conditions, such as when a vehicle is sliding.

[0082] The high-level observation module provides accurate location information even in situations where there are external factors that degrade GPS performance. It calculates an estimated value by fusing the predicted value obtained by applying GPS measurements, the speed of the vehicle currently in motion, and steering angle information to a four-wheel steering vehicle model.

[0083] The low-level observation module predicts the vehicle's longitudinal and lateral velocities based on acceleration and angular velocity information from the IMU sensor, and also predicts longitudinal and lateral velocities based on odometry information. Subsequently, it utilizes the observation module to compare the two sets of information to determine the vehicle's driving situation and fuses them to provide accurate velocities to the vehicle model used in the high-level observation module.

[0084] The algorithm of the present invention, which fuses various sensor signals to provide accurate location information to a mobility system, is divided into two layers, and verification of each layer was conducted by configuring a double lane-change test environment using a virtual environment simulator.

[0085] Fig. 5 is a four-wheel steering vehicle model.

[0086] General position estimation techniques differ in purpose from the localization of four-wheel steering vehicles.

[0087] Although there is a similarity in the method of predicting future position using a kinematic model of a four-wheel steering vehicle, the kinematic model has the problem that it is difficult to calculate parameters related to the vehicle's dynamic characteristics.

[0088] Therefore, the present invention differs from conventional technologies in that it uses a vehicle dynamic model.

[0089] In addition, the present invention includes a process of combining the data and IMU at a low-level observer to reduce uncertainty and increase the accuracy of the predicted value.

[0090] The lateral forces acting on the front and rear tires of the vehicle are, respectively , It is the same as. Here , These are the lateral forces of the front and rear wheels, respectively.

[0091] The equations for the vehicle's lateral force and moment are, respectively , It is the same as. Here is the mass of the vehicle, is the vehicle's moment of inertia, is the distance from the vehicle's center of gravity to the vehicle's front axle, and is the distance from the vehicle's center of gravity to the vehicle's rear axle.

[0092] ceremony , , , The combined 4-wheel steering dynamics model using , , It is the same as.

[0093] Figure 6 shows the result of estimating the longitudinal speed of a vehicle using a low-level observation module according to the present invention.

[0094] The longitudinal speed of a moving vehicle was predicted using a low-level observation module, and the result screen is as shown in Fig. 6. In Fig. 6, the result (True) is the true value provided by the virtual environment simulator, the result (Model) is the result predicted using a vehicle model, and the green result (Filter) is the estimated longitudinal speed value calculated using the low-level observation module proposed in the present invention.

[0095] As shown in the result of Figure 6, it can be seen that the longitudinal speed estimate calculated using the low-level observation module proposed in the present invention is closer to the true value than the longitudinal speed of the vehicle predicted using the vehicle model.

[0096] Figure 7 is the result of estimating the lateral velocity of a vehicle using a low-level observation module according to the present invention.

[0097] The lateral velocity of a moving vehicle was estimated using a low-level observation module, and the result screen is as shown in Fig. 7. In Fig. 7, Result (True) is the true value provided by the virtual environment simulator, Result (Model) is the result predicted using a vehicle model, and Result (Filter) is the lateral velocity estimate value calculated using the low-level observation module proposed in the present invention.

[0098] As shown in the result of Fig. 7, when lateral velocity occurs as the vehicle's movement changes, the vehicle model did not observe the change value, but in the case of the lateral velocity estimate value calculated using the low-level observation module proposed in the present invention, it can be seen that the vehicle's movement is observed.

[0099] Figure 8 is the result of precise position estimation of a vehicle using a high-level observation module according to the present invention.

[0100] Based on the vehicle speed information obtained from the low-level observation module, the position of the moving vehicle was estimated using the high-level observation module, and the result screen is as shown in Fig. 8. In Fig. 8, the result (Truth) is the true value provided by the virtual environment simulator, the result (Raw) is the GPS measurement value including error, the result (Model) is the position prediction result using the vehicle model, and the result (Filter) is the position estimation value calculated using the high-level observation module proposed in the present invention.

