Neuromorphic brain-like decision-making system

Abstract

The invention discloses a neuromorphic brain-like decision-making system, and relates to the technical field of artificial intelligence. Object information is abstracted into symbols and marked, the initial position of a robot is used as a starting point in combination with position coordinates, the coordinates of a task passing point and an end point are determined, and an executable path is generated through a graph search algorithm; calculating a path walking distance, and combining with a speed range set by a user to obtain an estimated walking duration; meanwhile, the relation between the distance and the energy consumption is preset, estimated energy consumption is determined through matching, on this basis, a final energy consumption performance value is calculated, the final energy consumption performance value and estimated walking duration of the path are extracted, and comprehensive evaluation indexes of all the paths are calculated in combination with the energy consumption deviation habit and the efficiency deviation habit of the user and a weight coefficient set of the energy consumption deviation habit and the efficiency deviation habit; the path with the highest index is selected as the final walking path, intelligent decision making according to user preferences and multi-dimensional evaluation is achieved, energy consumption and efficiency are balanced, and the operation efficiency of the robot and the user experience are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of artificial intelligence, and specifically to a neuromorphic brain-like decision-making system. Background Art

[0002] In the current era of rapid technological development, robot technology is increasingly widely used in many fields.

[0003] However, when the neuromorphic brain-like decision-making system in the prior art is applied to robots, there are still the following deficiencies: In a complex and changeable real environment, a robot needs to efficiently process a large amount of complex information to make accurate decisions. However, the information processing ability of traditional systems is limited, and it is difficult to quickly and accurately identify various objects in the environment, analyze their mutual relationships and the impact on its own tasks; for example, when performing tasks in an indoor scene, in the face of many obstacles and dynamic elements such as furniture and people, traditional decision-making systems often cannot quickly plan a reasonable path in combination with the user's decision-making habits, resulting in low task execution efficiency or even failure.

[0004] Therefore, a neuromorphic brain-like decision-making system is introduced. Summary of the Invention

[0005] The purpose of the present invention is to solve the problems pointed out in the background art, and to propose a neuromorphic brain-like decision-making system.

[0006] The purpose of the present invention can be achieved by the following technical solutions: A neuromorphic brain-like decision-making system, comprising: An environmental perception module: receiving commands issued by the user, determining the tasks required to be executed by the user using semantic analysis, collecting environmental data in all directions based on the tasks required to be executed by the user command, and obtaining the feature information of the object; wherein the feature information includes the shape, color, texture of the object and their position coordinates in space, extracting features from the collected environmental data, and identifying different types of objects; A derivation and decision module: predefined symbol domain and logical operator group, wherein the symbol domain is used to represent the object and position information elements in the environment, and the logical operator group comprehensively analyzes the mutual relationships of each object in the symbol domain using a preset algorithm, and plans the final walking path for the robot to execute the task; A reaction execution module: generating corresponding control instructions according to the final walking path planned by the derivation and decision module, and driving the robot to execute the planned final walking path.

[0007] As a preferred embodiment of the present invention, the logical operator group comprehensively analyzes the mutual relationships of each object in the symbol domain using a preset algorithm, specifically as: Abstract the information elements of each object in the environment into symbols, that is, mark the symbols according to the object type; obtain the coordinates of the position information of each object in the environment; Taking the initial position of the current robot as the starting point, obtain the waypoints and the end point required for the task, and determine the coordinates corresponding to the starting point, waypoints, and end point; Starting from the starting point of the robot, use a graph search-based algorithm to search for paths and generate executable paths corresponding to each current task.

[0008] As a preferred embodiment of the present invention, plan the final walking path for the robot to perform the task, specifically: Calculate the required walking distance of each executable path corresponding to the current task, and at the same time obtain the range of the feasible walking speed set by the user; extract the intermediate value within the range of the feasible walking speed and calculate it with the required walking distance of each executable path corresponding to the current task, that is, divide each group of required walking distances by the intermediate value to obtain the estimated required walking duration of each executable path corresponding to the current task; Preset groups of distance value ranges corresponding to the required walking distance, and set each group of distance value ranges to correspond to an estimated energy consumption; match the required walking distance of each executable path corresponding to the current task with the preset groups of distance value ranges to determine the estimated energy consumption of each executable path corresponding to the current task; Analyze the estimated energy consumption of each executable path corresponding to the current task to determine the final energy consumption performance value of each executable path corresponding to the current task; Extract the final energy consumption performance value and the estimated required walking duration of each executable path corresponding to the current task, and conduct a comprehensive evaluation in combination with the user's preference habits to determine the final walking path.

