Autonomous navigation obstacle avoidance performance evaluation method based on multi-dimensional comprehensive analysis

Abstract

The invention relates to an autonomous navigation obstacle avoidance performance evaluation method based on multi-dimensional comprehensive analysis, and belongs to the technical field of autonomous navigation system evaluation. Comprising the following steps: in a set scene, obtaining an expected target value through a global path planning algorithm; acquiring an actual value of autonomous navigation; processing the original data to generate intermediate data; calculating an index value by using the path planning data and the autonomous navigation data; weighting the basic indexes by using an analytic hierarchy process and carrying out consistency check to obtain reasonable weights of the indexes; and carrying out index fusion by using the weights to obtain a final evaluation result. According to the method, the autonomous navigation obstacle avoidance performance evaluation index system is established, the influence factors of the autonomous navigation obstacle avoidance performance are determined, so that the autonomous navigation obstacle avoidance performance evaluation model based on the analytic hierarchy process is constructed, index fusion is realized, the autonomous navigation obstacle avoidance performance evaluation result is obtained, and the obstacle avoidance performance of the autonomous navigation system is comprehensively, quantitatively and accurately evaluated.

Description

Technical Field

[0001] The invention relates to an obstacle avoidance performance evaluation method, belonging to the technical field of autonomous navigation system evaluation. Background Art

[0002] An ideal unmanned system should have the ability to sense the external environment with the help of sensors without external interference, and thus move autonomously to the target area. Autonomous navigation technology is the core technology to achieve this goal. Intelligent entities with autonomous navigation capabilities can be collectively referred to as autonomous navigation systems, such as autonomous mobile robots, autonomous unmanned submarines, etc.

[0003] Autonomous positioning and navigation technology is an important sign for robots to achieve autonomous walking, and obstacle avoidance is an important function for achieving autonomous positioning and navigation. It means that when the robot is walking, it senses the existence of obstacles in its route planning through sensors, updates the path in real time according to the corresponding algorithm, and finally bypasses the obstacles to reach the target point. The obstacle avoidance performance of the autonomous navigation system is related to the safety, system reliability, navigation efficiency and accuracy of the autonomous navigation system. Therefore, for various technologies and applications involving autonomous navigation, improving obstacle avoidance performance is one of the important goals of continuous optimization. Therefore, it is very important to accurately and reasonably evaluate the performance of the obstacle avoidance performance of the autonomous navigation system. For example: the publication number is CN113780849B, and the name of the invention is INS_GNSS combined navigation system performance evaluation method, system, device and storage medium. In its technical solution, first, the physical object of its evaluation is used to evaluate the INS / GNSS combined navigation system; INS is an inertial navigation system, which is an autonomous navigation system that does not rely on external information and does not radiate energy to the outside; the basic working principle of inertial navigation is based on Newton's laws of mechanics. By measuring the acceleration of the carrier in the inertial reference system, integrating it over time, and transforming it into the navigation coordinate system, it is possible to obtain information such as speed, yaw angle and position in the navigation coordinate system; GNSS is a global navigation satellite system that can be anywhere on the surface of the earth or in near-Earth space. An airborne radio navigation and positioning system that provides users with all-weather 3D coordinates, speed and time information, including one or more satellite constellations and their augmentation systems required to support specific tasks; in short, the INS / GNSS integrated navigation system requires both physical information in the navigation process and satellite support, and is a complex navigation system; secondly, its evaluation scope is to evaluate all aspects of the INS / GNSS integrated navigation system, the evaluation scope is relatively wide, and more information and original data are required; finally, the INS_GNSS integrated navigation system performance evaluation is to conduct a comprehensive evaluation of the INS / GNSS integrated navigation system performance, using a combination of fuzzy comprehensive evaluation and hierarchical analysis to conduct a multi-level analysis of the entire system and establish an evaluation model. The publication number is CN114529088A, and the name of the invention is a driving path planning method and system based on accident risk cost. Its technical solution aims to analyze road historical data and plan a traffic path with the lowest probability of traffic accidents and the shortest time for the driver. In short, it is to arrange and combine all the paths between the starting point and the end point, and then select a traffic path with the lowest probability of traffic accidents, with the focus on "selecting the path".The autonomous navigation system in this paper does not have an original path to choose from. It selects the next path based on the path planning algorithm and exploration algorithm in the autonomous navigation system and the data fed back by sensors such as radar and IMU in the mission environment. Each step is determined independently by the autonomous navigation system, rather than choosing a ready-made path. First, the data source is based on historical traffic accident data to establish an accident risk cost evaluation index system. When selecting roads, historical data is used as support, and the path with a low probability of traffic accidents is selected by analyzing the historical data. Secondly, the evaluation method is to use the entropy weight method for evaluation. The entropy weight method is an objective evaluation method that needs to be driven by data. The program is fixed but not flexible enough during evaluation and it is easy to ignore realistic factors. Thirdly, the navigation object it serves is the driver's path selection, that is, the navigation system it faces has a driver and can be manually controlled at critical moments. Finally, the focus of the implementation is to develop a driving path planning system with a low accident risk cost, rather than evaluating the existing system. The document "Autonomous Navigation for Mobile Service Robots in Urban Pedestrian Environments" discloses that the expected target value is obtained through a global path planning algorithm, and the actual navigation data is compared with the expected path to evaluate the navigation performance. However, first of all, the task it performs is to develop an urban service robot for automatic transportation, guidance or monitoring of personnel or goods, and the task is to develop; secondly, the full text has 10 sections, and the scope of the tasks performed and the main content of the work direction are to propose a fully automatic navigation solution suitable for urban pedestrian environments. Section 9 summarizes the experimental results of the developed platform, which is to propose an autonomous navigation solution instead of evaluating the existing solutions, that is, Section 9 summarizes the results of their experiments, but only summarizes and analyzes the results, and does not establish a fixed evaluation index and evaluation system to evaluate their autonomous navigation system; finally, Section 9 mentions the summary of the experimental results and the target data, but their target data is estimated by their autonomous navigation system itself, that is, it is estimated based on previous statistical data, which is a statistical estimation result of previous data.

