Control parameter automatically adjusting system and autonomous travel assisting device

Abstract

To promptly accomplish an operation start for an autonomous travel system by efficiently adjusting a parameter relating to travel control for a movable body.SOLUTION: While a movable body (a vehicle 101) is repeatedly traveling through predetermined routes (broad routes 401 and 402), a parameter updating unit 306 learns a relation between a control parameter (a weighting matrix Q, R) and one evaluation index selected from a plurality of evaluation indices c1, c2 and c3, and also learns a relation between the control parameter and other evaluation indices other than the one evaluation index among the plurality of evaluation indices c1, c2 and c3.SELECTED DRAWING: Figure 3

Description

[Technical field]

[0001] The present invention relates to an automatic adjustment function for control parameters, aimed at realizing safe and efficient autonomous driving of moving objects such as vehicles, and to an autonomous driving control technology utilizing the same. [Background technology]

[0002] With the labor shortage caused by the declining birthrate and aging population, and the expansion of the e-commerce market, labor saving and improvement of work efficiency in logistics warehouses and factories have become issues. To solve these issues, autonomous vehicles (hereinafter referred to as vehicles) that can operate without a driver and autonomous driving systems that operate multiple vehicles efficiently are being introduced. In conventional autonomous driving systems, the range in which a vehicle can move is limited from the perspective of safety, and the movement and work ranges of people (such as field workers) and vehicles are generally different. On the other hand, with the development of autonomous driving technology in the robotics field, even if the movement lines of vehicles and people overlap, the vehicles themselves are able to determine the need to avoid or detour, and flexibly select actions according to the surrounding situation.

[0003] In autonomous vehicle driving control, there are various requirements, such as the time to reach the destination, energy consumption, and the tolerance for approaching obstacles. Therefore, when constructing a control function, it is necessary to adjust the control parameters based on these multiple requirements. Traditionally, the adjustment of control parameters was determined largely by trial and error by hand, requiring a great deal of man-hours. On the other hand, with the recent development of machine learning technology, there are technologies that prepare evaluation indices such as rewards corresponding to driving results, and automatically update and adjust the control parameters based on update rules that maximize or minimize the evaluation indices.

[0004] As a specific example, Patent Document 1 proposes a method of preparing an evaluation index that scores fuel efficiency and exhaust gas performance based on changes in the state quantity (speed, etc.) of a vehicle (automobile), and determining a control command that improves the value of the evaluation index, as well as a vehicle control device. By adopting the configuration of Patent Document 1, it is possible to automatically and appropriately adjust the vehicle control parameters. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Patent Publication No. 2020-56737 Summary of the Invention [Problem to be solved by the invention]

[0006] Here, we will take a deeper look at the requirements for autonomous vehicle driving. Taking the autonomous driving of a transport robot as an example, the transport robot is required to track a target trajectory and stop at the target position with as little error as possible. On the other hand, even in the same field environment, there may be multiple requirements for tracking accuracy with respect to the target trajectory and stopping accuracy with respect to the target position depending on the driving area and driving purpose. For example, focusing on the stopping operation of a vehicle, consider the following two cases.

[0007] (A) As a result of traffic control, the vehicle stops in an intersection area to allow other vehicles to proceed.

[0008] (B) The conveyor stops when a load is to be placed on or received from the conveyor. In general, when it comes to vehicle stopping control, there is a trade-off between the time to stop, traveling speed, and accuracy. In case (A), the vehicle only needs to avoid interfering with other vehicles crossing in front of the vehicle, and so quick stopping is required rather than stopping accuracy. More specifically, even if the vehicle stops slightly tilted from the target stopping direction, it will not affect other vehicles crossing in front. On the other hand, in case (B), if the vehicle does not stop precisely at the specified position and direction, the load conveyed from the conveyor will not be received at the expected position, which could lead to the load dropping, so stopping accuracy is prioritized over quick stopping.

[0009] In this way, the requirements that are prioritized for the autonomous driving of a vehicle change depending on the target operation and the driving environment. In the present invention, the above-mentioned prioritized requirements are called "driving strategies". The control parameter automatic adjustment technology, including the technology of Patent Document 1, prepares an evaluation index for each driving strategy and adjusts the parameters. More specifically, for the two previous cases, an evaluation index for a driving strategy that prioritizes the time to stop and an evaluation index for a driving strategy that prioritizes the stopping accuracy are prepared, and a predetermined number of runs are performed for both cases. In the previous case, only two elements, the time to stop and the accuracy, are exemplified, but there are various types of vehicle state quantities and upper and lower limits of required performance. Therefore, depending on the number of driving strategies to be realized, there are many evaluation indexes, and the number of runs for parameter adjustment becomes large. Therefore, it is considered that the configuration of Patent Document 1 has room for improvement in terms of the efficiency of parameter adjustment and the labor (time) required for adjustment.

