WO2020233052A1 - 智能化路径规划方法、装置、设备及存储介质 - Google Patents

智能化路径规划方法、装置、设备及存储介质 Download PDF

Info

Publication number
WO2020233052A1
WO2020233052A1 PCT/CN2019/120802 CN2019120802W WO2020233052A1 WO 2020233052 A1 WO2020233052 A1 WO 2020233052A1 CN 2019120802 W CN2019120802 W CN 2019120802W WO 2020233052 A1 WO2020233052 A1 WO 2020233052A1
Authority
WO
WIPO (PCT)
Prior art keywords
path planning
grid
pulse
loudness
random number
Prior art date
Application number
PCT/CN2019/120802
Other languages
English (en)
French (fr)
Inventor
杜艳艳
Original Assignee
深圳壹账通智能科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 深圳壹账通智能科技有限公司 filed Critical 深圳壹账通智能科技有限公司
Publication of WO2020233052A1 publication Critical patent/WO2020233052A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/26—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network
    • G01C21/34—Route searching; Route guidance
    • G01C21/3446—Details of route searching algorithms, e.g. Dijkstra, A*, arc-flags, using precalculated routes

Definitions

  • This application relates to the field of artificial intelligence technology, in particular to intelligent path planning methods, devices, equipment, and storage media.
  • the intelligent path planning algorithm is a new meta-heuristic algorithm. Because of its unique advantages, it has attracted the attention of researchers in recent years and has gradually become a hot issue in the field of intelligent computing. The related scientific research results are also increasing. The FJSP scheduling problem , Function optimization, wireless sensors, cloud manufacturing supply chain and many other research results are emerging. However, the current intelligent path planning algorithm has low optimization accuracy, and it is easy to fall into the defects of partial area optimization and premature. Moreover, the current intelligent path planning algorithm has low optimization accuracy and slow convergence speed in the later stage, which is easy to fall into Local optimization. Although various scholars at home and abroad have made various improvements and achieved results, the inventor realizes that its accuracy is not the highest, and it may not be the optimal solution for path planning.
  • This application provides an intelligent path planning method, device, equipment, and storage medium, the main purpose of which is to provide an accurate intelligent path planning solution.
  • the first aspect of the present application provides an intelligent path planning method, including: Step A: The image processing layer receives a map image with marked obstacles and an initialization speed v t , where t represents time and is based on The grid method performs grid division on the map image with marked obstacles to obtain a grid map, and sets the initial position of the path in the grid map And end position And randomly generates a pulse frequency f t, the value of the pulse emission rate R t A t and the loudness of the grid map, the initialization velocity v t, the pulse frequency f t, the emission of the pulse and loudness A t R t
  • the value of is input to the path planning layer
  • Step B The path planning layer updates the time t to t+1, and solves the optimal direction solution x * in the initial grid based on the gradient descent algorithm, combined with the Optimal direction solution x * predict the initial position
  • the predicted position at time t+1 is The pulse frequency f t is f t+1 , and the initialization speed
  • the second aspect of the present application provides an intelligent path planning device, including: a map picture receiving module for receiving a map picture of marked obstacles and an initialization speed v t , where t represents time and is based on a grid method.
  • the map picture with the obstacles marked is divided into grids to obtain a grid map, and the initial position of the path is set in the grid map And end position And randomly generates a pulse frequency f t, a pulse emission rate R t values of A t and loudness.
  • the direction and path solving module is used to update the time t to t+1, and solve the optimal direction solution x * in the initial grid based on the gradient descent algorithm, and combine the optimal direction solution x * to predict the Initial position
  • the predicted position at time t+1 is
  • the pulse frequency f t is f t+1
  • the initialization speed v t is v t+1 .
  • Pulse is determined and output path means for displacement using a feedback register generator generates a uniformly distributed random number rand, determines the uniform random number rand R t, and pulses emissivity magnitude relation between the loudness A t of initialization, when the When the uniformly distributed random number rand is less than the initial pulse emission rate R t , the loudness judgment module 40 is executed, and when the uniformly distributed random number rand is greater than the initial pulse emission rate R t , the predicted position is determined With the end position Whether in the same grid, when the predicted position With the end position Not in the same grid, update the current position to the predicted position The current pulse frequency is f t+1 , the current initialization speed is v t+1 , and the direction and path solving module 20 and the pulse judgment module 30 are repeated.
  • the path planning layer When the predicted position With the end position In the same grid, the path planning layer combines the predicted positions at each time to complete the optimal path and output it.
  • Loudness determination module configured to, when the time of the uniformly distributed random number is less than the loudness A t rand, updates the initialization pulse emissivity R t is R t + 1, a loudness value A t A t + 1, and repeat path direction solution module 20, when the uniformly distributed random number rand is between the reset pulse and loudness emissivity R t a t, is repeated determination module 30 performs pulse.
  • a third aspect of the present application provides an intelligent path planning device, including: a memory and at least one processor, the memory stores instructions, and the memory and the at least one processor are interconnected by wires; the at least one The processor invokes the instructions in the memory, so that the intelligent path planning device executes the method described in the first aspect.
  • the fourth aspect of the present application provides a computer-readable storage medium, the computer-readable storage medium stores computer instructions, and when the computer instructions run on a computer, the computer executes the above-mentioned first aspect method.
  • the intelligent path planning method, device, equipment, and storage medium proposed in this application mainly divide the map by the grid method, and start the search based on the intelligent path planning method at the initial position, and find the best one to the next time point. And determine whether the optimal position meets the requirements according to a uniformly distributed random number, and finally output the optimal path plan.
  • This application resets the calculation method of loudness and pulse frequency, which can avoid falling into a local optimum during path planning, and introduces a gradient descent algorithm to improve the speed of path planning. Therefore, this application can realize the precise path planning function.