[0009] As a preferred embodiment of the present invention, determine the final energy consumption performance value of each executable path corresponding to the current task, specifically: Count the required number of turns and potential influence adjustment times when the robot simulates the execution of each executable path corresponding to the current task; Extract the symbols of the movable objects from the symbol domain, and obtain the specific coordinates corresponding to the symbols of the movable objects, denoted as influence coordinates. Obtain the shortest distance between the robot and each group of influence coordinates during the execution of the corresponding path at the current time point, denoted as the influence distance; Preset the reference distances corresponding to different movable objects; compare the influence distances in each group in the executable paths corresponding to each current task with the corresponding preset reference distances, and count the number of influence distances less than the corresponding preset reference distances in the executable paths corresponding to each current task as the potential influence adjustment times for each current task. Set the weight coefficients corresponding to the required number of turns and the potential influence adjustment times respectively. Multiply the required number of turns and the potential influence adjustment times of the executable paths corresponding to each current task by the corresponding set weight coefficients respectively, and then sum to obtain the additional energy consumption values of the executable paths corresponding to each current task. Preset the value range groups corresponding to the additional energy consumption values, and set an energy consumption additional coefficient corresponding to each value range group. Match the additional energy consumption values of the executable paths corresponding to each current task with the value range groups to obtain the energy consumption additional coefficients of the executable paths corresponding to each current task. Multiply the energy consumption additional coefficients of the executable paths corresponding to each current task by the estimated energy consumption to obtain the final energy consumption performance values of the executable paths corresponding to each current task.

[0010] As a preferred embodiment of the present invention, extract the final energy consumption performance values and the estimated required walking times of the executable paths corresponding to each current task, and conduct a comprehensive evaluation in combination with the user's preference habits. Specifically: Extract the final energy consumption performance values and the estimated required walking times of the executable paths corresponding to each current task, and preset the value range groups corresponding to the final energy consumption performance values and the value range groups corresponding to the estimated required walking times respectively. Set an energy consumption score and an efficiency score corresponding to each value range group of the performance value and each value range group of the duration respectively. Match the final energy consumption performance values and the estimated required walking times of the executable paths corresponding to each current task with the value range groups of the performance values and the value range groups of the durations respectively to determine the energy consumption scores and the efficiency scores of the executable paths corresponding to each current task.

[0011] As a preferred embodiment of the present invention, extract the final energy consumption performance values and the estimated required walking times of the executable paths corresponding to each current task, and conduct a comprehensive evaluation in combination with the user's preference habits, which also includes: Preset the weight coefficient sets corresponding to the energy consumption preference habit and the efficiency preference habit respectively. The weight coefficient sets corresponding to the energy consumption preference habit and the efficiency preference habit both include the weight coefficients corresponding to the energy consumption score and the efficiency score respectively. Extract the weight coefficients corresponding to the energy consumption score and the efficiency score within the energy consumption preference habit respectively. Multiply the energy consumption score and the efficiency score of each executable path corresponding to the current task by the corresponding weight coefficients within the energy consumption preference habit respectively, and then sum them to obtain the comprehensive evaluation index of each executable path corresponding to the current task within the energy consumption preference habit; Extract the weight coefficients corresponding to the energy consumption score and the efficiency score within the efficiency preference habit respectively. Multiply the energy consumption score and the efficiency score of each executable path corresponding to the current task by the corresponding weight coefficients within the efficiency preference habit respectively, and then sum them to obtain the comprehensive evaluation index of each executable path corresponding to the current task within the efficiency preference habit; Taking the current time point as the starting point, extract the preference habits selected by the user within the set time window before the starting point, and count the respective corresponding preference times of the energy consumption preference habit and the efficiency preference habit within the set time window; If the number of times of the energy consumption preference habit > the number of times of the efficiency preference habit, then select the executable path with the highest comprehensive evaluation index from the comprehensive evaluation indexes of each executable path corresponding to the current task within the energy consumption preference habit as the final walking path; If the number of times of the energy consumption preference habit < the number of times of the efficiency preference habit, then select the executable path with the highest comprehensive evaluation index from the comprehensive evaluation indexes of each executable path corresponding to the current task within the efficiency preference habit as the final walking path.