[0101] As shown in the result of Figure 8, if only GPS containing errors is used, it is difficult to measure the accurate location of the vehicle, and the location prediction result using the vehicle model shows that the error accumulates over time. However, the location estimation result calculated using the high-level observation module proposed in the present invention shows a result close to the true value.

[0102] Most existing precision position estimation algorithms are applied to front-wheel steering systems. If a precision position estimation system considering a four-wheel steering system is developed through the present invention, it is expected to ensure stability in areas where autonomous driving is deemed difficult due to the degraded performance of GPS systems, and furthermore, contribute to the advancement of autonomous driving systems, such as driving in narrow roads or parking in such areas.

[0103] FIG. 9 is a diagram illustrating a precise position estimation method of an autonomous driving platform according to one embodiment.

[0104] A precise position estimation method for an autonomous driving platform according to one embodiment collects GPS signals to estimate the position of a vehicle (step 901).

[0105] For example, to estimate the location of a vehicle by collecting GPS signals, an Extended Kalman Filter is applied to the collected GPS signals to estimate the vehicle's position, and the error covariance with the predicted value of the next step can be calculated based on the initial value of the collected GPS signals. Next, the Kalman gain, which serves as the weight in the estimation formula, is calculated to determine the ratio between the observed value and the predicted value, and then the estimated value is calculated to estimate the vehicle's position.

[0106] Meanwhile, the precise position estimation method of an autonomous driving platform according to one embodiment estimates a correction value for a dynamic difference greater than a reference value (step 902), and can correct the estimated position of the vehicle using the estimated correction value (step 903).

[0107] FIG. 10 is a diagram illustrating a method for estimating a correction value for a dynamic difference greater than a reference value according to one embodiment.

[0108] In the present invention, to estimate a correction value, a correction value for a dynamic difference greater than a reference value is estimated, and the position of the estimated vehicle can be corrected using the estimated correction value.

[0109] To this end, the precise position estimation method of an autonomous driving platform according to one embodiment can first predict the speed of a vehicle based on a vehicle model for the vehicle in the process of generating a correction value for a dynamic difference greater than a reference value (step 1001).

[0110] Next, the precise position estimation method of the autonomous driving platform according to one embodiment can secondarily predict the speed of the vehicle based on IMU (Inertial Measurement Unit) sensor values ​​regarding the movement of the vehicle in the process of generating a correction value for a dynamic difference greater than a reference value (step 1002).

[0111] In this process, the precise position estimation method can secondarily predict the speed of the vehicle by additionally utilizing odometry information regarding the vehicle's movement.

[0112] In addition, the precise position estimation method can estimate vehicle speed information by utilizing an Extended Kalman Filter in the process of comparing the first predicted vehicle speed and the second predicted vehicle speed (step 1003).

[0113] In addition, the estimated position of the vehicle can be corrected using speed information for the estimated vehicle (step 1004).

[0114] In order to estimate a correction value for a dynamic difference greater than a reference value and to correct the estimated position of the vehicle using the estimated correction value, the position of the vehicle can be corrected based on the estimated speed information.

[0115] In the process of secondarily predicting the speed of the vehicle based on IMU (Inertial Measurement Unit) sensor values ​​regarding the movement of the vehicle, the longitudinal speed or lateral speed of the vehicle can be secondarily predicted based on the acceleration and angular velocity information of the vehicle as IMU (Inertial Measurement Unit) sensor values.

[0116] In addition, to predict the speed of the vehicle, the longitudinal speed or lateral speed of the vehicle can be predicted based on the odometry information of the vehicle.

[0117] For example, to correct the estimated position of a vehicle, the speed of the vehicle can be estimated by using an Extended Kalman Filter on the vehicle's IMU (Inertial Measurement Unit) sensor values ​​and odometry information. Additionally, in this process, the error covariance between the predicted value of the next step and the initial values ​​of the vehicle's IMU sensor values ​​and odometry information is calculated based on the initial values; the Kalman gain, which is a weighting factor in the estimation formula, is calculated to determine the ratio between the observed value and the predicted value; and then the estimated value is calculated to estimate the speed of the vehicle.

[0118] Ultimately, by using the present invention, the position of a vehicle can be determined by fusing multiple sensor signals to reduce dependence on the GPS system in performing precise position estimation.