  • FIG. 1 is a schematic flowchart of an intelligent path planning method provided by an embodiment of this application
  • FIG. 2 is a schematic diagram of the internal structure of an intelligent path planning device provided by an embodiment of the application.
  • FIG. 3 is a schematic diagram of modules of an intelligent path planning program in an intelligent path planning device provided by an embodiment of the application.
  • the embodiments of the present application provide an intelligent path planning method, device, equipment and storage medium, which are mainly used to divide the map by the grid method, and start searching at the initial position based on the intelligent path planning method, and find The optimal position at the next time point, and judge whether the optimal position meets the requirements according to a uniformly distributed random number, until the optimal path plan is finally output.
  • This application resets the calculation method of loudness and pulse frequency, which can avoid falling into a local optimum during path planning, and introduces a gradient descent algorithm to improve the speed of path planning. Therefore, this application can realize the precise path planning function.
  • FIG. 1 is a schematic flowchart of the intelligent path planning method provided by an embodiment of the present application.
  • the method can be executed by a device, and the device can be implemented by software and/or hardware.
  • the intelligent path planning method includes:
  • the image processing layer receives the map image of the marked obstacle and the initialization speed v t , where t represents time, and performs grid division on the map image of the marked obstacle based on the grid method to obtain a grid map, And set the initial position of the path in the raster map And end position And randomly generates a pulse frequency f t, the value of the pulse emission rate R t A t and the loudness of the grid map, the initialization velocity v t, the pulse frequency f t, the emission of the pulse and loudness A t R t
  • the value of is input to the route planning layer.
  • the method for obtaining the map picture with marked obstacles includes: first mapping the pixels of the original map picture to [0-155], and marking the obstacle with the pixel set in [155-200].
  • the image processing layer performs unit division on the map picture with marked obstacles to obtain multiple units, and the multiple units are squares of fixed size and the same.
  • Each unit performs image preprocessing operations.
  • the image preprocessing operations include expansion, erosion, and binarization.
  • the expansion processing includes the use of 3*3 structural elements and each pixel in the map picture where the obstacles have been marked to perform an "OR" operation
  • the corrosion processing includes the use of 3*3 structural elements and all pixels.
  • Each pixel in the map picture where the obstacle has been marked performs an AND operation
  • the binarization operation includes changing the pixel value of each pixel in the map picture where the obstacle has been marked to 0 or 255.
  • a Moore field tracking algorithm is used to extract the contours of obstacles on multiple units completed by the image preprocessing operation to obtain the grid map.
  • the Moore domain tracking algorithm is also called an indirect neighborhood algorithm.
  • the basic idea is to find a black pixel and define it as the starting pixel. There are many ways to locate the starting pixel, one of which is: starting from the bottom left pixel of the multiple units, scanning each column of pixels from bottom to top to the top pixel, and then according to the top Starting from the pixel, scan each column of pixels from left to right until a black pixel is encountered, and it is used as the starting pixel.
  • the path planning layer updates the time t to t+1, and solves the optimal direction solution x * in the initial grid based on a gradient descent algorithm, and predicts the optimal direction solution x * in combination with the optimal direction solution x * initial position
  • the predicted position at time t+1 is
  • the pulse frequency f t is f t+1
  • the initialization speed v t is v t+1 .
  • the gradient descent algorithm includes a loss function and an iterative function, wherein the loss function is:
  • the iteration function is:
  • ⁇ is the iterative parameter of the gradient descent algorithm
  • T represents the transposition of the matrix
  • ⁇ is the noise parameter of the iterative function.
  • the initial position The predicted position at time t+1 is
  • the pulse frequency prediction is:
  • f t+1 f min +(f max -f min ) ⁇
  • the initialization speed prediction is:
  • ⁇ is a random number generated in [0,1]
  • f min and f max represent the minimum and maximum pulse frequency respectively.
  • the displacement path planning storage layer using a feedback generator generates a uniformly distributed random number size relationship rand, determines the uniform random number rand and the emissivity of the reset pulse R t and A t of loudness.
  • the feedback shift register generator is composed of a shift register and a combinational logic feedback device, and can randomly generate the uniformly distributed random number rand.
  • step S6 when the uniformly distributed random number rand is less than the loudness A t, said updating is pulsed emission rate R t R t + 1, updating the value of loudness A t A t + 1, and returns to step S4 to continue Analyzing The relationship between the uniformly distributed random number rand and the initial pulse emission rate R t .
  • return to S3 regenerate a uniformly distributed random number rand.
  • the preferred embodiment of the present application uses the following function to update the initial pulse emission rate R t to R t+1 :
  • R t+1 R t (1-e - ⁇ t )
  • the enhancement coefficient of ⁇ pulse emission rate ⁇ is the attenuation coefficient of volume, and e is an infinite loop irrational number.
  • the path planning layer combines the predicted positions at each time to complete the optimal path and output it.
  • the position at time t is t+1 time position t+n time position Combining the positions at all times, the optimal plan of the path can be determined and output.
  • This application also provides an intelligent path planning device.
  • FIG. 2 it is a schematic diagram of the internal structure of an intelligent path planning device provided by an embodiment of this application.
  • the intelligent path planning device 1 may be a PC (Personal Computer, personal computer), or a terminal device such as a smart phone, a tablet computer, or a portable computer, or a server.
  • the intelligent path planning device 1 at least includes a memory 11, a processor 12, a communication bus 13, and a network interface 14.
  • the memory 11 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), magnetic memory, magnetic disk, optical disk, etc.
  • the memory 11 may be an internal storage unit of the intelligent path planning device 1 in some embodiments, for example, the hard disk of the intelligent path planning device 1.