[0004] Therefore, the current research on obstacle avoidance performance evaluation of autonomous navigation systems mainly has the following problems:

[0005] (1) How to determine the factors affecting autonomous navigation and obstacle avoidance performance;

[0006] (2) How to quantitatively calculate the evaluation indicators that reflect the obstacle avoidance performance of the autonomous navigation system;

[0007] (3) How to obtain the final obstacle avoidance performance evaluation results through various evaluation indicators.

[0008] The autonomous navigation system will generate time data and path point data during the execution of tasks, and the evaluation indicators are calculated by processing and processing these basic data. In order to reasonably evaluate the obstacle avoidance performance of the autonomous navigation system, choosing a suitable evaluation algorithm to fuse the indicators is a new problem currently faced.

[0009] Therefore, it is urgent to propose an autonomous navigation obstacle avoidance performance evaluation method based on multi-dimensional comprehensive analysis to solve the above technical problems. Summary of the invention

[0010] In order to solve the above technical problems, an autonomous navigation obstacle avoidance performance evaluation method based on multi-dimensional comprehensive analysis is provided. A brief overview of the present invention is given below to provide a basic understanding of certain aspects of the present invention. It should be understood that this overview is not an exhaustive overview of the present invention. It is not intended to determine the key or important parts of the present invention, nor is it intended to limit the scope of the present invention.

[0011] The technical solution of the present invention:

[0012] An autonomous navigation obstacle avoidance performance evaluation method based on multi-dimensional comprehensive analysis includes the following steps:

[0013] Under the set scene terrain, the expected target value is obtained through the global path planning algorithm; the actual value of autonomous navigation is obtained; the target value and the actual value both include the path point set (spatial data), path time (time data) and other data; the original data of the scene terrain is processed to generate intermediate data such as the slope matrix and the roughness matrix;

[0014] The index values ​​are calculated using path planning data and autonomous navigation data. The basic indicators are weighted using the analytic hierarchy process and a consistency test is performed. After the consistency test, the reasonable weights of each indicator can be obtained. The indicators are reasonably integrated using these weights to obtain the final evaluation results.

[0015] Preferably: according to the time and space data generated during the operation of the system, which are the most important basic information, based on the key factors affecting the obstacle avoidance performance, and taking into account that the autonomous navigation obstacle avoidance performance has a direct impact on the time, path selection and energy consumption when the system performs tasks, the present invention extracts evaluation indicators from three dimensions: time (time consumption ratio, time stability), path (steering stability, path stability, safety margin, safety distance) and energy (energy consumption ratio), and comprehensively evaluates the obstacle avoidance performance of the autonomous navigation system; the technical indicators for obstacle avoidance performance evaluation are designed through indicator calculation and hierarchical analysis method, including time consumption ratio, time stability, steering stability, path stability, safety margin, safety distance and energy consumption ratio.

[0016] Preferred: Time consumption ratio: the ratio of the actual time consumed to reach the target location to the ideal time spent to reach the target;

[0017]

[0018] Among them, T L Indicates the actual time taken to reach the target, T A Indicates the time it takes to reach the target under ideal circumstances; the closer the ratio is to 1, the less time it takes to plan the path.

[0019] Preferred: Time stability: Calculate the time intervals between all two adjacent time points on the path, subtract the maximum time interval from the minimum time interval, and divide it by the minimum time interval; the larger the value, the more uneven the time intervals are, and the larger the gap is, the lower the score of this indicator; assuming that a navigation task has a total of n path points, there should be n time points and n-1 time intervals. Traverse the entire time interval array and find Δt max and Δt min , take the difference and divide by Δt min , and get T s , the smaller the value, the better the obstacle avoidance performance;

[0020] Δt1=|t2-t1| (2-2)

[0021]

[0022] Among them, △t1 represents the time interval between two adjacent time points, t1 and t2 represent two adjacent time points, and Δt max Indicates the maximum time interval, Δt min Indicates the minimum time interval.

[0023] Preferred: Steering stability: connect two adjacent path points, extend the line, and calculate the angle between the extended line and the next line, which is the steering angle, i.e. θ1, θ2, θ3... in the schematic diagram. If the autonomous navigation system has good obstacle avoidance performance, the navigation path should be continuous and smooth, i.e. the steering angle should be relatively uniform and not much different. Steering stability is the difference between the maximum steering angle and the minimum steering angle divided by the minimum steering angle. The smaller the value, the better the obstacle avoidance function and the higher the index score.

[0024]

[0025] Among them, θ max represents the maximum steering angle, θ min Indicates the minimum steering angle.