[0012] Compared with the prior art, the beneficial effects of the present invention are: The present invention receives a user command, determines a task by using semantic analysis, and omnidirectionally collects environmental data through sensors such as lidar, cameras, and depth cameras, extracts feature information such as the shape, color, texture, and position coordinates of an object, and identifies different types of objects to provide basic data for subsequent decision-making; The present invention abstracts object information into symbols and marks them, combines position coordinates, takes the initial position of the robot as the starting point, determines the passing points and end point coordinates of the task, and generates an executable path by using a graph search algorithm; then calculates the walking distance of the path, combines the set speed range of the user to obtain the estimated walking duration; at the same time, presets the relationship between distance and energy consumption, matches and determines the estimated energy consumption, calculates the final energy consumption performance value on this basis, extracts the final energy consumption performance value and the estimated walking duration of the path, and then combines the user's energy consumption preference habit and efficiency preference habit and their weight coefficient sets, calculates the comprehensive evaluation index of each path, and selects the path with the highest index as the final walking path, realizing intelligent decision-making based on user preferences and multi-dimensional evaluation, balancing energy consumption and efficiency, and improving the running efficiency of the robot and the user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] For the convenience of those skilled in the art to understand, the present invention will be further described below with reference to the accompanying drawings.

[0014] Figure 1 This is the principle block diagram of the present invention. Detailed implementation manners

[0015] Next, the technical solutions of the present invention will be clearly and completely described in conjunction with the embodiments. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts shall fall within the protection scope of the present invention.

[0016] Please refer to Figure 1 As shown, a neuromorphic brain-like decision-making system includes an environmental perception module, a derivation decision-making module, and a reaction execution module; The environmental perception module is used to receive the commands issued by the user, determine the tasks required to be executed by the user through semantic analysis, collect environmental data in all directions based on the tasks required to be executed by the user commands, and obtain the feature information of the objects; the feature information includes the shape, color, texture of the objects and their position coordinates in space, and uses image recognition technology to extract features from the collected environmental data and identify different types of objects; different types such as people and items, etc. The environmental perception module is equipped with a variety of high-precision sensors such as lidar, cameras, and depth cameras; these sensors work together to provide the robot with all-round environmental perception capabilities; Perform semantic analysis and grammar disassembling on the commands issued by the user, and convert the commands into task descriptions that can be understood by the machine. For example, for the command "Take the book on the living room table to the bookshelf in the study", it can identify key elements such as "book", "living room table", "bookshelf in the study", etc., so as to determine the tasks required to be executed by the user; The derivation decision-making module is used to pre-define a symbol domain and a logical operator group. The symbol domain is used to represent the object and position information elements in the environment, and the logical operator group comprehensively analyzes the mutual relationships of the objects in the symbol domain using a preset algorithm and plans the final walking path for the robot to execute the task; Specifically: Abstract the object information elements in the environment into symbols, that is, mark the symbols according to the object types; for example, use specific symbols to represent people (P), furniture (V), obstacles (O), etc.; obtain the coordinate of the position information of each object in the environment through a spatial perception algorithm combined with sensor data; the coordinate is (x, y, z); Taking the initial position of the current robot as the starting point, the passing points and the end point required for the task, determine the coordinates corresponding to the starting point, passing points, and end point; Start from the starting point of the robot and use a graph search-based algorithm to search for paths, and generate executable paths corresponding to each current task; It should be noted that, for example, for the task of "taking the book on the living room table to the bookshelf in the study", the following paths are generated: Path 1: Bypass other furniture in the living room and go directly from the table to the bookshelf in the study. The path is relatively short, but it requires frequent changes in direction to avoid furniture.