  • the memory 11 may also be an external storage device of the intelligent path planning device 1, for example, a plug-in hard disk equipped on the intelligent path planning device 1, a smart media card (SMC), and a secure digital (Secure Digital, SD) card, flash card (Flash Card), etc.
  • the memory 11 may also include both an internal storage unit of the intelligent path planning device 1 and an external storage device.
  • the memory 11 can be used not only to store application software and various data installed in the intelligent path planning device 1, such as the code of the intelligent path planning program 01, etc., but also to temporarily store data that has been output or will be output.
  • the processor 12 may be a central processing unit (CPU), controller, microcontroller, microprocessor, or other data processing chip, and is used to run the program code or processing stored in the memory 11 Data, such as execution of intelligent path planning program 01, etc.
  • CPU central processing unit
  • controller microcontroller
  • microprocessor or other data processing chip
  • the communication bus 13 is used to realize the connection and communication between these components.
  • the network interface 14 may optionally include a standard wired interface and a wireless interface (such as a WI-FI interface), and is usually used to establish a communication connection between the apparatus 1 and other electronic devices.
  • the device 1 may also include a user interface.
  • the user interface may include a display (Display) and an input unit such as a keyboard (Keyboard).
  • the optional user interface may also include a standard wired interface and a wireless interface.
  • the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode, organic light emitting diode) touch device, etc.
  • the display can also be called a display screen or a display unit as appropriate, and is used to display the information processed in the intelligent route planning device 1 and to display a visualized user interface.
  • FIG. 2 only shows the intelligent path planning device 1 with components 11-14 and the intelligent path planning program 01. Those skilled in the art can understand that the structure shown in FIG. 1 does not constitute an intelligent path planning device
  • the definition of 1 may include fewer or more components than shown, or a combination of certain components, or different component arrangements.
  • the memory 11 stores an intelligent path planning program 01; when the processor 12 executes the intelligent path planning program 01 stored in the memory 11, the following steps are implemented:
  • Step 1 The image processing layer receives the map picture of the marked obstacle and the initial speed v t , where t represents time, and performs grid division on the map picture of the marked obstacle based on the grid method to obtain a grid map , And set the initial position of the path in the grid map And end position And randomly generates a pulse frequency f t, the value of the pulse emission rate R t A t and the loudness of the grid map, the initialization velocity v t, the pulse frequency f t, the emission of the pulse and loudness A t R t The value of is input to the route planning layer.
  • the method for obtaining the map picture with marked obstacles includes: first mapping the pixels of the original map picture to [0-155], and marking the obstacle with the pixel set in [155-200].
  • the image processing layer performs unit division on the map picture with marked obstacles to obtain multiple units, and the multiple units are squares of fixed size and the same.
  • Each unit performs image preprocessing operations.
  • the image preprocessing operations include expansion, erosion, and binarization.
  • the expansion processing includes the use of 3*3 structural elements and each pixel in the map picture where the obstacles have been marked to perform an "OR" operation
  • the corrosion processing includes the use of 3*3 structural elements and all pixels.
  • Each pixel in the map picture where the obstacle has been marked performs an AND operation
  • the binarization operation includes changing the pixel value of each pixel in the map picture where the obstacle has been marked to 0 or 255.
  • a Moore field tracking algorithm is used to extract the contours of obstacles on multiple units completed by the image preprocessing operation to obtain the grid map.
  • the Moore domain tracking algorithm is also called an indirect neighborhood algorithm.
  • the basic idea is to find a black pixel and define it as the starting pixel. There are many ways to locate the starting pixel, one of which is: starting from the bottom left pixel of the multiple units, scanning each column of pixels from bottom to top to the top pixel, and then according to the top Starting from the pixel, scan each column of pixels from left to right until a black pixel is encountered, and it is used as the starting pixel.
  • Step 2 The path planning layer updates the time t to t+1, and solves the optimal direction solution x * in the initial grid based on the gradient descent algorithm, and combines the optimal direction solution x * to predict the Initial position
  • the predicted position at time t+1 is
  • the pulse frequency f t is f t+1
  • the initialization speed v t is v t+1 .
  • the gradient descent algorithm includes a loss function and an iterative function, wherein the loss function is:
  • the iteration function is:
  • ⁇ is the iterative parameter of the gradient descent algorithm
  • T represents the transposition of the matrix
  • ⁇ is the noise parameter of the iterative function.
  • the initial position The predicted position at time t+1 is
  • the pulse frequency prediction is:
  • f t+1 f min +(f max -f min ) ⁇
  • the initialization speed prediction is:
  • ⁇ is a random number generated in [0,1]
  • f min and f max represent the minimum and maximum pulse frequency respectively.
  • Step three the displacement path planning storage layer using a feedback generator generates a uniformly distributed random number size relationship rand, determines the uniform random number rand and the emissivity of the reset pulse R t and A t of loudness.
  • the feedback shift register generator is composed of a shift register and a combinational logic feedback device, and can randomly generate the uniformly distributed random number rand.
  • Step 4 Determine the magnitude relationship between the uniformly distributed random number rand and the initial pulse emission rate R t .
  • Step five when the uniformly distributed random number rand is smaller than the initializing pulse when the emission rate R t, is further determined the uniformly distributed random number rand magnitude relation between the loudness of A t.
  • Step six when the uniformly distributed random number rand is less than the loudness A t, said updating is pulsed emission rate R t R t + 1, updating the value of loudness A t A t + 1, and returns to step four Continue to determine the magnitude relationship between the uniformly distributed random number rand and the initial pulse emission rate R t . When the uniform random number rand is larger than the loudness A t, is returned to step three, regenerate a uniformly distributed random number rand.