[0026] Preferred: Path stability: For autonomous navigation systems, the factors that affect the stability of the path are mainly the slope, roughness and steering angle of the path; whether it is uphill or downhill, the greater the slope, the more dangerous it is; the greater the path roughness, the more dangerous it is for the lunar rover, which is subject to much less gravity than the earth; the steering angle of the path is measured by the angle between adjacent grids. Generally speaking, the smaller the angle, the more likely the lunar rover is to roll over and the more dangerous it is to drive; therefore, the weights of the path slope, roughness and steering angle in safe driving can be quantified and calculated respectively, and the weighted safety factor can be obtained, as shown in formula (2-5);

[0027]

[0028] Among them, the slope matrix and roughness matrix of the map can be calculated by processing the map data. On the basis of the path points that have been obtained, the slope values ​​θ corresponding to the path points are calculated respectively. i ,θ max Indicates the maximum slope value, θ min Indicates the minimum slope value and roughness value i ,rou max Indicates the maximum roughness, rou min Indicates the minimum roughness and can calculate the angle φ between adjacent grid points i ,φ max represents a grid point, φ min Represents another adjacent grid point. Since the dimensions of these three are different, they need to be normalized. The inverse of the sum of the three is the stability S. It can be seen that the larger S is, the higher the path stability.

[0029] Preferred: Safety margin: With each path point as the center, take 4 grids upward, downward, left and right, that is, 1m*1m square. In this range, there are 25 coordinate points of 5*5. The number of coordinate points occupied by obstacles in the area is detected by the algorithm. The fewer coordinate points occupied by obstacles, the better the obstacle avoidance performance. The safety margin is the number of obstacle coordinate points in the target area. The smaller the index, the better the obstacle avoidance performance of the autonomous navigation system. Assuming that there are x obstacle coordinate points in a certain path point area, the safety margin formula is as follows;

[0030]

[0031] Preferred: Safety distance: With each path point as the center, take 4 grids upward, downward, left and right, that is, 1m*1m square. In this range, there are 25 coordinate points of 5*5. The coordinate points occupied by obstacles in the area are detected by the algorithm, and the distance between each obstacle coordinate point and the path point is calculated respectively; the safety distance is the minimum distance between all obstacle coordinate points and path points; the safety distance of an autonomous navigation system with good obstacle avoidance performance should be within a reasonable range; the formula for calculating the safety distance is as follows;

[0032]

[0033] Among them, (x, y) is the coordinate of the path point, (x i ,y i ) are the coordinates of all obstacle points in the traversal area of ​​the path point.

[0034] Preferred: Energy consumption ratio: refers to the ratio of the energy actually consumed to reach the target location to the energy consumed to reach the target under ideal conditions; the closer the ratio is to 1, the less energy is consumed by path planning;

[0035]

[0036] Among them, E L It represents the energy consumed to reach the target position, E A It represents the energy consumed to reach the target under ideal conditions.

[0037] Optimized: Obstacle avoidance performance evaluation technical indicator fusion: Use the hierarchical analysis method to weight the seven basic indicators and obtain the time consumption ratio R T , time stability T S , Steering stabilityθ S , path stability S, safety margin M S , Safety distance L S , energy consumption ratio R E The weights are represented by α, β, γ, δ, ε, ζ, and η respectively, and the obstacle avoidance performance evaluation model is shown in formula (2-9);

[0038] Score=α×R T +β×T S +γ×θ S +δ×S+ε×M s +ζ×L s +η×R E (2-9).

[0039] Beneficial effects of the present invention:

[0040] The present invention establishes an autonomous navigation obstacle avoidance performance evaluation index system, determines the influencing factors of autonomous navigation obstacle avoidance performance, thereby constructing an autonomous navigation obstacle avoidance performance evaluation model based on the hierarchical analysis method, realizes index fusion, obtains autonomous navigation obstacle avoidance performance evaluation results, and comprehensively, quantitatively and accurately evaluates and assesses the obstacle avoidance performance of the autonomous navigation system;

[0041] The obstacle avoidance performance evaluation method of the autonomous navigation system proposed in the present invention is suitable for the comprehensive evaluation of obstacle avoidance performance in different scenarios and tasks. The obstacle avoidance performance is comprehensively evaluated from the perspectives of time, space and energy of the selected indicators, and weighted by the hierarchical analysis method, so as to obtain quantitative autonomous navigation obstacle avoidance performance evaluation results. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 It is a flow chart of an autonomous navigation obstacle avoidance performance evaluation method based on multi-dimensional comprehensive analysis of the present invention;

[0043] Figure 2 It is a design scheme diagram of technical indicators for obstacle avoidance performance evaluation;

[0044] Figure 3 It is a schematic diagram of steering stability;

[0045] Figure 4 It is a schematic diagram of safety margin and safety distance;

[0046] Figure 5 It is a schematic diagram of the rationality of the evaluation method. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical scheme and advantages of the present invention clearer, the present invention is described below by the specific embodiments shown in the accompanying drawings. However, it should be understood that these descriptions are only exemplary and are not intended to limit the scope of the present invention. In addition, in the following description, the description of well-known structures and technologies is omitted to avoid unnecessary confusion of the concept of the present invention.