[0017] Path 2: Move along the wall of the living room, avoiding most furniture and areas with human activities, but the path is relatively long; Path 3: First approach the pedestrians in the living room. Without disturbing the pedestrians, use the relatively empty area next to the pedestrians to go to the bookshelf in the study. This path needs to consider the movement and behavior of pedestrians, but it has an advantage in overall efficiency; Calculate the required walking distance of each executable path corresponding to the current task, and at the same time obtain the range of feasible walking speeds set by the user for the robot; it is set by the user, that is, the maximum walking speed range allowed for the robot in the current application area; extract the median value within the range of feasible walking speeds, and calculate it with the required walking distance of each executable path corresponding to the current task, that is, divide each group of required walking distances by the median value, so as to obtain the estimated required walking time for each executable path corresponding to the current task; Preset groups of distance value ranges corresponding to the required walking distance, and set each group of distance value ranges to correspond to an estimated energy consumption; match the required walking distance of each executable path corresponding to the current task with the preset groups of distance value ranges, so as to determine the estimated energy consumption of each executable path corresponding to the current task; Count the required number of turns and potential impact adjustment times when the robot simulates the execution of each executable path corresponding to the current task; Extract the symbols of movable objects from the symbol domain, and obtain the specific coordinates corresponding to the symbols of movable objects, denoted as influence coordinates. Obtain the shortest distance between the robot and each group of influence coordinates during the execution of the corresponding path at the current time point, denoted as the influence distance; Preset the reference distances corresponding to the influence distances of different movable objects; for example, if the movable object is a pedestrian and the activity of the pedestrian is relatively strong, the set value of the reference distance corresponding to the influence distance is relatively high. If the influence distance is less than the reference distance, it means that the possibility of the robot needing to adjust the turning during the execution of the corresponding path is relatively high; compare each group of influence distances in each executable path corresponding to the current task with the corresponding preset reference distances, and count the number of influence distances less than the corresponding preset reference distances in each executable path corresponding to the current task as the potential impact adjustment times of each executable path corresponding to the current task; Set the weight coefficients corresponding to the required number of turns and the potential number of impact adjustments respectively. Multiply the required number of turns and the potential number of impact adjustments of each executable path corresponding to the current task by the corresponding set weight coefficients, and then sum them to obtain the additional energy consumption values of each executable path corresponding to the current task; Preset the value ranges of each group of additional values corresponding to the additional energy consumption values, and set an energy consumption additional coefficient corresponding to each group of value ranges; the range of the energy consumption additional coefficient is set between 1.086 - 1.138, and the higher the additional energy consumption value, the higher the corresponding energy consumption additional coefficient obtained; Match the additional energy consumption values of each executable path corresponding to the current task with each group of value ranges of the additional values, so as to obtain the energy consumption additional coefficients of each executable path corresponding to the current task; Multiply the energy consumption additional coefficients of each executable path corresponding to the current task by the estimated energy consumption, so as to obtain the final energy consumption performance values of each executable path corresponding to the current task; It should be noted that the energy consumption corresponding to the preset required walking distance is preset, and then the estimated energy consumption is matched according to the actual path distance, so that the robot can more accurately estimate the energy required to complete the task, which helps to reasonably plan the energy use and avoid task interruption due to insufficient energy; By counting the required number of turns and the potential number of impact adjustments, the influence of environmental factors (such as movable objects) on the robot during the execution of the path is considered, and the complexity of path execution is comprehensively evaluated, providing more realistic data for energy consumption calculation.