[0010] The present invention has been made to solve the above problems, and aims to quickly realize the start of operation of an autonomous driving system by efficiently adjusting parameters related to the driving control of a moving object. [Means for solving the problem]

[0011] In order to achieve the above-mentioned object, the present invention provides an automatic control parameter adjustment system for adjusting control parameters of a moving body, the system comprising: a sensor for detecting a state quantity of the moving body; a target state setting unit for setting a target state quantity of the moving body; a control command generation unit for generating a control command for the moving body using the control parameters so as to reduce a deviation of the state quantity from the target state quantity; an evaluation index design unit for designing a plurality of evaluation indexes for calculating an evaluation value for a driving result of the moving body; and a parameter updating unit for learning the relationship between the plurality of evaluation indexes and the control parameters and updating the control parameters so as to improve the evaluation value calculated using the plurality of evaluation indexes, wherein the parameter updating unit learns the relationship between the control parameters and one evaluation index selected from the plurality of evaluation indexes while the moving body repeatedly travels a predetermined route, and learns the relationship between the control parameters and other evaluation indexes among the plurality of evaluation indexes other than the one evaluation index. Effect of the Invention

[0012] According to the present invention, it is possible to quickly start operation of an autonomous driving system by efficiently adjusting parameters related to the driving control of a moving object. [Brief description of the drawings]

[0013] [Figure 1] A conceptual diagram showing an example of a driving environment inside a warehouse or factory [Figure 2A] Conceptual diagram of the vehicle [Figure 2B] Simplified functional block diagram of the vehicle and on-board controller [Diagram 3] Example of functional configuration of control parameter automatic tuning system [Figure 4] A conceptual diagram showing the overall route for the driving environment in Figure 1. [Diagram 5] Conceptual diagram showing the relationship between global and local paths toward autonomous vehicle driving [Figure 6] A function expressing the relationship between control parameters and evaluation indices using Bayesian optimization [Figure 7]A conceptual diagram showing how the function of multiple driving strategies is updated in parameter updating by Bayesian optimization, which is a feature of the present invention. [Figure 8] A flow chart showing the process of adjusting the weight parameters of model predictive control using the control parameter automatic tuning system. [Figure 9] Example of circular path combining global paths for control parameter tuning [Figure 10] A conceptual diagram showing how the circular route in Figure 9 is changed according to the progress of parameter adjustment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0014] Hereinafter, an embodiment of the present invention will be described in detail with reference to Fig. 1 to Fig. 10. Note that in this embodiment, an explanation will be given of inter-process transport in a warehouse or factory as an example.

[0015] <Driving Environment 100> FIG. 1 illustrates a travel environment 100 in a warehouse or factory assumed in this embodiment. As an example, a trolley-type transport vehicle 101 (hereinafter, vehicle 101) is used for the transport work. Alternatively, a towing vehicle or a forklift may be used. The travel environment is an environment with a series of narrow roads surrounded by shelves 102 and the like. There may be loads 103 on the shelves 102, pallets 104 temporarily placed on the route, and loads on the pallets 104. The environment is also assumed to include a vehicle charging unit 105, a conveyor 106 transporting loads from other areas, and the like.

[0016] <Vehicle 101> A conceptual diagram of a vehicle 101 used in this embodiment is shown in Fig. 2A. In this embodiment, a differential two-wheel vehicle is used as an example, but as mentioned above, the present invention is not limited to this and can handle moving bodies of various drive modes such as omni-wheel type, forklifts, and four-wheel vehicles. The equation of motion of a differential two-wheel vehicle can be given by Equation 1.

[0017]

number

[0018] Here, x is the x coordinate of the vehicle 101, y is the y coordinate of the vehicle 101, θ is the heading (direction) of the vehicle 101, v is the traveling speed of the vehicle 101, and ω is the angular velocity of the vehicle 101. Note that in the control design of the vehicle 101, we deal with Equation 2, which is obtained by discretizing Equation 1 with a sampling period Δt. Note that k in Equation 2 means a processing step (time).

[0019]

number

[0020] <Functional blocks of the vehicle 101 and the in-vehicle controller 201> In order to explain the control operation of the vehicle 101, a simplified functional block diagram of the vehicle 101 and the on-board controller 201 is shown in FIG. 2B. The on-board controller 201 that performs control calculations is installed in the vehicle 101. The vehicle 101 also includes a drive unit 207 and an encoder 208. The on-board controller 201 includes a recording unit 202, an on-board sensor 203, a driving map management unit 204, a self-position calculation unit 205, and a control command generation unit 206.

[0021] The recording unit 202 is a recording device for recording the results of arithmetic processing provided in the in-vehicle controller 201, and corresponds to an HDD (hard disk drive), an SSD (solid state drive), or a memory. Various control programs and parameters used in the control programs are recorded in the recording unit 202.

[0022] The on-board sensor 203 corresponds to an external sensor mounted on the vehicle 101, and in this embodiment, as an example, a LiDAR (Light Detection And Ranging) is mounted. The LiDAR is a sensor that measures the distance to an object present within the irradiation range while changing the irradiation angle of a laser beam.

[0023] The driving map management unit 204 manages a driving map (grid map) showing obstacle information in the driving environment 100. The map is created in advance before the start of operation using SLAM (Simultaneous Localization and Mapping) with input of measurement data during driving acquired from the on-board sensor 203 (LiDAR) and vehicle travel distance information acquired from the encoder 208 described later. Since the creation of a grid map using SLAM is a known technique, a description thereof will be omitted.

[0024] The self-position calculation unit 205 combines the measurement results obtained from the on-board sensor 203 and the encoder 208 to calculate the self-position (x, y coordinates on a two-dimensional plane) and orientation of the vehicle itself. Specifically, the self-position and orientation of the vehicle 101 are acquired using a map matching technique. Map matching is a technique that takes the measurement information (point cloud set) acquired by the on-board sensor 203 and the driving map as input, and estimates the vehicle position on the driving map by utilizing the similarity between the point cloud set and the driving map. For example, there is an Adaptive Monte Carlo Localization method that utilizes a particle filter. Since the self-position estimation technique using map matching is also a publicly known technique, a detailed explanation will be omitted.

[0025] The control command generating unit 206 performs calculations related to driving control of the vehicle 101. The details of the calculations will be described later.