  • the preferred embodiment of the present application uses the following function to update the initial pulse emission rate R t to R t+1 :
  • R t+1 R t (1-e - ⁇ t )
  • the enhancement coefficient of ⁇ pulse emission rate ⁇ is the attenuation coefficient of volume, and e is an infinite loop irrational number.
  • Step 7 When the uniformly distributed random number rand is greater than the initial pulse emission rate R t , determine the predicted position With the end position Whether it is in the same grid.
  • Step 8 When the predicted position With the end position Not in the same grid, update the current position to the predicted position
  • the current pulse frequency is f t+1
  • the current initialization speed is v t+1 , and return to step two.
  • Step 9 When the predicted position With the end position In the same grid, the path planning layer combines the predicted positions at each time to complete the optimal path and output it.
  • the position at time t is t+1 time position t+n time position Combining the positions at all times, the optimal plan of the path can be determined and output.
  • the intelligent path planning program can also be divided into one or more modules, and the one or more modules are stored in the memory 11 and are executed by one or more processors (this embodiment It is executed by the processor 12) to complete this application.
  • the module referred to in this application refers to a series of computer program instruction segments that can complete specific functions, used to describe the execution process of the intelligent path planning program in the intelligent path planning device .
  • FIG. 3 a schematic diagram of program modules of an intelligent path planning method program in an embodiment of an intelligent path planning method and apparatus of this application.
  • the intelligent path planning method program can be divided into The map image receiving module 10, the direction and path solving module 20, the pulse judgment and output path module 30, and the loudness judgment module 40 are exemplary:
  • the map picture receiving module 10 is configured to: receive a map picture with marked obstacles and an initialization speed v t , where t represents time, and perform grid division on the map picture with marked obstacles based on a grid method Obtain a grid map, and set the initial position of the path in the grid map And end position And randomly generates a pulse frequency f t, a pulse emission rate R t values of A t and loudness.
  • the direction and path solving module 20 is configured to: update the time t to t+1, and solve the optimal direction solution x * in the initial grid based on a gradient descent algorithm, and combine the optimal direction solution x * Predict the initial position
  • the predicted position at time t+1 is
  • the pulse frequency f t is f t+1
  • the initialization speed v t is v t+1 .
  • Determining an output path of the pulse module 30 and configured to: register generator using a feedback displacement uniform random number rand, determines the uniform random number rand magnitude relation between the reset pulse R t emissivity and loudness of A t, When the uniformly distributed random number rand is less than the initialization pulse emission rate R t , the loudness judgment module 40 is executed, and when the uniformly distributed random number rand is greater than the initialization pulse emission rate R t , the predicted position is determined With the end position Whether in the same grid, when the predicted position With the end position Not in the same grid, update the current position to the predicted position The current pulse frequency is f t+1 , the current initialization speed is v t+1 , and the direction and path solving module 20 and the pulse judgment module 30 are repeated. When the predicted position With the end position In the same grid, the path planning layer combines the predicted positions at each time to complete the optimal path and output it.
  • the loudness determination module 40 is configured to: when the rand uniformly distributed random number is less than the loudness A t, said updating initialization pulse emission rate is R t R t + 1, a loudness value A t + A t 1, the direction of the path and repeat the solution module 20, when the uniformly distributed random number rand is between the reset pulse and loudness emissivity R t a t, is repeated determination module 30 performs pulse.
  • the embodiment of the present application also proposes a computer-readable storage medium, the computer-readable storage medium stores an intelligent path planning program, and the intelligent path planning program can be executed by one or more processors to Implement the following operations:
  • Step A The image processing layer receives the map image of the marked obstacle and the initial speed v t , where t represents time, and performs grid division on the map image of the marked obstacle based on the grid method to obtain a grid map , And set the initial position of the path in the grid map And end position And randomly generates a pulse frequency f t, the value of the pulse emission rate R t A t and the loudness of the grid map, the initialization velocity v t, the pulse frequency f t, the emission of the pulse and loudness A t R t Enter the value of to the path planning layer;
  • Step B The path planning layer updates the time t to t+1, and solves the optimal direction solution x * in the initial grid based on the gradient descent algorithm, and combines the optimal direction solution x * to predict the Initial position
  • the predicted position at time t+1 is The pulse frequency f t is f t+1
  • the initialization speed v t is v t+1 ;
  • Step C the displacement path planning storage layer using a feedback generator generates a uniformly distributed random number size relationship rand, determines the uniform random number rand and the emissivity of the reset pulse R t and A t of loudness;
  • Step D When the uniform random number rand reset pulse is smaller than the emission rate R T, a step E: When the uniform random number rand is less than the loudness A t, said updating initialization pulse emission rate R t is an R t + 1, a loudness value a t a t + 1, and repeat the procedure D, when the value of the rand is a uniformly distributed random number between the reset pulse and loudness emissivity R & lt t a t , Then repeat steps C and D;
  • the path planning layer combines the predicted positions at each time to complete the optimal path and output it.
  • the present application also provides an intelligent path planning device, including: a memory and at least one processor, the memory stores instructions, the memory and the at least one processor are interconnected through a wire; the at least one processor is called The instructions in the memory enable the intelligent path planning device to execute the steps in the above intelligent path planning method.
  • the present application also provides a computer-readable storage medium.
  • the computer-readable storage medium may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium.