[0048] Specific implementation method 1: Combination Figure 1-5The present embodiment is described. The present embodiment is an autonomous navigation obstacle avoidance performance evaluation method based on multi-dimensional comprehensive analysis. The overall process of the autonomous navigation system obstacle avoidance performance evaluation is as follows: Figure 1 As shown in the figure, the expected target value is first obtained through the global path planning algorithm in the specified scenario, including the path point set, path time and other data, and the actual value of autonomous navigation is obtained. Then the original data is processed to generate intermediate data such as slope matrix and roughness matrix, and then the index value is calculated using the path planning data and autonomous navigation data. Then the basic indicators are weighted and consistency tested using the hierarchical analysis method. After the consistency test, the reasonable weight of each indicator can be obtained, and finally the indicators are fused to obtain the final evaluation result; specifically, the following steps are included:

[0049] Under the set scene terrain, the expected target value is obtained through the global path planning algorithm; the actual value of autonomous navigation is obtained; the target value and the actual value both include the path point set (spatial data), path time (time data) and other data; the original data of the scene terrain is processed to generate intermediate data such as the slope matrix and the roughness matrix;

[0050] The index values ​​are calculated using the path planning data and the autonomous navigation data; the basic indexes are weighted using the hierarchical analysis method and a consistency test is performed. After passing the consistency test, the reasonable weights of the various indexes can be obtained; these weights are reasonably used to fuse the indexes and obtain the final evaluation results; the present invention analyzes and studies the working environment and internal mechanism of the autonomous navigation system, establishes an autonomous navigation obstacle avoidance performance evaluation index system, thereby constructing an autonomous navigation obstacle avoidance performance evaluation model based on the hierarchical analysis method, realizes the fusion of indicators, obtains the autonomous navigation obstacle avoidance performance evaluation results, and realizes a comprehensive, quantitative and accurate evaluation of the obstacle avoidance performance of the autonomous navigation system; it can provide a scientific basis and guidance for the further research and optimization of the autonomous navigation technology; at the same time, the method also has a certain reference value and can provide a reference for performance evaluation research in other fields;

[0051] Design of technical indicators for obstacle avoidance performance evaluation: The obstacle avoidance function of the autonomous navigation system is crucial and directly affects the safety, efficiency and navigation accuracy of the system. In order to reasonably evaluate the obstacle avoidance performance of the system and realize its quantitative expression, it is necessary to deeply analyze the working mechanism of the autonomous navigation system, explore the key factors affecting the obstacle avoidance performance, and establish a corresponding evaluation model based on these factors. In this process, the time and space data generated when the system is running are the most important basic information. In view of the fact that the obstacle avoidance performance of autonomous navigation has a direct impact on the time, path selection and energy consumption when the system performs tasks, the present invention extracts evaluation indicators from three dimensions of time, path and energy to comprehensively evaluate the obstacle avoidance performance of the autonomous navigation system.

[0052] Time consumption ratio: the ratio of the actual time consumed to reach the target location to the ideal time spent to reach the target;

[0053]

[0054] Among them, T L Indicates the actual time taken to reach the target, T A Indicates the time it takes to reach the target under ideal conditions. The closer the ratio is to 1, the less time it takes to plan the path.

[0055] Efficient path planning not only considers obstacle avoidance, but also optimizes time factors. A high-performance navigation system can dynamically adjust the path while avoiding obstacles, minimize the distance from the original route, and optimize the overall navigation time. For example, detouring around obstacles may cause deviation from the target path, thereby increasing navigation time. Excellent obstacle avoidance performance can ensure the optimization of the obstacle avoidance path, making the obstacle avoidance process as short as possible and not causing large deviations from the final target position.

[0056] Time stability: Calculate the time intervals between all two adjacent time points on the path, subtract the maximum time interval from the minimum time interval, and divide it by the minimum time interval; the larger the value, the more uneven the time intervals are, and the larger the gap is, the lower the score of this indicator; assuming that a navigation task has a total of n path points, there should be n time points and n-1 time intervals. Traverse the entire time interval array and find Δt max and Δt min , take the difference and divide by Δt min , and get T s , the smaller the value, the better the obstacle avoidance performance;

[0057] Δt1=|t2-t1| (2-2)

[0058]

[0059] Among them, △t1 represents the time interval between two adjacent time points, t1 and t2 represent two adjacent time points, and Δt max Indicates the maximum time interval, Δt min Indicates the minimum time interval;

[0060] When the autonomous navigation system identifies an obstacle, the system must take measures to avoid it; this usually means temporarily interrupting the original path and adjusting the driving route; obstacle avoidance actions will cause changes to the original route, thereby increasing the required navigation time; then the time interval between the time at that point and the next path point recorded by the system will increase; good obstacle avoidance performance should be able to ensure that the location of obstacles is predicted in advance and the path is planned well, so as to ensure a smooth path and a relatively uniform time interval between path points;

[0061] Steering stability: connect two adjacent path points, extend the line, and calculate the angle between the extended line and the next line, which is the steering angle, i.e. θ1, θ2, θ3... in the schematic diagram. If the autonomous navigation system has good obstacle avoidance performance, the navigation path should be continuous and smooth, i.e. the steering angle should be relatively uniform with little difference. Steering stability is the difference between the maximum steering angle and the minimum steering angle divided by the minimum steering angle. The smaller the value, the better the obstacle avoidance function and the higher the index score.

[0062]

[0063] Among them, θ max represents the maximum steering angle, θ min Indicates the minimum steering angle;

[0064] The quality of obstacle avoidance performance has a direct impact on the smoothness and stability of the path. If the system frequently adjusts its direction or makes too sharp turns during obstacle avoidance, the smoothness of the path may be reduced. An efficient obstacle avoidance algorithm should not only avoid obstacles, but also minimize sharp turns in the path to maintain the continuity and smoothness of the path. The system can smoothly adjust the direction according to the current obstacle avoidance situation so that the path is not too tortuous.