[0119] Furthermore, rather than simply fusing the three pieces of information, a methodology can be developed that considers vehicle characteristics and driving conditions by designing a hierarchical Kalman filter structure; and by utilizing an advanced vehicle model and a hierarchical Kalman filter module, movements such as vehicle slippage can be accurately estimated even in various driving situations, such as crab walking.

[0121] The device described above may be implemented as a hardware component, a software component, and / or a combination of a hardware component and a software component. For example, the device and components described in the embodiments may be implemented using one or more general-purpose or special-purpose computers, such as, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable array (FPA), a programmable logic unit (PLU), a microprocessor, or any other device capable of executing and responding to instructions. The processing unit may execute an operating system (OS) and one or more software applications executed on said operating system. Additionally, the processing unit may access, store, manipulate, process, and generate data in response to the execution of the software. For ease of understanding, the processing unit may be described as being used as a single unit, but those skilled in the art will understand that the processing unit may include multiple processing elements and / or multiple types of processing elements. For example, the processing unit may include multiple processors or one processor and one controller. In addition, other processing configurations, such as parallel processors, are also possible.

[0122] Software may include computer programs, code, instructions, or a combination of one or more of these, and may configure a processing unit to operate as desired or command the processing unit independently or collectively. Software and / or data may be permanently or temporarily embodied in any type of machine, component, physical device, virtual equipment, computer storage medium or device, or transmitted signal wave so as to be interpreted by the processing unit or to provide instructions or data to the processing unit. Software may be distributed over networked computer systems and may be stored or executed in a distributed manner. Software and data may be stored on one or more computer-readable recording media.

[0123] The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded on a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc., either alone or in combination. The program instructions recorded on the medium may be those specifically designed and configured for the embodiment, or they may be those known and available to those skilled in the art of computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes; optical recording media such as CD-ROMs and DVDs; magneto-optical media such as floptical disks; and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, and flash memory. Examples of program instructions include machine code, such as that generated by a compiler, as well as high-level language code that can be executed by a computer using an interpreter, etc. The hardware devices described above may be configured to operate as one or more software modules to perform the operation of the embodiment, and vice versa.

[0124] Although the embodiments have been described above with reference to the limited drawings, those skilled in the art can make various modifications and variations from the description above. For example, suitable results can be achieved even if the described techniques are performed in a different order than described, and / or the components of the described system, structure, device, circuit, etc. are combined or assembled in a form different from described, or replaced or substituted by other components or equivalents.

[0125] Therefore, other implementations, other embodiments, and equivalents to the claims also fall within the scope of the claims set forth below.