  • the computer-readable storage medium stores computer instructions, and when the computer instructions are executed on the computer, the computer executes the following steps:
  • Step A The image processing layer receives the map image of the marked obstacle and the initial speed v t , where t represents time, and performs grid division on the map image of the marked obstacle based on the grid method to obtain a grid map , And set the initial position of the path in the grid map And end position And randomly generates a pulse frequency f t, the value of the pulse emission rate R t A t and the loudness of the grid map, the initialization velocity v t, the pulse frequency f t, the emission of the pulse and loudness A t R t Enter the value of to the path planning layer;
  • Step B The path planning layer updates the time t to t+1, and solves the optimal direction solution x * in the initial grid based on the gradient descent algorithm, and combines the optimal direction solution x * to predict the Initial position
  • the predicted position at time t+1 is The pulse frequency f t is f t+1
  • the initialization speed v t is v t+1 ;
  • Step C the displacement path planning storage layer using a feedback generator generates a uniformly distributed random number size relationship rand, determines the uniform random number rand and the emissivity of the reset pulse R t and A t of loudness;
  • Step D When the uniform random number rand reset pulse is smaller than the emission rate R T, a step E: When the uniform random number rand is less than the loudness A t, said updating initialization pulse emission rate R t is an R t + 1, a loudness value a t a t + 1, and repeat the procedure D, when the value of the rand is a uniformly distributed random number between the reset pulse and loudness emissivity R & lt t a t , Then repeat steps C and D;
  • the path planning layer combines the predicted positions at each time to complete the optimal path and output it.
  • the disclosed system, device, and method may be implemented in other ways.
  • the device embodiments described above are only illustrative.
  • the division of the units is only a logical function division, and there may be other divisions in actual implementation, for example, multiple units or components can be combined or It can be integrated into another system, or some features can be ignored or not implemented.
  • the displayed or discussed mutual coupling or direct coupling or communication connection may be indirect coupling or communication connection through some interfaces, devices or units, and may be in electrical, mechanical or other forms.

Landscapes

  • Engineering & Computer Science (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Automation & Control Theory (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Image Analysis (AREA)

Abstract

提供了一种人工智能技术，具体为一种智能化路径规划方法、装置、设备及存储介质，其中该方法包括：接收已标记出障碍物的地图图片和初始化速度，对已标记出障碍物的地图图片进行栅格划分并初始化位置xi t与终点位置xt j，随机产生脉冲频率ft、脉冲发射率Rt和响度At后（S1），求解初始栅格内的最优方向解x*和下一时间位置xt+1 i后（S2），产生均匀分布随机数rand，并判断均匀分布随机数rand与初始化脉冲发射率Rt和响度At的大小关系（S3），直至预测位置 xt+1 i与终点位置xt j在同一栅格内（S7），最终结合各时间的预测位置完成最优路径并输出（S9）。该方法可以实现精准的智能路径规划功能。