[0026] The driving unit 207 converts the control command input from the control command generating unit 206 into a predetermined current value in order to transmit power to the driving wheels of the vehicle 101. In addition, an encoder 208 provided on the driving wheels acquires the amount of rotation of the driving wheels and estimates the travel distance and current speed of the vehicle. The estimated value is transmitted to the control command generating unit 206, the self-position calculating unit 205, and the recording unit 202 as a response value from the driving unit 207. The above is an overview of the functions provided in the vehicle 101 and the in-vehicle controller 201.

[0027] In this embodiment, a method and system are proposed for efficiently adjusting control parameters set in a control command generator 206 that satisfy the target performance of each driving strategy set for each driving area and target operation in a vehicle 101 operating in a driving environment 100 shown in Fig. 1. Hereinafter, an overview of the control parameter automatic adjustment system will be described.

[0028] <Control parameter automatic tuning system 300> FIG. 3 is a functional block configuration example of the control parameter automatic adjustment system 300. The control parameter automatic adjustment system 300 in this embodiment includes an on-board controller 201 and a vehicle 101. In addition, as functional blocks not shown in FIG. 2B, a target position setting unit 301, a global route planning unit 302, a local route planning unit 303, an evaluation index design unit 304, a main evaluation index setting unit 305, a parameter update unit 306, and a driving result management unit 307 are added. In FIG. 3, lines with arrows represent data flows. In addition, solid lines represent data transmitted at each control cycle, and dotted lines represent data transmitted at a certain timing, such as before the system is operated or after the vehicle 101 has reached a predetermined position. In the following, an overview of each functional block newly added in FIG. 3 will be described.

[0029] <Target position setting section 301> The target position setting unit 301 sets the task currently being performed by the vehicle 101 or a target position for task execution (coordinates on the driving map) based on the driving map managed by the driving map management unit 204. As an example, the target position may be directly input by an operator via an external device (such as a mobile display) that displays the vehicle state or the driving map, or, although not shown in Fig. 3, the target position may be automatically calculated from the operating state of a task set in advance on the map distribution server side.

[0030] <Global route planning unit 302> The global route planning unit 302 receives the driving map managed by the driving map management unit 204, the vehicle's own position (current position) calculated by the self-position calculation unit 205, and the target position set by the target position setting unit 301 as inputs, and performs global route planning from the own position to the target position. The global route is generated based on the driving map expressed as a lattice. There are many existing methods for planning the global route, such as the Dijkstra algorithm and the A* (A star) algorithm, so the explanation will be omitted here. The generated global route is expressed as a set of coordinate values ​​(such as row and column numbers) of each lattice included in the driving map, and arrival at the target position is guaranteed by traveling in the order of the coordinate values. In addition, a target traveling speed is set for the coordinate values ​​of each lattice constituting the global route. There are various possible policies for setting the target traveling speed, and it may be set according to the curvature in each section of the global route, or it may be set according to the size of the area to be traveled or the speed limit set for on-site operation. In addition, the target travel speed may be set in consideration of the vehicle's motion characteristics by making the method of planning the global route a method that takes into account the vehicle motion model. Routes 401 and 402 shown in FIG. 4 show the generated global routes. The target travel speed of the global route 401 may be set higher since the global route 401 travels linearly in an area with few obstacles around. On the other hand, the target travel speed of the route 402 may be set lower than that of the global route 401 since the route 402 travels in a narrow road surrounded by shelves with turns. The target travel speed may be set to a constant value for the entire route, or may be set individually for each grid (each coordinate value) that constitutes the global route. The target travel speed set here is one element when considering a travel strategy, which will be described later.

[0031] <Local route planning unit 303> The local path planning unit 303 receives the global path calculated by the global path planning unit 302 and the self-position as input, sets the lattice coordinates of the global path that exists, for example, a distance D [m] ahead of the self-position as a local target position, and generates a local path to the local target position. The local path is not a coordinate value indicating the global path, but may be a set of finer coordinate values ​​or a function expressing a curve, and is a smoother path than the global path.

[0032] In this embodiment, a curve obtained by approximating the global path with a cubic function is used. The local path 501 shown in FIG. 5 is obtained by approximating the corresponding coordinate group of the global path 401 with a cubic function. The points on the global path 401 indicate the center positions of the grids, and their coordinate values ​​(x r ,y r ) with the target driving speed v r The local path 501 can be expressed as in Equation 3 by inputting a set of coordinate values ​​of the global path 401. s ,y s are coordinate values ​​on the global route transformed into a vehicle coordinate system with the vehicle's 101 own position as the origin.

[0033]

number

[0034] <Control command generating unit 206> The control command generation unit 206 receives as input the local route planned by the local route planning unit 303 and the vehicle's own position, and generates a control command for the drive unit 207 equipped in the vehicle 101 according to a route tracking algorithm so as to pass along the local route as much as possible.

[0035] In this embodiment, the path tracking algorithm uses a framework of MPC (Model Predictive Control). MPC minimizes an evaluation function set in advance in a section (vehicle behavior from the current time to N steps ahead) for each control period, and obtains a control command to be input to the vehicle 101.

[0036] The evaluation function is the vehicle position at each time and the lateral deviation c e , azimuth deviation θ e , deviation from the target running speed v e The deviation vector e k Define the deviation vector e k The function to evaluate (the first term in the right-hand side of Equation 4) and the control command input u k =(v k ,ω k) is constructed using a function to evaluate the magnitude of

[0037]

number

[0038] In addition, T in Equation 4 is a symbol representing transposition. Also, the matrix Q is the deviation vector e k The matrix R is the weighting matrix (weighting parameter) related to the control command u k and N is a parameter representing the evaluation interval (prediction interval). The weighting matrices Q and R are generally expressed as diagonal matrices as shown in Equation 5.