[0065] Path stability: For autonomous navigation systems, the factors that affect their stability are mainly the slope, roughness and turning angle of the path. Whether it is uphill or downhill, the greater the slope, the more dangerous it is. The greater the path roughness, the more dangerous it is for the lunar rover, which is subject to much less gravity than the earth. The turning angle of the path is measured by the angle between adjacent grids. Generally speaking, the smaller the angle, the more likely the lunar rover is to roll over and the more dangerous it is to drive. Therefore, the weights of the slope, roughness and turning angle of the path in safe driving can be quantified and calculated respectively, and the weighted safety factor can be obtained, as shown in formula (2-5);

[0066]

[0067] Among them, the slope matrix and roughness matrix of the map can be calculated by processing the map data. On the basis of the path points that have been obtained, the slope values ​​θ corresponding to the path points are calculated respectively. i ,θ max Indicates the maximum slope value, θ min Indicates the minimum slope value and roughness value i ,rou max Indicates the maximum roughness, rou min Indicates the minimum roughness and can calculate the angle φ between adjacent grid points i ,φ max represents a grid point, φ minrepresents another adjacent grid point. Since the dimensions of these three are different, they need to be normalized. The inverse of the sum of the three is the stability S. It can be seen that the larger the S, the higher the path stability.

[0068] Safety margin: Take each path point as the center, and select 4 grids upward, downward, left and right, that is, 1m*1m square. There are 25 coordinate points in this range, 5*5, such as Figure 4 As shown in the yellow dotted box in the figure, the number of coordinate points occupied by obstacles in the area is detected by the algorithm. The fewer coordinate points occupied by obstacles, the better the obstacle avoidance performance. The safety margin is the number of obstacle coordinate points in the target area. The smaller the index is, the better the obstacle avoidance performance of the autonomous navigation system. Assuming that there are x obstacle coordinate points in a certain path point area, the safety margin formula is as follows;

[0069]

[0070] Safety distance: Take each path point as the center, and select 4 grids upward, downward, left and right, that is, 1m*1m square. There are 25 coordinate points in this range, 5*5, such as Figure 4 As shown in the yellow dotted box, the algorithm traverses and detects the coordinate points occupied by obstacles in the area, and calculates the distance between each obstacle coordinate point and the path point respectively; the safe distance is the minimum distance between all obstacle coordinate points and the path point; the safe distance of an autonomous navigation system with good obstacle avoidance performance should be within a reasonable range; the safe distance calculation formula is as follows;

[0071]

[0072] Among them, (x, y) is the coordinate of the path point, (x i ,y i ) are the coordinates of all obstacle points in the traversal area of ​​the path point;

[0073] In some cases, the obstacle avoidance system may adopt a more conservative obstacle avoidance strategy, that is, maintaining a larger safety distance and taking multiple small-scale adjustments to avoid obstacles. Although this strategy can ensure the safety of the system, it usually leads to larger changes in the path and may also increase the navigation time. Conservative obstacle avoidance strategies will increase the complexity of the path, causing the system to detour more times and the curvature of the path. Therefore, the autonomous navigation system should maintain an appropriate interval with obstacles to avoid path length redundancy caused by too large an interval, and also avoid collision risks caused by too small an interval.

[0074] Energy consumption ratio: refers to the ratio of the energy actually consumed to reach the target location to the energy consumed to reach the target under ideal conditions. The closer the ratio is to 1, the less energy is consumed by path planning.

[0075]

[0076] The obstacle avoidance process usually requires the system to re-plan the path to bypass obstacles. If the obstacle avoidance performance is poor, the system may need to choose a longer or more complex detour, resulting in an increase in driving distance and thus more energy consumption. An efficient obstacle avoidance algorithm can minimize the length and complexity of the path while avoiding collisions, thereby reducing the energy required for navigation. Optimized path planning allows the system to avoid obstacles through the shortest and most direct path, avoiding unnecessary detours or repeated adjustments, thereby improving energy utilization.

[0077] Integration of technical indicators for obstacle avoidance performance evaluation: The obstacle avoidance performance evaluation indicator system of the autonomous navigation system is mainly composed of time consumption ratio, time stability, steering stability, path stability, safety margin, safety distance, and energy consumption ratio; among them, time consumption ratio and time stability characterize the quality of obstacle avoidance performance from the time dimension, steering stability, path stability, safety margin, and safety distance measure the quality of navigation system obstacle avoidance performance from the spatial dimension, and energy consumption ratio quantifies obstacle avoidance performance from the perspective of energy consumption;

[0078] The time consumption ratio R is obtained by weighting the seven basic indicators using the hierarchical analysis method. T , time stability T S , Steering stabilityθ S , path stability S, safety margin M S , Safety distance L S , energy consumption ratio R E The weights of are represented by α, β, γ, δ, ε, ζ, and η respectively, and the obstacle avoidance performance evaluation model is shown in formula (2-9);

[0079] Score=α×R T +β×T S +γ×θ S +δ×S+ε×M s +ζ×L s +η×R E (2-9)

[0080] The present invention processes the expected target value obtained by the global path planning algorithm and the actual value data of the autonomous navigation to generate intermediate data and calculate the index value; uses the hierarchical analysis method to weight the basic indicators and perform consistency tests to obtain the reasonable weights of each indicator; uses these weights to fuse the indicators to obtain the final evaluation results; the present invention analyzes the influencing factors of the autonomous navigation obstacle avoidance performance and establishes an autonomous navigation obstacle avoidance performance evaluation index system, thereby constructing an autonomous navigation obstacle avoidance performance evaluation model based on the hierarchical analysis method, realizing index fusion, obtaining the autonomous navigation obstacle avoidance performance evaluation results, and comprehensively, quantitatively and accurately evaluating the obstacle avoidance performance of the autonomous navigation system;

[0081] The present invention can accurately and comprehensively evaluate the obstacle avoidance capability of the autonomous navigation system. Through this evaluation method, a scientific basis and guidance can be provided for the further research and optimization of the autonomous navigation technology. At the same time, this method also has a certain reference value and can provide a reference for the performance evaluation research in other fields.