Description

智能化路径规划方法、装置、设备及存储介质
本申请要求于2019年5月21日提交中国专利局、申请号为201910421933.0、发明名称为“智能化路径规划方法、装置及计算机可读存储介质”的中国专利申请的优先权，其全部内容通过引用结合在申请中。
技术领域
本申请涉及人工智能技术领域，尤其涉及智能化路径规划方法、装置、设备及存储介质。
背景技术
智能化路径规划算法属于新兴的元启发式算法，因其特有的优点，近几年备受研究者的关注，逐渐成为智能计算领域的热点问题，相关的科研成果也日益增加，在FJSP调度问题、函数优化、无线传感器、云制造供应链等诸多方面的研究成果日益浮现。然而，目前的智能化路径规划算法存在的寻优准确度不高，容易陷入部分区域最优，早熟等缺陷，且当前的智能化路径规划算法寻优精度不高，后期收敛速度慢，易陷入局部最优，虽然国内外各种学者进行了不同程度的改进，取得了成果，但是发明人意识到其精度还是不是最高的，在路径规划问题上，可能不是最优解。
发明内容
本申请提供一种智能化路径规划方法、装置、设备及存储介质，其主要目的在于提供一种精准的智能路径规划方案。
为实现上述目的，本申请第一方面提供了一种智能化路径规划方法，包括：步骤A：图片处理层接收已标记出障碍物的地图图片和初始化速度v t，其中t表示时间，并基于栅格法对所述已标记出障碍物的地图图片进行栅格划分得到栅格地图，并在所述栅格地图内设置路径的初始位置
Figure PCTCN2019120802-appb-000001
与终点位置
Figure PCTCN2019120802-appb-000002
并随机产生脉冲频率f t、脉冲发射率R t和响度A t的值，将所述栅格地图、所述初始化速度v t、所述脉冲频率f t、脉冲发射率R t和响度A t的值输入至路径规划层；步骤B：所述路径规划层将所述时间t更新至t+1，并基于梯度下降算法求解所述初始栅格内的最优方向解x *，结合所述最优方向解x *预测所述初始位置
Figure PCTCN2019120802-appb-000003
在t+1时间的预测位置为
Figure PCTCN2019120802-appb-000004
所述脉冲频率f t为f t+1，所述初始化速度v t为v t+1；步骤C：所述路径规划层利用反馈位移寄存发生器产生均匀分布随机数rand，判断所述均匀分布随机数rand与所述初始化脉冲发射率R t和响度A t的大小关系；步骤D：当所述均匀分布随机数rand小于所述初始化脉冲发射率R t时，执行步骤E：当所述均匀分布随机数rand小于所述响度A t时，更新所述初始化脉冲发射率R t为R t+1，所述响度A t值为A t+1，并重复步骤D，当所述均匀分布随机数rand的值介于所述初始化脉冲发射率R t和响度A t之间，则重复步骤C与步骤D；当所述均匀分布随机数rand大于所述初始化脉冲发射率R t时， 判断所述预测位置
Figure PCTCN2019120802-appb-000005
与所述终点位置
Figure PCTCN2019120802-appb-000006
是否在同一栅格内，当所述预测位置
Figure PCTCN2019120802-appb-000007
与所述终点位置
Figure PCTCN2019120802-appb-000008
不在同一栅格内，更新当前位置为所述预测位置
Figure PCTCN2019120802-appb-000009
当前脉冲频率为f t+1及当前初始化速度为v t+1，并重复步骤B和步骤C，当所述预测位置
Figure PCTCN2019120802-appb-000010
与所述终点位置
Figure PCTCN2019120802-appb-000011
在同一栅格内，所述路径规划层结合各时间的预测位置完成最优路径并输出。
本申请第二方面提供了一种智能化路径规划装置，包括：地图图片接收模块，用于接收已标记出障碍物的地图图片和初始化速度v t，其中t表示时间，并基于栅格法对所述已标记出障碍物的地图图片进行栅格划分得到栅格地图，并在所述栅格地图内设置路径的初始位置
Figure PCTCN2019120802-appb-000012
与终点位置
Figure PCTCN2019120802-appb-000013
并随机产生脉冲频率f t、脉冲发射率R t和响度A t的值。方向与路径求解模块，用于将所述时间t更新至t+1，并基于梯度下降算法求解所述初始栅格内的最优方向解x *，结合所述最优方向解x *预测所述初始位置
Figure PCTCN2019120802-appb-000014
在t+1时间的预测位置为
Figure PCTCN2019120802-appb-000015
所述脉冲频率f t为f t+1，所述初始化速度v t为v t+1。脉冲判断并输出路径模块，用于利用反馈位移寄存发生器产生均匀分布随机数rand，判断所述均匀分布随机数rand与所述初始化脉冲发射率R t和响度A t的大小关系，当所述均匀分布随机数rand小于所述初始化脉冲发射率R t时，执行响度判断模块40，当所述均匀分布随机数rand大于所述初始化脉冲发射率R t时，判断所述预测位置
Figure PCTCN2019120802-appb-000016
与所述终点位置
Figure PCTCN2019120802-appb-000017
是否在同一栅格内，当所述预测位置
Figure PCTCN2019120802-appb-000018
与所述终点位置
Figure PCTCN2019120802-appb-000019
不在同一栅格内，更新当前位置为所述预测位置
Figure PCTCN2019120802-appb-000020
当前脉冲频率为f t+1，当前初始化速度为v t+1，并重复方向与路径求解模块20和脉冲判断模块30，当所述预测位置
Figure PCTCN2019120802-appb-000021
与所述终点位置
Figure PCTCN2019120802-appb-000022
在同一栅格内，所述路径规划层结合各时间的预测位置完成最优路径并输出。响度判断模块，用于当所述均匀分布随机数rand小于所述响度A t时，更新所述初始化脉冲发射率R t为R t+1，所述响度A t值为A t+1，并重复执行方向与路径求解模块20，当所述均匀分布随机数rand的值介于所述初始化脉冲发射率R t和响度A t之间，则重复执行脉冲判断模块30。
本申请第三方面提供了一种智能化路径规划设备，包括：存储器和至少一个处理器，所述存储器中存储有指令，所述存储器和所述至少一个处理器通过线路互联；所述至少一个处理器调用所述存储器中的所述指令，以使得所述智能化路径规划设备执行上述第一方面所述的方法。