[0039]

number

[0040] However, it should be noted that the present invention can also be applied to the case where off-diagonal elements are included in the weighting matrices Q and R. Thus, the control command generator 206 includes a plurality of weighting parameters, and in this embodiment, these weighting parameters are successively updated and optimized according to each driving strategy.

[0041] Deviation vector e k The initial value c in e (0),θ e (0),v e (0) is obtained from the local path and global path. The lateral deviation is expressed as the intercept term d of the cubic function (Equation 3) expressed as the local path, as shown in Equation 6.

[0042]

number

[0043] The azimuth deviation is the coordinate (x r ,y r) with respect to the tangent direction of the global path 401, and the coordinate (x r ,y r ) and the Nr-1 coordinate (x r+Nr-1 ,y r+Nr-1 ) and the azimuth angle θr of the line segment connecting them (Equation 7).

[0044]

number

[0045] Here, the coordinates on the global route nearest to the current self-position are (x r ,y r The speed deviation is expressed as the target running speed v r+Nr and the current speed v. In addition, MPC needs to handle the change in vehicle behavior in the evaluation section shown in N, so the deviation vector e k It is also necessary to consider the changes within the evaluation section for . Therefore, using the vehicle motion shown in Equation 2 and the local path shown in Equation 3, the changes in position deviation and heading deviation within the evaluation section N are expressed as Equation 8 and Equation 9, respectively.

[0046]

number

[0047]

number

[0048] In addition, d / ds in Equation 8 indicates the first derivative of the cubic function that indicates the local path. Regarding the change in speed deviation, the running speed v k and the target driving speed v r+Nr This deals with the difference between (see equation 4).

[0049] In addition, MPC can easily handle constraint conditions. For example, in the case of the vehicle 101 (differential two-wheel vehicle) used in this embodiment, the characteristics of the drive unit 207, including the vehicle 101's running speed v and the upper and lower limits of the turning angular velocity ω, can be incorporated as constraint conditions. This makes it possible to generate control commands that the vehicle 101 can actually achieve. An MPC that takes such constraint conditions into consideration can be expressed as in equation 10.

[0050]

number

[0051] In addition, [st] in Equation 10 is an abbreviation of [subject to], and means that the optimization problem is solved under the conditions after [st]. Also, the subscripts max and min represent the control command u k Each element of (velocity v k , angular velocity ω k ) corresponds to the upper and lower limit values. In addition, the constraint conditions include the formula 2, the formula 8, and the formula 9. Note that the MPC in this embodiment is configured to obtain a control command u that minimizes the evaluation function J, but it may also be configured to set a target value for the evaluation function J and obtain a control command u that falls below the target value.

[0052] <Parameter adjustment (characteristic part of the present invention)> In MPC, even a slight change in the weight parameters (weight matrices Q, R, etc.) may significantly change the actual vehicle operation, so designing the weight parameters is very important. On the other hand, as the number of weight parameters increases, the number of combinations increases, and so labor is required to adjust the values. In this embodiment, a case is dealt with in which the above-mentioned weight parameters are adjusted more efficiently for the multiple driving strategies mentioned at the beginning. Specifically, a case is dealt with in which, even for the same global route, a situation is assumed in which the allowable maximum lateral deviation, directional deviation, and deviation of the target driving speed are different, and an optimal weight parameter corresponding to each situation (driving strategy) is efficiently obtained.

[0053] The proposal of the present invention can also be applied to global routes that are planned each time according to the destination during actual operation of an autonomous driving system. However, in this embodiment, for the sake of simplicity, the following description focuses on global routes 401 and 402 as the target global routes.

[0054] <Evaluation Index Design Department 304> In recent years, a method has been devised for adjusting control parameters such as the weight parameters of MPC using Bayesian Optimization (BO). BO is an algorithm that searches for parameters that maximize or minimize a specific evaluation index c by repeating simulations and experiments.

[0055] In this embodiment, multiple evaluation indexes c are set based on various driving strategies, and need to be described by some mathematical formula, similar to the evaluation function J of the MPC. However, unlike the evaluation function of the MPC, it is not necessary to use state quantities or control commands, and it is also possible to use, for example, the time it takes for the vehicle to reach a target driving position as the evaluation index c. In parameter adjustment of the MPC by BO, a vector ρ that combines the diagonal elements of the weighting matrices Q and R that correspond to the adjustment parameters is given by Equation 11.

[0056]

number

[0057] As an example of the evaluation index c, the time it takes for the vehicle to reach the target position t end Then, the optimal ρ that minimizes this value is * An optimization problem such as the one shown in Eq. 12 can be given as a problem for finding

[0058]

number

[0059] In general, this optimization problem cannot be solved directly. In BO, several solution candidates ρ iWhile checking the value of the evaluation index c actually obtained by applying i FIG. 6 is a diagram showing a schematic diagram of this solution candidate search method. The black circles in the diagram represent the solution candidates ρ i (i=1...6) is the evaluation value obtained. Using the solution candidates obtained so far, it is possible to calculate the estimated value of the function shape (solid line in the figure) and its variance (dashed line in the figure). The variance is the solution candidate ρ i The next solution candidate ρ 7 (White triangle in the figure) is determined. By performing the above process, the solution candidate ρ i It is known that this method allows for very efficient parameter search (weight design) compared to brute force verification.

[0060] The evaluation index design unit 304 in this embodiment designs a plurality of evaluation indexes c. In this embodiment, evaluation indexes corresponding to the following three driving strategies are prepared.

[0061] (1) Emphasis on lateral deviation An example of an evaluation index corresponding to a driving strategy that places importance on lateral deviation is shown in Equation 13.