[0018] Extract the final energy consumption performance values and the estimated required walking duration of each executable path corresponding to the current task, and preset the value ranges of each group of performance values and the value ranges of each group of durations corresponding to the final energy consumption performance values and the estimated required walking duration respectively; Set a group of energy consumption scores and efficiency scores corresponding to each group of performance value ranges and each group of duration ranges respectively; the higher the final energy consumption performance value, the lower the corresponding energy consumption score obtained, and the longer the estimated required walking duration, the lower the corresponding efficiency score obtained; the ranges of both the energy consumption score and the efficiency score are set between 1 - 10; Match the final energy consumption performance values and the estimated required walking duration of each executable path corresponding to the current task with each group of performance value ranges and each group of duration ranges respectively, so as to determine the energy consumption scores and efficiency scores of each executable path corresponding to the current task; The set of weight coefficients corresponding to the preset energy consumption preference habit and the efficiency preference habit respectively. Both the set of weight coefficients corresponding to the energy consumption preference habit and the efficiency preference habit include the weight coefficients corresponding to the energy consumption score and the efficiency score respectively. Among them, the weight coefficient of the energy consumption score in the energy consumption preference habit is greater than the weight coefficient of the efficiency score, and the weight coefficient of the energy consumption score in the efficiency preference habit is less than the weight coefficient of the efficiency score. Extract the weight coefficients corresponding to the energy consumption score and the efficiency score in the energy consumption preference habit respectively. Multiply the energy consumption score and the efficiency score of each executable path corresponding to the current task by the corresponding weight coefficients in the energy consumption preference habit respectively, and then sum them up to obtain the comprehensive evaluation index of each executable path corresponding to the current task in the energy consumption preference habit. Extract the weight coefficients corresponding to the energy consumption score and the efficiency score in the efficiency preference habit respectively. Multiply the energy consumption score and the efficiency score of each executable path corresponding to the current task by the corresponding weight coefficients in the efficiency preference habit respectively, and then sum them up to obtain the comprehensive evaluation index of each executable path corresponding to the current task in the efficiency preference habit. Taking the current time point as the starting point, extract the preference habit selected by the user within the set time window before the starting point, and count the respective corresponding preference times of the energy consumption preference habit and the efficiency preference habit within the set time window. If the number of times of the energy consumption preference habit > the number of times of the efficiency preference habit, then select the executable path with the highest comprehensive evaluation index from the comprehensive evaluation indexes of each executable path corresponding to the current task in the energy consumption preference habit as the final walking path. If the number of times of the energy consumption preference habit < the number of times of the efficiency preference habit, then select the executable path with the highest comprehensive evaluation index from the comprehensive evaluation indexes of each executable path corresponding to the current task in the efficiency preference habit as the final walking path. It should be noted that by presetting the value range of the final energy consumption performance value and the estimated required walking duration, and respectively matching the energy consumption score and the efficiency score, a quantitative evaluation of the path is carried out from two key dimensions of energy consumption and time consumption, providing comprehensive and intuitive data support for the robot path selection. Setting the energy consumption preference habit and the efficiency preference habit and their corresponding sets of weight coefficients can flexibly adjust the evaluation focus according to different usage scenarios and demand preferences of users. For example, in a scenario with limited energy, users may be more inclined to the energy consumption preference, and at this time the weight of the energy consumption score is greater; while in a scenario with high time requirements, the efficiency preference habit is more applicable, highlighting the weight of the efficiency score. Starting from the current time point, the bias habits of the user within a set time window before the starting point are counted, and the final walking path is selected according to the number of times of energy consumption bias habits and efficiency bias habits. This method can automatically adapt to the user's long-term preference habits without the user having to manually select each time, improving the autonomy and intelligence level of the robot's decision-making; Taking into comprehensive consideration energy consumption and efficiency, and combining with user preferences, the path with the highest comprehensive evaluation index is finally selected as the final walking path, enabling the robot to no longer be limited to a single factor when choosing a path, but to find a balance among multiple important factors, ensuring that the selected path can meet the user's preferences while achieving the optimal combination of energy consumption and efficiency as much as possible, improving the overall operation efficiency of the robot and the user experience.