本申请的第四方面提供了一种计算机可读存储介质，所述计算机可读存储介质中存储有计算机指令，当所述计算机指令在计算机上运行时，使得计算机执行上述第一方面所述的方法。
本申请提出的智能化路径规划方法、装置、设备及存储介质，主要通过栅格法对地图进行划分，并在初始位置基于所述智能化路径规划方法开始搜索，寻找到下一个时间点的最佳位置，且根据均匀分布随机数判断所述最佳位置是否满足要求，直至最后输出最优的路径规划。本申请重设了响度和脉冲频率的计算方法，可避免在进行路径规划时陷入局部最优，且引入梯度下降算法提高 路径规划速度。因此本申请可以实现精准的路径规划功能。
附图说明
图1为本申请一实施例提供的智能化路径规划方法的流程示意图；
图2为本申请一实施例提供的智能化路径规划装置的内部结构示意图；
图3为本申请一实施例提供的智能化路径规划装置中智能化路径规划程序的模块示意图。
具体实施方式
本申请实施例提供了一种智能化路径规划方法、装置、设备及存储介质，主要用于通过栅格法对地图进行划分，并在初始位置基于所述智能化路径规划方法开始搜索，寻找到下一个时间点的最佳位置，且根据均匀分布随机数判断所述最佳位置是否满足要求，直至最后输出最优的路径规划。本申请重设了响度和脉冲频率的计算方法，可避免在进行路径规划时陷入局部最优，且引入梯度下降算法提高路径规划速度。因此本申请可以实现精准的路径规划功能。
为了使本技术领域的人员更好地理解本申请方案，下面将结合本申请实施例中的附图，对本申请实施例进行描述。
为便于理解，下面对本申请实施例的具体流程进行描述，参照图1所示，为本申请一实施例提供的智能化路径规划方法的流程示意图。该方法可以由一个装置执行，该装置可以由软件和/或硬件实现。
在本实施例中，智能化路径规划方法包括：
S1、图片处理层接收已标记出障碍物的地图图片和初始化速度v t，其中t表示时间，并基于栅格法对所述已标记出障碍物的地图图片进行栅格划分得到栅格地图，并在所述栅格地图内设置路径的初始位置
Figure PCTCN2019120802-appb-000023
与终点位置
Figure PCTCN2019120802-appb-000024
并随机产生脉冲频率f t、脉冲发射率R t和响度A t的值，将所述栅格地图、所述初始化速度v t、所述脉冲频率f t、脉冲发射率R t和响度A t的值输入至路径规划层。
所述已标记出障碍物的地图图片的获得方法包括：先将原始地图图片像素映射至[0-155]，用[155-200]内的像素集标记出障碍物。
本申请较佳实施例中，所述图片处理层对所述已标记出障碍物的地图图片进行单元分割，得到多个单元，所述多个单元为大小固定且相同的方块，对所述多个单元进行图像预处理操作，所述图像预处理操作包括膨胀、腐蚀与二值化等处理。其中，所述膨胀处理包括使用3*3的结构元素与所述已标记出障碍物的地图图片内的每个像素做“或”操作；所述腐蚀处理包括使用3*3的结构元素与所述已标记出障碍物的地图图片内的每个像素做“与”操作；所述二值化操作包括将所述已标记出障碍物的地图图片内的每个像素的像素值变为0或255。
本申请较佳实施例对所述图像预处理操作完成的多个单元使用摩尔领域跟踪算法提取障碍物的轮廓，得到所述栅格地图。所述摩尔领域跟踪算法也称间接邻域算法，基本思想为找到一个黑色像素，并将它定义为起始像素。其中， 定位所述起始像素可用多种方式，其中一种方法为：从所述多个单元的左下角像素开始，自下而上扫描每一列像素直至最上方像素，然后依据所述最上方像素开始，自左向右扫描每一列像素，直至遇到一个黑色的像素，并将其作为起始像素。
S2、所述路径规划层将所述时间t更新至t+1，并基于梯度下降算法求解所述初始栅格内的最优方向解x *，结合所述最优方向解x *预测所述初始位置
Figure PCTCN2019120802-appb-000025
在t+1时间的预测位置为
Figure PCTCN2019120802-appb-000026
所述脉冲频率f t为f t+1，所述初始化速度v t为v t+1。
本申请较佳实施例中，所述梯度下降算法包含损失函数与迭代函数，其中，所述损失函数为：
Figure PCTCN2019120802-appb-000027
所述迭代函数为：
Figure PCTCN2019120802-appb-000028
其中，θ为所述梯度下降算法的迭代参数，T表示矩阵的转置，δ为迭代函数的噪声参数。
本申请较佳实施例中，所述初始位置
Figure PCTCN2019120802-appb-000029
在t+1时间预测后的位置为
Figure PCTCN2019120802-appb-000030
Figure PCTCN2019120802-appb-000031
所述脉冲频率预测为：
f t+1＝f min+(f max-f min)β
所述初始化速度预测为：
Figure PCTCN2019120802-appb-000032
其中，β为[0,1]内产生的随机数，f min，f max分别表示脉冲频率的最小值与最大值。
S3、所述路径规划层利用反馈位移寄存发生器产生均匀分布随机数rand，判断所述均匀分布随机数rand与所述初始化脉冲发射率R t和响度A t的大小关系。
本申请较佳实施例中，所述反馈位移寄存发生器由移位寄存器和组合逻辑反馈器组成，可随机生成所述均匀分布随机数rand。
S4、判断所述均匀分布随机数rand与所述初始化脉冲发射率R t的大小关系。
S5、当所述均匀分布随机数rand小于所述初始化脉冲发射率R t时，进一步判断所述均匀分布随机数rand与所述响度A t的大小关系。
S6、当所述均匀分布随机数rand小于所述响度A t时，更新所述脉冲发射率R t为R t+1，更新所述响度A t值为A t+1，并返回S4继续判断所述均匀分布随机数rand与所述初始化脉冲发射率R t的大小关系。当所述均匀分布随机数rand大于所述响度A t时，返回S3，重新生成一个均匀分布随机数rand。
本申请较佳实施例利用下述函数更新所述初始化脉冲发射率R t为R t+1：
R t+1＝R t(1-e -γt)
利用下述函数更新所述响度A t值为A t+1：
A t+1＝αA t
其中，γ脉冲发射率的增强系数，α为音量的衰减系数，e为无限循环无理数。
S7、当所述均匀分布随机数rand大于所述初始化脉冲发射率R t时，判断所述预测位置
Figure PCTCN2019120802-appb-000033
与所述终点位置
Figure PCTCN2019120802-appb-000034
是否在同一栅格内。
S8、当所述预测位置
Figure PCTCN2019120802-appb-000035
与所述终点位置
Figure PCTCN2019120802-appb-000036
不在同一栅格内，则更新当前位置为所述预测位置
Figure PCTCN2019120802-appb-000037
当前脉冲频率为f t+1，当前初始化速度为v t+1，并返回S2。
S9、当所述预测位置
Figure PCTCN2019120802-appb-000038
与所述终点位置
Figure PCTCN2019120802-appb-000039