[0082] The present invention has the following advantages:

[0083] 1. Accuracy of evaluation

[0084] The present invention discloses an autonomous navigation obstacle avoidance performance evaluation method based on multi-dimensional comprehensive analysis, that is, the purpose is "evaluation", the object of evaluation is "obstacle avoidance performance of autonomous navigation system", and an evaluation model of "multi-dimensional comprehensive analysis" is adopted; it is a method that specifically studies the external working environment and internal system of the autonomous navigation system, specifically analyzes and studies the "obstacle avoidance performance" of the autonomous navigation system, determines the influencing factors of the autonomous navigation obstacle avoidance performance, and then establishes an index system and an evaluation model after conducting a multi-dimensional and comprehensive comprehensive analysis around the "obstacle avoidance performance" of the autonomous navigation system; that is, all the work of the present invention is concentrated on the aspect of "obstacle avoidance performance of autonomous navigation system", and a multi-dimensional comprehensive analysis method is adopted when studying the "obstacle avoidance performance of autonomous navigation system", rather than evaluating the performance of various aspects of the entire autonomous navigation system, so our work is more accurate and targeted, and the evaluation indicators and evaluation methods used are more suitable for evaluating the "obstacle avoidance performance of autonomous navigation system";

[0085] 2. Data Source

[0086] In order to more reasonably evaluate the obstacle avoidance performance of the autonomous navigation system, the present invention proposes that for the target scene of the autonomous navigation task, a path planning algorithm is used to calculate an "ideal value" of navigation, that is, under the most ideal situation, the path "target value" of the autonomous navigation system for this navigation task; then real autonomous navigation is performed, and the corresponding navigation path and time data are recorded; this data is the "true value" of the navigation task; through processing, different indicators are used to compare the difference between the "target value" and the "true value", so as to measure the obstacle avoidance performance of the autonomous navigation system; in short, when the autonomous navigation system performs the task, the closer the "true value" of the navigation is to the "target value" calculated by the path planning algorithm, the better the obstacle avoidance performance of the autonomous navigation system;

[0087] That is, the data sources of the present invention are two aspects, one is the "ideal value" calculated by the path planning algorithm, and the other is the "real value" generated during the navigation process of the autonomous navigation system;

[0088] 3. Reasonableness of indicators

[0089] The obstacle avoidance performance of the autonomous navigation system is extremely important and is directly related to the system safety, system efficiency and navigation accuracy. In order to reasonably evaluate the obstacle avoidance performance of the system and quantitatively express the obstacle avoidance performance, it is necessary to conduct an in-depth analysis of the autonomous navigation system, analyze the factors affecting the obstacle avoidance performance, and establish an obstacle avoidance performance evaluation model for the autonomous navigation system. Therefore, there are two issues that need to be considered in the present invention. One is what will affect the obstacle avoidance performance of the autonomous navigation system? The other is in what aspects can the good or bad obstacle avoidance performance of the autonomous navigation system be reflected? The former is the influencing factor of the autonomous navigation system, and the latter is to reflect the obstacle avoidance performance of the autonomous navigation system in the form of an indicator. Through analysis, it is found that the obstacle avoidance performance of the autonomous navigation system will affect the navigation time, navigation path and energy loss. Therefore, the present invention constructs an indicator system from three aspects: time, space and energy consumption. The time and space data generated during the working process of the autonomous navigation system are used as basic information. The influence of the obstacle avoidance performance of the autonomous navigation system on time, space and energy consumption is studied to analyze the obstacle avoidance performance level of the autonomous navigation system.

[0090] 3.1 Time

[0091] The impact of autonomous navigation obstacle avoidance performance on navigation time is a key factor, which is directly related to the efficiency and practicality of the system in practical applications; there is a close correlation between autonomous navigation obstacle avoidance performance and navigation time; effective obstacle avoidance performance can optimize the path as much as possible while ensuring safety and reduce unnecessary time loss; if the obstacle avoidance system is properly designed, it can ensure that obstacles are avoided quickly and accurately in complex environments, thereby reducing the increase in navigation time; and poor obstacle avoidance performance, especially untimely or inaccurate path adjustment, may cause a significant extension of navigation time and even affect the overall task completion efficiency; therefore, when designing an autonomous navigation system, optimizing obstacle avoidance performance, improving reaction speed, and balancing the relationship between path planning and obstacle avoidance decisions are crucial to improving navigation efficiency and reducing time consumption; therefore, the present invention proposes indicators of two time dimensions, namely time consumption ratio and time stability, to measure the obstacle avoidance performance of the autonomous navigation system from the time dimension;

[0092] 3.2 Space

[0093] The impact of autonomous navigation obstacle avoidance performance on the navigation path is a complex and critical factor, involving the length, shape, stability and realization of the ultimate goal of the path; obstacle avoidance performance directly determines how the system responds to obstacles, which path to choose and how to maintain efficiency and accuracy during obstacle avoidance; an effective obstacle avoidance system can not only ensure safety, but also optimize the path as much as possible, reduce redundancy, maintain the smoothness and stability of the path, and shorten navigation time; in a dynamic environment, the system needs to have the ability to respond quickly, plan paths in real time and make efficient obstacle avoidance decisions; poor obstacle avoidance performance may lead to lengthy and complex paths, increase navigation time and affect the realization of the ultimate goal; therefore, improving obstacle avoidance performance and optimizing path planning algorithms can significantly improve the overall efficiency, stability and user experience of the system; therefore, the present invention proposes four indicators, namely steering stability, path smoothness, safety distance and safety margin, to evaluate the obstacle avoidance performance of the autonomous navigation system from a spatial dimension;