[0019] The reaction execution module is used to generate corresponding control instructions according to the final walking path planned by the derivation decision module, and drive the robot to execute the planned final walking path; It also includes a somatosensory module, a visual processing module, a semantic logic module, a network and positioning module, and a working memory module; The somatosensory module is used to help the robot perceive and adapt to the surrounding environmental changes. Through the somatosensory sensors, the robot can perceive physical characteristics such as temperature, humidity, and pressure; The visual processing module is used to utilize image recognition technology to enable the robot to real-time identify and analyze objects, spatial layouts, and their positions in the environment; The semantic logic module is used to enable the robot to understand and reason about the user's language instructions and perceive human emotions through semantic understanding and natural language (NLP) processing technologies, as well as DQN (Deep Q-Network); The network and positioning module is used to combine network nerves and topological maps to locate the position of the robot in space; The working memory module is used to store and adjust the data generated during the execution process and build a database; It should be noted that through the mutual cooperation of the above-mentioned various modules, the following effects are achieved: Intelligent decision-making: It can intelligently plan the task path according to the user's commands and environmental information, optimize the decision-making strategy, and balance energy consumption and efficiency; Flexible adaptation: It has a strong perception and adaptation ability, and can adjust the task execution strategy in real time according to environmental changes, improving the stability of task execution; Personalized service: According to the user's preference habits, it dynamically adjusts the path planning to provide a personalized service experience.

[0020] Multi-dimensional evaluation: Combining multiple factors such as walking duration, energy consumption, and path optimization, comprehensively evaluate the path selection to ensure the efficient completion of the task; The preferred embodiments of the present invention disclosed above are only used to help illustrate the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to the specific embodiments only. Obviously, many modifications and variations can be made according to the content of this specification. These embodiments are selected and specifically described in this specification in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can well understand and utilize the present invention. The present invention is only limited by the claims and their full scope and equivalents.

Claims

1. A neuromorphic brain-like decision-making system, characterized in that: include: Environmental perception module: Receives commands from users, uses semantic analysis to determine the tasks that users need to perform, collects environmental data in all directions based on the tasks that need to be performed by user commands, and obtains feature information of objects; the feature information includes the shape, color, texture and position coordinates of objects in space, extracts features from the collected environmental data, and identifies different types of objects; Inference and decision module: pre-defines the symbol domain and logic operator group, where the symbol domain is used to represent the objects and position information elements in the environment, and the logic operator group uses the preset algorithm to comprehensively analyze the relationship between the objects in the symbol domain and plan the final walking path for the robot to perform the task; Reaction execution module: Generates corresponding control instructions based on the final walking path planned by the derivation decision module, and drives the robot to execute the planned final walking path.

2. A neuromorphic brain-like decision-making system according to claim 1, characterized in that: The logical operator group uses a preset algorithm to conduct a comprehensive analysis of the relationships between objects in the symbol domain, specifically: Abstract the information elements of each object in the environment into symbols, that is, mark the symbols according to the object type; obtain the coordinates of the position information of each object in the environment; Taking the current initial position of the robot as the starting point, obtain the waypoints and end point required for the task, and determine the coordinates corresponding to the starting point, waypoints, and end point; Starting from the starting point of the robot, a graph-based search algorithm is used to search for paths and generate executable paths corresponding to each current task.

3. A neuromorphic brain-like decision-making system according to claim 2, characterized in that: Plan the robot's final walking path to perform the task, specifically: Calculate the required walking distance of the executable path corresponding to each current task, and obtain the robot's walking speed range set by the user; extract the middle value within the walking speed range, and calculate it with the required walking distance of the executable path corresponding to each current task, that is, divide each group of required walking distance by the middle value, so as to obtain the estimated required walking time of the executable path corresponding to each current task; Preset each group of distance value ranges corresponding to the required walking distance, and set each group of distance value ranges to correspond to an estimated energy consumption; Matching the required walking distance of each executable path corresponding to each current task with the preset distance value ranges of each group, thereby determining the estimated energy consumption of each executable path corresponding to each current task; Analyze the estimated energy consumption of the executable paths corresponding to each current task, so as to determine the final energy consumption performance value of the executable paths corresponding to each current task; The final energy consumption performance value and estimated walking time of each executable path corresponding to the current task are extracted, and a comprehensive evaluation is performed based on the user's preference habits to determine the final walking path.