[0062]

number

[0063] In narrow or complicated areas such as area 112 in FIG. 1, if the lateral deviation amount becomes large with respect to the target overall route, it will lead to interference with surrounding obstacles. Therefore, the evaluation index c1 shown in formula 13 is set to a smaller maximum allowable value for the lateral deviation amount compared to other evaluation indexes described later. There are two types of lateral deviations to be evaluated, and the average value e of the lateral deviations obtained in each control cycle is ave_c and the maximum lateral deviation e determined from the results up to the end of the run max_c Here, the maximum lateral deviation e max_cWhen this occurs, the evaluation index value (hereafter referred to as the driving score) is designed to be significantly worsened (increased). t1 ,W c1 indicates the weight parameters of various evaluation items, which are values ​​that are set in advance. Although these weights themselves are also adjustment items, the number is small compared to the number of parameters of the weight matrices Q and R that you ultimately want to adjust, and since the conditions that must be achieved (required performance) can be easily expressed as a formula, it is easier to design than a comprehensive search for the weight matrices Q and R.

[0064] (2) Emphasis on azimuth deviation An example of an evaluation index corresponding to a driving strategy that places importance on the heading deviation is shown in Equation 14.

[0065]

number

[0066] When transporting a load, if the direction deviation becomes large and the vehicle behaves in a meandering manner with respect to the target overall route, this can lead to excessive shaking of the transported load, which may cause the load to collapse or fall. Therefore, the evaluation index c2 shown in formula 14 is designed to worsen the travel score according to the average direction deviation compared to other evaluation indexes (see the third term of evaluation index c2).

[0067] (3) Emphasis on driving time An example of an evaluation index corresponding to a driving strategy that places importance on driving time is shown in Equation 15.

[0068]

number

[0069] Even in an environment with few obstacles around, such as in area 111 in Fig. 1, if there is a material handling device such as a conveyor 106 at the destination, if the item does not arrive at the material handling device within a specified time, the item will be stuck on the conveyor 106, lowering the operational efficiency of the entire warehouse. Therefore, using the same design guidelines as in (1), an upper limit is set for the difference from the target running time, and if this upper limit is exceeded (if the item arrives late), the running score is designed to deteriorate significantly. However, the lateral deviation from the target route is also evaluated, and parameters that achieve both route tracking accuracy and running speed are found.

[0070] The feature of the present invention is that, as described above, multiple evaluation indexes are set, and when updating parameters by BO for any evaluation index, the relationship between the input weight parameters and driving scores for other evaluation indexes is calculated at the same time, and the function shapes of the evaluation indexes for other driving strategies are also updated. Details will be explained below.

[0071] <Main evaluation index setting unit 305> The main evaluation index setting unit 305 sets a main evaluation index to be used for actual parameter update among the multiple evaluation indexes designed by the evaluation index design unit 304. The selection policy of the main evaluation index may be random, or may be selected intentionally by the on-site operator. For example, a policy of preferentially selecting an evaluation index for which it is difficult to obtain a good driving score as the main evaluation index is considered. This is because the more difficult it is to obtain a good driving score for an evaluation index, the more likely it is that the number of trials will increase, and therefore the mechanism of the parameter update unit 306 described in detail below increases the possibility of obtaining a good driving score for other evaluation indexes. As an example, among the various evaluation indexes shown in Equation 13, Equation 14, and Equation 15, the evaluation index c1 shown in Equation 13 is designed to lead to a deterioration in the score even with a small lateral deviation compared to the other indexes, so it is determined to be the index that is most difficult to adjust, and in this embodiment, the evaluation index c1 is set as the main evaluation index.

[0072] <Parameter update section 306> The parameter update unit 306 receives the change in state quantity associated with the current run recorded in the recording unit 202, the function shape of the BO recorded in the running result management unit 307 described later, the weight parameters implemented in the past, and the running score at that time as input, and obtains the weight parameters to be input in the next run according to the update rules of the BO. Note that here, the update rules of the BO are also applied to evaluation indexes other than the main evaluation index (hereinafter, other evaluation indexes). FIG. 7 is a diagram showing this situation diagrammatically. This situation is the focus of the present invention, and the function shape of the evaluation index c corresponding to each running strategy is updated by inputting the relationship between the weight parameters and the running score in addition to the main evaluation index in FIG. 7. In the case of the original BO, the flow in which the next parameters are proposed according to the estimated function shape is common, but in the present invention, the weight parameters updated according to the rules of the main evaluation index are only applied to the function shape update of the other evaluation indexes. As a result, the other evaluation indexes are different from the flow in which the optimal parameters (next parameters) estimated from the current function shape are obtained. Specifically, in the situation shown in Figure 7, if the original BO was used, the other driving strategy (evaluation index c2) would be ρ' 7 is output as the next parameter candidate, but here, ρ 7 is the next parameter. Therefore, the other driving strategies will not be updated to efficiently estimate the function shape, but since the relationship between the weight parameters and the driving score has already been estimated in advance when the other evaluation index itself becomes the main evaluation index, it is highly likely that the parameters that lead to a worsening of the score will not need to be updated. By using the above mechanism, the number of trials (number of runs) of the other driving strategies can be reduced, and the labor required for parameter adjustment can be reduced.

[0073] The newly proposed weighting parameters for the main evaluation index based on the update rules of BO are used the next time the vehicle travels along a similar global route, and are transmitted to the control command generator 206 .

[0074] The above is an explanation of each functional block constituting the control parameter automatic adjustment system 300 proposed in the present invention. Hereinafter, the actual flow of parameter automatic adjustment will be explained according to the flowchart shown in FIG.