[0094] 3.3 Energy consumption

[0095] The performance of autonomous navigation and obstacle avoidance has a significant impact on energy consumption, especially in autonomous systems that need to work for a long time and continuously; factors such as path length, speed changes, and obstacle avoidance strategies during the obstacle avoidance process will directly affect the efficiency of energy use; efficient obstacle avoidance algorithms can optimize paths, reduce unnecessary acceleration and braking, and thus reduce energy consumption while ensuring safety; poor obstacle avoidance performance, especially frequent path adjustments and sudden changes in motion, will lead to a significant increase in energy consumption; therefore, optimizing obstacle avoidance performance and path planning strategies, and balancing energy efficiency and safety, are key tasks in the design of autonomous navigation systems. Therefore, we proposed an energy consumption ratio indicator to evaluate the obstacle avoidance performance of autonomous navigation systems from an energy perspective;

[0096] 3.4 Rationality of evaluation method

[0097] The object to be evaluated in the present invention is the obstacle avoidance performance of the autonomous navigation system. After establishing a reasonable evaluation index system, a reasonable evaluation method should be selected to weight the indexes to obtain the final evaluation result.

[0098] The analytic hierarchy process is a relatively common and accurate evaluation method. It can simplify complex problems, decompose decision-related elements into multiple levels such as goals, criteria, and plans, and conduct qualitative and quantitative analysis on this basis. It is a systematic, simple, flexible and effective decision-making method. The principle and operation are relatively simple and easy to understand. It also introduces a consistency verification mechanism to verify the consistency of the decision maker's judgment, thereby avoiding decision-making errors caused by inconsistent judgments, thereby improving the reliability of decision results. Therefore, the analytic hierarchy process is very suitable for evaluating the obstacle avoidance performance of autonomous navigation systems, that is, our target layer is the obstacle avoidance performance of autonomous navigation systems, and the criterion layer is time consumption ratio, time stability, steering stability, path stability, safety margin, safety distance, and energy consumption ratio.

[0099] Then, the weights corresponding to the indicators are calculated by following the steps of constructing a judgment matrix, finding the maximum eigenvalue, eigenvector, normalization, and consistency test. Finally, the weights are multiplied by the corresponding indicator values ​​and then summed up to obtain the final obstacle avoidance performance score of the autonomous navigation system.

[0100] The autonomous navigation system faced by the present invention only needs the time and position information during the navigation process, and does not need other complex support such as satellites. The present invention is specifically used to evaluate the obstacle avoidance performance of the autonomous navigation system, with a more accurate scope, a deeper excavation of the factors affecting the obstacle avoidance performance of the autonomous navigation system, and a more detailed and intuitive evaluation index. After a multi-dimensional comprehensive analysis of the autonomous navigation system, a concise and clear hierarchical system is established, and an evaluation model is established. The constructed hierarchical structure is more concise, and the evaluation method is more concise and effective.

[0101] The autonomous navigation system faced by the present invention has no original path to choose from. It selects the next path in the mission environment based on the path planning algorithm and exploration algorithm in the autonomous navigation system and the data fed back by sensors such as radar and IMU. Each step is determined independently by the autonomous navigation system, rather than selecting a ready-made path. When establishing the indicator system and constructing the evaluation model, no historical data support is required. Only the "ideal value" and "real value" of each navigation are required. The focus is on the processing and analysis of the current navigation data, which has nothing to do with historical data. The evaluation method adopted is the hierarchical analysis method of multi-dimensional comprehensive analysis. The hierarchical analysis method is a subjective evaluation method, but we set up a consistency test to avoid the influence of subjective factors, and the evaluation model is more flexible and reliable. The autonomous navigation system faced is completely autonomous, without a driver, and is completely not controlled by human will. The present invention evaluates the obstacle avoidance performance of the existing autonomous navigation system, that is, the task of this article is to evaluate.

[0102] The present invention is to construct an obstacle avoidance performance evaluation model for an autonomous navigation system, and to evaluate the navigation obstacle avoidance performance of existing autonomous navigation systems, i.e., unmanned vehicles, navigation robots, etc., that is, the task of the invention is to evaluate; the present invention focuses on the obstacle avoidance performance of autonomous navigation systems, proposes a set of autonomous navigation evaluation index systems, and constructs an obstacle avoidance performance evaluation model for autonomous navigation systems. That is, the task of the present invention is to analyze and study existing autonomous navigation systems, and then to reasonably evaluate them, rather than to implement a set of autonomous navigation systems; the target value proposed by the present invention is to calculate the scene of the autonomous navigation task using a path planning algorithm to obtain an ideal value under ideal conditions, that is, the target value of the present invention is obtained through real calculation of the scene, rather than estimation, and the real value is the real path and time data generated when the autonomous navigation system actually performs the task, and the calculated "ideal value" and the "real value" of the autonomous navigation system are used to calculate the index.

[0103] It should be noted that in the above embodiments, as long as the technical solutions are not contradictory, they can be arranged and combined, and those skilled in the art can exhaust all possibilities based on the mathematical knowledge of arrangement and combination. Therefore, the present invention will no longer describe the technical solutions after arrangement and combination one by one, but it should be understood that the technical solutions after arrangement and combination have been disclosed by the present invention.