4. A neuromorphic brain-like decision-making system according to claim 3, characterized in that: Determine the final energy consumption performance value of each executable path corresponding to the current task, specifically: Count the number of turns and potential impact adjustments required by the robot when simulating the execution of each executable path corresponding to the current task; Extract the symbol of the movable object from the symbol domain, obtain the specific coordinates corresponding to the symbol of the movable object, record them as the influence coordinates, obtain the shortest distance between the robot and each group of influence coordinates during the execution of the corresponding path at the current time point, record them as the influence distance; Preset reference distances for the impact distances corresponding to different movable objects; Compare each group of impact distances in the executable path corresponding to each current task with the corresponding preset reference distance, and count the number of impact distances in the executable path corresponding to each current task that are smaller than the corresponding preset reference distance as the number of potential impact adjustments for the executable path corresponding to each current task; The weight coefficients corresponding to the required number of turns and the number of potential impact adjustments are set respectively, and the required number of turns and the number of potential impact adjustments of the executable paths corresponding to each current task are multiplied by the corresponding set weight coefficients respectively, and then the sum is obtained to obtain the energy consumption added value of the executable paths corresponding to each current task; Preset the value ranges of each group of additional values ​​corresponding to the energy consumption additional value, and set each group of additional value ranges to correspond to an energy consumption additional coefficient; Match the energy consumption additional value of the executable path corresponding to each current task with the value range of each group of additional values, so as to obtain the energy consumption additional coefficient of the executable path corresponding to each current task; The energy consumption additional coefficient of the executable path corresponding to each current task is multiplied by the estimated energy consumption to obtain the final energy consumption performance value of the executable path corresponding to each current task.

5. A neuromorphic brain-like decision-making system according to claim 4, characterized in that: The final energy consumption performance value and estimated walking time of each executable path corresponding to the current task are extracted, and a comprehensive evaluation is performed based on the user's preference habits, specifically: Extract the final energy consumption performance value and the estimated required walking time of each executable path corresponding to the current task, and preset the performance value ranges of each group and the duration value ranges of each group corresponding to the final energy consumption performance value and the estimated required walking time; Set each group of performance value ranges and each group of duration value ranges to correspond to a group of energy consumption scores and efficiency scores respectively; The final energy consumption performance value and the estimated walking time of the executable paths corresponding to each current task are matched with the value ranges of each group of performance values ​​and the value ranges of each group of duration values, respectively, so as to determine the energy consumption score and efficiency score of the executable paths corresponding to each current task.

6. A neuromorphic brain-like decision-making system according to claim 5, characterized in that: Extract the final energy consumption performance value and estimated walking time of each executable path corresponding to the current task, and make a comprehensive evaluation based on the user's preference habits, including: Preset weight coefficient sets corresponding to energy consumption biased habits and efficiency biased habits, respectively, wherein the weight coefficient sets corresponding to energy consumption biased habits and efficiency biased habits include weight coefficients corresponding to energy consumption scores and efficiency scores, respectively; Extract the weight coefficients corresponding to the energy consumption score and efficiency score in the energy consumption bias habit, multiply the energy consumption score and efficiency score of the executable path corresponding to each current task by the corresponding weight coefficient in the energy consumption bias habit, and then sum them up to obtain the comprehensive evaluation index of the executable path corresponding to each current task in the energy consumption bias habit; Extract the weight coefficients corresponding to the energy consumption score and efficiency score in the efficiency bias habit, multiply the energy consumption score and efficiency score of the executable path corresponding to each current task by the corresponding weight coefficient in the efficiency bias habit, and then sum them up to obtain the comprehensive evaluation index of the executable path corresponding to each current task in the efficiency bias habit; Taking the current time point as the starting point, extract the preference habits selected by the user within the set time window before the starting point, and count the number of preference times corresponding to the energy consumption preference habit and the efficiency preference habit within the set time window; If the number of energy consumption biased habits is greater than the number of efficiency biased habits, then the executable path with the highest comprehensive evaluation index among the executable paths corresponding to the current tasks in the energy consumption biased habits is selected as the final walking path; If the number of energy consumption biased habits is less than the number of efficiency biased habits, then the executable path with the highest comprehensive evaluation index is selected as the final walking path from the comprehensive evaluation indexes of the executable paths corresponding to the current tasks in the efficiency biased habits.

Images