[0075] <Driving environment during automatic parameter adjustment> In explaining the flow of the process of automatic parameter adjustment, in this embodiment, the description will be focused on the global routes 401 and 402 in FIG. 4. The accuracy of the self-position calculation unit 205 changes somewhat depending on the shape of surrounding obstacles, and the global route is also generated for each destination during on-site operation. Therefore, it is desirable to adjust the weight parameters each time for each global route generated during on-site operation shown in FIG. 1 and FIG. 4. However, especially in the initial stage of adjusting the weight parameters, parameters that lead to excessive meandering or lateral deviation may be tried, so there is a risk of interference with surrounding obstacles when traveling on a narrow road between the shelves 102 in the area 112 in FIG. 1. Therefore, as an example, a mechanism is considered in which the distance to the surrounding obstacles is measured using an on-board sensor (LiDAR) mounted on the vehicle 101, and the control command generation unit 206 generates a control command to avoid or decelerate as necessary. On the other hand, in such a case, it is also necessary to consider the design of an evaluation index when avoidance or deceleration occurs. Therefore, in order to focus on the main points of the present invention, in this embodiment, a travel area 901 is separately prepared in which a circular route 902 can be formed by connecting the global routes 401, 402, as shown in Fig. 9. The dotted line part of the circular route 902 indicates a buffer route 903, and the results while traveling on the buffer route 903 are not used for evaluation, and since a low speed is acceptable, it is possible to realize traveling that follows the center of the route as much as possible. Note that the black diamond and black circle that connect the routes in the figure indicate the travel start position and the target value position (end position), respectively.

[0076] <Flow of automatic parameter adjustment process> When the control parameter automatic adjustment system 300 starts the process, it acquires the own position of the vehicle 101 (process 801 in FIG. 8). Here, the process described in the own position calculation unit 205 is executed.

[0077] Next, a travel target position of the vehicle 101 is set (process 802). Here, the process described in the target position setting unit 301 is executed. In this embodiment, since the circular route 902 (including the general routes 401, 402) shown in Fig. 9 has already been prepared, for example, the end position of the general route 401 is set as the target position here.

[0078] Next, a global route to the set driving target position is generated (process 803). The contents of this process are as described in the process outline of the global route planning unit 302. However, in this embodiment, since the global route 401 has already been prepared, a coordinate value group of the global route 401 is set.

[0079] Next, the update parameters are obtained (process 804). In this embodiment, the driving result management unit 307 manages the driving results (function shapes expressing the relationship between weight parameters and driving scores) of the global routes 401 and 402. Here, the weight parameters updated from the driving results of the global route 401 are transmitted to the control command generation unit 206 via the parameter update unit 306 (parameter update by BO). Note that, when driving for the first time, if there are no past driving results, the parameter update unit 306 may use parameters randomly selected from the search space of various weight parameters set by the evaluation index design unit 304.

[0080] Next, evaluation indexes are set (process 805). Here, various evaluation indexes (Equations 13, 14, and 15) designed by the evaluation index design unit 304 and the main evaluation index (Equation 13) set by the main evaluation index setting unit 305 are set.

[0081] After this, actual autonomous driving starts. First, a local route is generated (process 806). Here, the process described in the local route planning unit 303 is executed. Next, the vehicle 101's own position is acquired (process 807). The own position is acquired here in order to acquire the change in the vehicle's own position while the vehicle is traveling and to calculate a control command.

[0082] Next, a path tracking control process is performed (process 808). Here, the process described in the control command generating unit 206 is executed. Note that the values ​​acquired in process 804 are used as various weight parameters of the MPC used in this process.

[0083] Next, various state quantities during vehicle travel are recorded (process 809). Here, values ​​necessary for calculating various evaluation indexes, such as lateral deviation from the local route, directional deviation, deviation from the target travel speed, and elapsed time from the travel start time, are acquired. The acquired values ​​are recorded in the recording unit 202.

[0084] Next, it is determined whether the vehicle has reached the travel target position (process 810). For example, if the self-position acquired in process 807 is within a radius of 0.1 m from the coordinate value indicating the travel target position set in process 802, a stop command (speed 0 m / s command) is sent to the vehicle, the vehicle ends its travel, and the process proceeds to process 811. On the other hand, if the travel target position has not been reached, the process returns to process 806, and processes 806, 807, 808, and 809 are repeatedly executed until the target position is reached.

[0085] In the next process 811, the values ​​recorded in process 809 are input and the evaluation values ​​of the multiple evaluation indexes are obtained. After that, the parameters to be tried in the next run are obtained only for the main evaluation index according to the update rules of BO. Note that for the indexes other than the main evaluation index, only the function shapes are updated according to the rules of BO.

[0086] Next, it is confirmed whether the adjustment of the weight parameters is completed (process 812). Specifically, it is determined whether the adjustment of the weight parameters of the main evaluation index for the global route 401 is completed. There are several possible conditions for the determination. For example, the evaluation index design unit 304 may set a target score for each evaluation index, and if a score equal to or less than the target score (or equal to or more than the target score in some cases) is obtained in this trial, it may be determined that the adjustment is completed. Alternatively, the average and variance of the driving scores for the past several times managed by the driving result management unit 307 may be obtained, and if the average value is equal to or less than the target score and the variance value is relatively small, it may be determined that the adjustment is completed. Alternatively, the evaluation index design unit 304 may simply set a maximum number of trials, and if the number of trials is reached, it may be determined that the adjustment is completed. If it is determined that the adjustment of the global route 401 is completed, the process proceeds to process 813.

[0087] On the other hand, if the adjustment is not completed, the process returns to process 801. In the case of this embodiment, since automatic parameter adjustment is performed along the circular route 902 in Fig. 9, the vehicle 101 should be at the end point of the global route 401 at this time. After this, since the vehicle 101 will enter the buffer route 903, the parameter adjustment is temporarily stopped, and the autonomous driving process except for process 809 is performed to move to the driving start position of the global route 402. After the vehicle 101 has reached the driving start position of the global route 402, processes 801 to 812 are performed again.