[0104] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An autonomous navigation obstacle avoidance performance evaluation method based on multi-dimensional comprehensive analysis, characterized by: The following steps are involved: In the set scenario, the expected target value is obtained through the global path planning algorithm; the actual value of autonomous navigation is obtained; Process the original data to generate intermediate data; Calculate the index value using path planning data and autonomous navigation data; use the analytic hierarchy process to weight the basic indicators and conduct consistency tests to obtain the reasonable weight of each indicator; These weights are used to fuse indicators and obtain the final evaluation results.

2. The autonomous navigation obstacle avoidance performance evaluation method based on multi-dimensional comprehensive analysis according to claim 1, characterized in that: The generated time and space data are used as basic information. Based on the factors affecting the obstacle avoidance performance, the influence of time, path selection and energy consumption is considered. The technical indicators for obstacle avoidance performance evaluation are designed through indicator calculation and hierarchical analysis method, including time consumption ratio, time stability, steering stability, path smoothness, safety margin, safety distance and energy consumption ratio.

3. The autonomous navigation obstacle avoidance performance evaluation method based on multi-dimensional comprehensive analysis according to claim 2 is characterized in that: Time consumption ratio: the ratio of the actual time consumed to reach the target location to the ideal time spent to reach the target; Among them, T L Indicates the actual time taken to reach the target, T A Indicates the time it takes to reach the target under ideal circumstances.

4. The autonomous navigation obstacle avoidance performance evaluation method based on multi-dimensional comprehensive analysis according to claim 3 is characterized in that: Time stability: Calculate the time intervals between all two adjacent time points on the path, subtract the maximum time interval from the minimum time interval, and divide it by the minimum time interval; the larger the value, the more uneven the time intervals, and the larger the gap, the lower the score of this indicator; traverse the entire time interval array and find Δt max and Δt min , take the difference and divide by Δt min , and get T s , the smaller the value, the better the obstacle avoidance performance; Δt1=|t2-t1| (2-2) Among them, △t1 represents the time interval between two adjacent time points, t1 and t2 represent two adjacent time points, and Δt max Indicates the maximum time interval, Δt min Indicates the minimum time interval.

5. The autonomous navigation obstacle avoidance performance evaluation method based on multi-dimensional comprehensive analysis according to claim 1, characterized in that: Steering stability: the difference between the maximum steering angle and the minimum steering angle divided by the minimum steering angle; the smaller the value, the better the obstacle avoidance function and the higher the index score; Among them, θ max represents the maximum steering angle, θ min Indicates the minimum steering angle.

6. The autonomous navigation obstacle avoidance performance evaluation method based on multi-dimensional comprehensive analysis according to claim 5 is characterized in that: Path smoothness: Quantify the weights of the path slope, roughness and steering angle in safe driving, and obtain the weighted safety factor, as shown in formula (2-5); Among them, the slope matrix and roughness matrix of the map can be calculated by processing the map data. On the basis of the path points that have been obtained, the slope values ​​θ corresponding to the path points are calculated respectively. i , roughness value i , and the angle φ between adjacent grid points can be calculated i , because the dimensions of these three are different, they need to be normalized; the inverse of the result of adding the three is the stability S. It can be seen that the larger the S, the higher the path stability.

7. The autonomous navigation obstacle avoidance performance evaluation method based on multi-dimensional comprehensive analysis according to claim 6 is characterized in that: Safety margin: With each path point as the center, 4 grids are taken upward, downward, left and right. There are 25 coordinate points in this range of 5*5. The number of coordinate points occupied by obstacles in the area is detected by the algorithm. The fewer coordinate points occupied by obstacles, the better the obstacle avoidance performance. The safety margin is the number of obstacle coordinate points in the target area. The smaller the index is, the better the obstacle avoidance performance of the autonomous navigation system. Assuming that there are x obstacle coordinate points in a certain path point area, the safety margin formula is as follows; 8. The autonomous navigation obstacle avoidance performance evaluation method based on multi-dimensional comprehensive analysis according to claim 7, characterized in that: Safety distance: With each path point as the center, take 4 grids upward, downward, left and right. There are 25 coordinate points in this range of 5*5. The coordinate points occupied by obstacles in the area are detected by the algorithm, and the distance between each obstacle coordinate point and the path point is calculated respectively; the safety distance is the minimum distance between all obstacle coordinate points and path points; the safety distance of an autonomous navigation system with good obstacle avoidance performance should be within a reasonable range; the formula for calculating the safety distance is as follows; Among them, (x, y) is the coordinate of the path point, (x i ,y i ) are the coordinates of all obstacle points in the traversal area of ​​the path point.

9. The autonomous navigation obstacle avoidance performance evaluation method based on multi-dimensional comprehensive analysis according to claim 8, characterized in that: Energy consumption ratio: refers to the ratio of the energy actually consumed to reach the target position to the energy consumed to reach the target under ideal conditions; The closer the ratio is to 1, the less energy is consumed in path planning; 10. The autonomous navigation obstacle avoidance performance evaluation method based on multi-dimensional comprehensive analysis according to claim 9, characterized in that: Fusion of technical indicators for obstacle avoidance performance evaluation: Using the analytic hierarchy process to weight the seven basic indicators, we can get the time consumption ratio R T , time stability T S , Steering stabilityθ S , path stability S, safety margin M S , Safety distance L S , energy consumption ratio R E The weights are represented by α, β, γ, δ, ε, ζ, and η respectively, and the obstacle avoidance performance evaluation model is shown in formula (2-9); Score=α×R T +β×T S +γ×θ S +δ×S+ε×M s +ζ×L s +η×R E (2-9)。

Images