Claims

Claim 1 A precision position estimation system for an autonomous driving platform comprising a high-level observation module and a low-level observation module for observing the movement of a vehicle to estimate a precise position, wherein the high-level observation module collects GPS signals to estimate the position of the vehicle, the low-level observation module estimates a correction value for a dynamic difference greater than a reference value, the high-level observation module corrects the position of the vehicle using the estimated correction value, and in the process of generating a correction value for a dynamic difference greater than the reference value, the low-level observation module first predicts the speed of the vehicle based on a vehicle model, secondarily predicts the speed of the vehicle based on IMU (Inertial Measurement Unit) sensor values ​​regarding the movement of the vehicle, and estimates speed information of the vehicle by comparing the first and second predicted speeds of the vehicle using an Extended Kalman Filter, and the high-level observation module corrects the position of the vehicle based on the estimated speed information. Claim 2 delete Claim 3 A precision position estimation system for an autonomous driving platform according to claim 1, wherein the low-level observation module secondarily predicts the longitudinal speed or lateral speed of the vehicle based on acceleration and angular velocity information of the vehicle as the IMU (Inertial Measurement Unit) sensor value. Claim 4 A precision position estimation system for an autonomous driving platform, characterized in that, in paragraph 3, the low-level observation module secondarily predicts the longitudinal speed or lateral speed of the vehicle based on the IMU (Inertial Measurement Unit) sensor value and the odometry information of the vehicle. Claim 5 A precision position estimation system for an autonomous driving platform, wherein, in claim 1, the high-level observation module estimates the position of the vehicle using an Extended Kalman Filter on the collected GPS signal. Claim 6 A precision position estimation system for an autonomous driving platform, characterized in that, in the process of estimating the position of the vehicle using the Extended Kalman Filter, the high-level observation module calculates the error covariance with the predicted value of the next step based on the initial value of the collected GPS signal, calculates the Kalman gain which is the weight of the estimation value calculation formula to determine the ratio of the observed value to the predicted value, and then calculates the estimated value to estimate the position of the vehicle. Claim 7 A precision position estimation system for an autonomous driving platform according to claim 1, wherein the low-level observation module estimates the speed of the vehicle using an Extended Kalman Filter on the Inertial Measurement Unit (IMU) sensor values ​​and odometry information of the vehicle. Claim 8 In claim 7, the low-level observation module is characterized by, in the process of estimating the speed of the vehicle using the Extended Kalman Filter, calculating the error covariance with the predicted value of the next step based on the initial values ​​of the vehicle's IMU (Inertial Measurement Unit) sensor values ​​and odometry information, calculating the Kalman gain which is the weight of the estimation value calculation formula to determine the ratio of the observed value to the predicted value, and then calculating the estimated value to estimate the speed of the vehicle. Claim 9 A method for precise position estimation of an autonomous driving platform, comprising: a step of collecting GPS signals to estimate the position of a vehicle; and a step of estimating a correction value for a dynamic difference greater than a reference value and correcting the estimated position of the vehicle using the estimated correction value, wherein the step of estimating a correction value for a dynamic difference greater than a reference value and correcting the estimated position of the vehicle using the estimated correction value includes, in the process of generating a correction value for a dynamic difference greater than a reference value, a step of first predicting the speed of the vehicle based on a vehicle model; a step of secondarily predicting the speed of the vehicle based on IMU (Inertial Measurement Unit) sensor values ​​regarding the movement of the vehicle; and a step of estimating speed information of the vehicle by comparing the first and second predicted speeds of the vehicle using an Extended Kalman Filter, wherein the step of estimating a correction value for a dynamic difference greater than a reference value and correcting the estimated position of the vehicle using the estimated correction value includes a step of correcting the position of the vehicle based on the estimated speed information. Claim 10 delete Claim 11 delete Claim 12 A method for precise position estimation of an autonomous driving platform according to claim 9, wherein the step of secondarily predicting the speed of the vehicle based on IMU (Inertial Measurement Unit) sensor values ​​regarding the movement of the vehicle includes the step of secondarily predicting the longitudinal speed or lateral speed of the vehicle based on acceleration and angular velocity information of the vehicle as the IMU (Inertial Measurement Unit) sensor values. Claim 13 A method for precise position estimation of an autonomous driving platform, characterized in that, in claim 12, the step of secondarily predicting the speed of the vehicle includes the step of secondarily predicting the longitudinal speed or lateral speed of the vehicle based on the IMU (Inertial Measurement Unit) sensor value and the odometry information of the vehicle. Claim 14 A method for precise position estimation of an autonomous driving platform according to claim 9, wherein the step of collecting GPS signals to estimate the position of a vehicle includes the step of estimating the position of the vehicle by using an Extended Kalman Filter on the collected GPS signals, calculating the error covariance with the predicted value of the next step based on the initial value of the collected GPS signals, calculating the Kalman gain which is the weight of the estimation value calculation formula to determine the ratio of the observed value to the predicted value, and then calculating the estimated value to estimate the position of the vehicle. Claim 15 In claim 9, the step of estimating a correction value for a dynamic difference greater than a reference value and correcting the estimated position of the vehicle using the estimated correction value comprises estimating the speed of the vehicle using an Extended Kalman Filter on the vehicle's IMU (Inertial Measurement Unit) sensor values ​​and odometry information, calculating the error covariance with the predicted value of the next step based on the initial values ​​of the vehicle's IMU (Inertial Measurement Unit) sensor values ​​and odometry information, calculating the Kalman gain which is the weight of the estimation value calculation formula to determine the ratio of the observed value to the predicted value, and then calculating the estimated value to estimate the speed of the vehicle.

Images