[0088] In process 813, it is confirmed whether parameter adjustment for all driving strategies is completed. Specifically, it is confirmed whether an end judgment has been made for all evaluation indexes set by the evaluation index design unit 304. If adjustment of all evaluation indexes is completed, the process proceeds to process 814.

[0089] On the other hand, if adjustment has not been completed for all evaluation indexes, the process proceeds to process 815. In process 815, the main evaluation index is changed (processing of the main evaluation index setting unit 305 is performed). If there is only one evaluation index left for which adjustment has not been completed, the relevant evaluation index is changed. The change policy when there are multiple evaluation indexes is as described in the explanation of the main evaluation index setting unit 305, and one possible policy is to switch to an evaluation index for which it is difficult to obtain a good score. Also, if the adjustment end condition for another driving strategy happens to be met, the adjustment of that driving strategy is also determined to be complete.

[0090] Next, in process 814, it is confirmed whether the adjustment of the weight parameters for all global paths is completed. If it is completed, all the processes of the control parameter automatic tuning system 300 are completed. If there is a global path for which the adjustment has not been completed, the process proceeds to process 816.

[0091] In process 816, the layout of the global route is changed (however, this is optional). Specifically, in the circular route 902 shown in Fig. 9, for example, when adjustments for all evaluation indexes of the global route 402 have been completed, a circular route 1001 consisting of only the global route 401 may be set as shown in Fig. 10. This allows parameter adjustments for the remaining evaluation indexes to be achieved in a shorter time. However, since the above-mentioned change in the layout of the global route may not be possible in the actual operation environment shown in Figs. 1 and 4 due to the influence of obstacles in the environment, process 816 is not necessarily a process that can be performed, and should be determined according to the situation at the site.

[0092] The above is the content of the autonomous driving system utilizing the control parameter automatic adjustment system 300 proposed in the present invention.

[0093] <Other use cases> In this embodiment, as an example, multiple driving strategies are established for the path following function, and corresponding evaluation indices are designed, but the use case is not limited to this, and it can be used for, for example, the stopping control problem described at the beginning of the specification. The BO exemplified in this embodiment is known as one of the metaheuristic techniques that does not depend on a specific calculation problem in the algorithm for combinatorial optimization problems. If the evaluation index to be realized can be expressed as a formula, and there is a mechanism for measuring the state quantity from which the value of the evaluation index can be obtained from the vehicle, it can be applied to any control problem / algorithm, regardless of the MPC framework exemplified in this embodiment.

[0094] In this embodiment, the evaluation index is designed with a focus on the arrival time to the destination and the lateral deviation from the target route, but the evaluation index is not limited to this and may be designed with a focus on the energy consumption of the vehicle, for example. In this case, it is possible to add a term indicating the energy consumption, specifically, the integral value of the speed and acceleration, to the evaluation functions shown in Equation 13, Equation 14, and Equation 15.

[0095] In addition, the control parameter automatic adjustment system 300 described in this embodiment aims to speed up the start of operation of the autonomous driving system by repeatedly driving the vehicle before the autonomous driving system starts operating. On the other hand, the application method of the present invention is not limited to this, and may be implemented, for example, during the operation of the autonomous driving system. Specifically, in the case of the inter-process transport work described in this embodiment, it is considered that the vehicle transports loads of various weights and shapes. These changes in loads often affect the vehicle's motion dynamic characteristics, leading to changes in the driving score of the evaluation index. Therefore, for example, when the rough weight of the transported load can be measured, the relationship between the weight of the transported load and the change in the driving score of the evaluation index may be learned using BO.

[0096] <Effects of this embodiment> In this embodiment, the problem of setting multiple driving strategies and finding control parameters that satisfy each driving strategy in autonomous driving control of a vehicle is addressed. As an invention proposal, a technology is proposed in which, when learning the relationship between the evaluation index (driving score) of the main driving strategy and the control parameters, the relationship between the evaluation index and control parameters of other driving strategies is also learned based on the update rule of the evaluation index of the main driving strategy.

[0097] As a result, when the other driving strategy itself becomes the primary evaluation index, the parameters that lead to a deterioration in the evaluation index can be learned in advance, so there is a high possibility that the number of runs (trials) required for parameter adjustment can be reduced. This makes it possible to efficiently adjust parameters related to the driving control of the vehicle (moving body), leading to a speedier start of operation of the autonomous driving system.

[0098] <Summary> In this embodiment, a control parameter automatic adjustment system 300 that adjusts the control parameters (weighting matrices Q, R) of a moving body (vehicle 101) includes a sensor (on-board sensor 203, encoder 208) that detects a state quantity of the moving body, a target state setting unit (target position setting unit 301, global route planning unit 302) that sets a target state quantity of the moving body, and a deviation e of the state quantity from the target state quantity. k The control command u of the moving object is calculated using the control parameters so that k the control command generation unit 206 that generates a control command for generating a plurality of evaluation indexes c1, c2, c3 for calculating an evaluation value (driving score) for the driving result of the moving body; and a parameter updating unit 306 that learns the relationship (function shape) between the plurality of evaluation indexes c1, c2, c3 and the control parameters and updates the control parameters so as to improve the evaluation value calculated using the plurality of evaluation indexes c1, c2, c3. While the moving body repeatedly travels along a predetermined route (global routes 401, 402), the parameter updating unit 306 learns the relationship between the control parameters and one evaluation index selected from the plurality of evaluation indexes c1, c2, c3, and also learns the relationship between the control parameters and other evaluation indexes among the plurality of evaluation indexes c1, c2, c3 other than the one evaluation index.

[0099] According to the present embodiment configured as described above, parameter adjustment related to the travel control of a moving object can be efficiently performed, thereby making it possible to quickly start operation of an autonomous travel system.

[0100] Moreover, the control parameter automatic adjustment system 300 in this embodiment includes a recording unit 202 that records the time-series changes in the state quantities and the control parameters. This makes it possible to perform offline parameter adjustment targeting a separately considered driving strategy (evaluation index) by using the time-series changes in the state quantities of the vehicle 101 and the control parameters recorded in the recording unit 202.

[0101] In addition, the evaluation index design unit 304 in this embodiment sets a target value (target score) for each of the multiple evaluation indexes c1, c2, and c3, and the parameter update unit 306 ends the update of the control parameters for the evaluation indexes that have reached the target value among the multiple evaluation indexes c1, c2, and c3. This makes it possible to efficiently adjust the parameters for the multiple evaluation indexes c1, c2, and c3.

[0102] The control parameter automatic adjustment system 300 in this embodiment includes a main evaluation index setting unit 305 that sets, as a main evaluation index, an evaluation index that is expected to have the greatest number of runs on the specified route required until the update of the control parameters is completed, from among the multiple evaluation indexes c1, c2, c3, based on the function shapes of the multiple evaluation indexes c1, c2, c3, the constraint conditions, and the target values, and the parameter update unit 306 prioritizes the update of the control parameters targeted at the main evaluation index. This reduces the number of trials (runs) targeted at other driving strategies, making it possible to reduce the man-hours required for parameter adjustment.

[0103] In addition, the control command generator 206 in this embodiment calculates the deviation e of the state quantity from the target state quantity. k and the control command u k and deviation e k and the control command u kThe control command u is set so that the value of the evaluation function J defined by the weighting parameters (weighting matrices Q, R) for k and the parameter update unit 306 updates the weighting parameters as the control parameters. This makes it possible to efficiently adjust the parameters related to the travel control of the moving object by the MPC, thereby quickly realizing the start of operation of the autonomous travel system.

[0104] The control parameter automatic adjustment system 300 in this embodiment may be provided in an autonomous driving assistance device that realizes functions such as an advanced driver assistance system (ADAS).

[0105] The present invention is not limited to the above-described embodiment, but includes various modified examples. For example, the above-described embodiment is shown for the purpose of explaining the present invention in an easy-to-understand manner, and the present invention is not necessarily limited to the embodiment having all of the described configurations. [Explanation of symbols]

[0106] 100...driving environment, 101...vehicle (moving object), 102...shelf, 103...load, 104...pallet, 105...charging unit, 106...conveyor, 111, 112...area, 201...vehicle controller, 202...recording unit, 203...vehicle sensor, 204...driving map management unit, 205...self-position calculation unit, 206...control command generation unit, 207...driving unit, 208...encoder (sensor), 300...control parameter automatic adjustment system , 301...target position setting unit (target state setting unit), 302...global path planning unit (target state setting unit), 303...local path planning unit, 304...evaluation index design unit, 305...main evaluation index setting unit, 306...parameter update unit, 307...driving result management unit, 401, 402...global path, 501...local path, 801...processing, 801-816...processing, 901...driving area, 902...annular path, 903...buffer path, 1001...annular path.

Claims

1. In a control parameter automatic adjustment system for adjusting a control parameter of a moving object, A sensor for detecting a state quantity of the moving object; a target state setting unit that sets a target state quantity of the moving body; a control command generating unit that generates a control command for the moving object by using the control parameters so as to reduce a deviation of the state quantity from the target state quantity; an evaluation index design unit that designs a plurality of evaluation indexes for calculating an evaluation value for a travel result of the moving object; a parameter update unit that learns a relationship between the plurality of evaluation indexes and the control parameters and updates the control parameters so that an evaluation value calculated using the plurality of evaluation indexes is improved; The parameter update unit learns a relationship between the control parameter and one evaluation index selected from the plurality of evaluation indexes while the mobile body repeatedly travels along a predetermined route, and also learns a relationship between the control parameter and other evaluation indexes among the plurality of evaluation indexes other than the one evaluation index. A control parameter automatic tuning system comprising:

2. 2. The control parameter automatic tuning system according to claim 1, A recording unit is provided for recording the time series changes in the state quantities and the control parameters. A control parameter automatic tuning system comprising:

3. 2. The control parameter automatic tuning system according to claim 1, the evaluation index design unit sets a target value for each of the plurality of evaluation indexes; The parameter update unit terminates updating of the control parameters for the evaluation index that has reached the target value among the plurality of evaluation indexes. A control parameter automatic tuning system comprising:

4. 2. The control parameter automatic tuning system according to claim 1, a main evaluation index setting unit that sets, as a main evaluation index, an evaluation index that is expected to have the largest number of travels of the predetermined route required until updating of the control parameters is completed, from among the plurality of evaluation indexes, based on the function shapes of the plurality of evaluation indexes, the constraint conditions, and the target values; The parameter update unit performs update of the control parameters targeted at the primary evaluation index with priority. A control parameter automatic tuning system comprising:

5. 2. The control parameter automatic tuning system according to claim 1, the control command generating unit determines the control command so as to improve a value of an evaluation function defined by a deviation of the state quantity from the target state quantity, the control command, and weighting parameters for the deviation and the control command; The parameter update unit updates the weight parameters as the control parameters. A control parameter automatic tuning system comprising:

6. An autonomous driving support device comprising the control parameter automatic adjustment system according to claim 1.

Images