在同一栅格内，所述路径规划层结合各时间的预测位置完成最优路径并输出。
本申请较佳实施例，如t时间的位置为
Figure PCTCN2019120802-appb-000040
t+1时间的位置
Figure PCTCN2019120802-appb-000041
t+n时间的位置
Figure PCTCN2019120802-appb-000042
结合所有时间的位置，可确定所述路径最优规划并输出。
本申请还提供一种智能化路径规划装置。参照图2所示，为本申请一实施例提供的智能化路径规划装置的内部结构示意图。
在本实施例中，所述智能化路径规划装置1可以是PC(Personal Computer，个人电脑)，或者是智能手机、平板电脑、便携计算机等终端设备，也可以是一种服务器等。该智能化路径规划装置1至少包括存储器11、处理器12，通信总线13，以及网络接口14。
其中，存储器11至少包括一种类型的可读存储介质，所述可读存储介质包括闪存、硬盘、多媒体卡、卡型存储器(例如，SD或DX存储器等)、磁性存储器、磁盘、光盘等。存储器11在一些实施例中可以是智能化路径规划装置1的内部存储单元，例如该智能化路径规划装置1的硬盘。存储器11在另一些实施例中也可以是智能化路径规划装置1的外部存储设备，例如智能化路径规划装置1上配备的插接式硬盘，智能存储卡(Smart Media Card，SMC)，安全数字(Secure Digital，SD)卡，闪存卡(Flash Card)等。进一步地，存储器11还可以既包括智能化路径规划装置1的内部存储单元也包括外部存储设备。存储器11不仅可以用于存储安装于智能化路径规划装置1的应用软件及各类数据，例如智能化路径规划程序01的代码等，还可以用于暂时地存储已经输出或者将要输出的数据。
处理器12在一些实施例中可以是一中央处理器(Central Processing Unit，CPU)、控制器、微控制器、微处理器或其他数据处理芯片，用于运行存储器11中存储的程序代码或处理数据，例如执行智能化路径规划程序01等。
通信总线13用于实现这些组件之间的连接通信。
网络接口14可选的可以包括标准的有线接口、无线接口(如WI-FI接口)， 通常用于在该装置1与其他电子设备之间建立通信连接。
可选地，该装置1还可以包括用户接口，用户接口可以包括显示器(Display)、输入单元比如键盘(Keyboard)，可选的用户接口还可以包括标准的有线接口、无线接口。可选地，在一些实施例中，显示器可以是LED显示器、液晶显示器、触控式液晶显示器以及OLED(Organic Light-Emitting Diode，有机发光二极管)触摸器等。其中，显示器也可以适当的称为显示屏或显示单元，用于显示在智能化路径规划装置1中处理的信息以及用于显示可视化的用户界面。
图2仅示出了具有组件11-14以及智能化路径规划程序01的智能化路径规划装置1，本领域技术人员可以理解的是，图1示出的结构并不构成对智能化路径规划装置1的限定，可以包括比图示更少或者更多的部件，或者组合某些部件，或者不同的部件布置。
在图2所示的装置1实施例中，存储器11中存储有智能化路径规划程序01；处理器12执行存储器11中存储的智能化路径规划程序01时实现如下步骤：
步骤一、图片处理层接收已标记出障碍物的地图图片和初始化速度v t，其中t表示时间，并基于栅格法对所述已标记出障碍物的地图图片进行栅格划分得到栅格地图，并在所述栅格地图内设置路径的初始位置
Figure PCTCN2019120802-appb-000043
与终点位置
Figure PCTCN2019120802-appb-000044
并随机产生脉冲频率f t、脉冲发射率R t和响度A t的值，将所述栅格地图、所述初始化速度v t、所述脉冲频率f t、脉冲发射率R t和响度A t的值输入至路径规划层。
所述已标记出障碍物的地图图片的获得方法包括：先将原始地图图片像素映射至[0-155]，用[155-200]内的像素集标记出障碍物。
本申请较佳实施例中，所述图片处理层对所述已标记出障碍物的地图图片进行单元分割，得到多个单元，所述多个单元为大小固定且相同的方块，对所述多个单元进行图像预处理操作，所述图像预处理操作包括膨胀、腐蚀与二值化等处理。其中，所述膨胀处理包括使用3*3的结构元素与所述已标记出障碍物的地图图片内的每个像素做“或”操作；所述腐蚀处理包括使用3*3的结构元素与所述已标记出障碍物的地图图片内的每个像素做“与”操作；所述二值化操作包括将所述已标记出障碍物的地图图片内的每个像素的像素值变为0或255。
本申请较佳实施例对所述图像预处理操作完成的多个单元使用摩尔领域跟踪算法提取障碍物的轮廓，得到所述栅格地图。所述摩尔领域跟踪算法也称间接邻域算法，基本思想为找到一个黑色像素，并将它定义为起始像素。其中，定位所述起始像素可用多种方式，其中一种方法为：从所述多个单元的左下角像素开始，自下而上扫描每一列像素直至最上方像素，然后依据所述最上方像素开始，自左向右扫描每一列像素，直至遇到一个黑色的像素，并将其作为起始像素。
步骤二、所述路径规划层将所述时间t更新至t+1，并基于梯度下降算法 求解所述初始栅格内的最优方向解x *，结合所述最优方向解x *预测所述初始位置
Figure PCTCN2019120802-appb-000045
在t+1时间的预测位置为
Figure PCTCN2019120802-appb-000046
所述脉冲频率f t为f t+1，所述初始化速度v t为v t+1。
本申请较佳实施例中，所述梯度下降算法包含损失函数与迭代函数，其中，所述损失函数为：
Figure PCTCN2019120802-appb-000047
所述迭代函数为：
Figure PCTCN2019120802-appb-000048
其中，θ为所述梯度下降算法的迭代参数，T表示矩阵的转置，δ为迭代函数的噪声参数。
本申请较佳实施例中，所述初始位置
Figure PCTCN2019120802-appb-000049
在t+1时间预测后的位置为
Figure PCTCN2019120802-appb-000050
Figure PCTCN2019120802-appb-000051
所述脉冲频率预测为：
f t+1＝f min+(f max-f min)β
所述初始化速度预测为：
Figure PCTCN2019120802-appb-000052
其中，β为[0,1]内产生的随机数，f min，f max分别表示脉冲频率的最小值与最大值。
步骤三、所述路径规划层利用反馈位移寄存发生器产生均匀分布随机数rand，判断所述均匀分布随机数rand与所述初始化脉冲发射率R t和响度A t的大小关系。
本申请较佳实施例中，所述反馈位移寄存发生器由移位寄存器和组合逻辑反馈器组成，可随机生成所述均匀分布随机数rand。
步骤四、判断所述均匀分布随机数rand与所述初始化脉冲发射率R t的大小关系。
步骤五、当所述均匀分布随机数rand小于所述初始化脉冲发射率R t时，进一步判断所述均匀分布随机数rand与所述响度A t的大小关系。
步骤六、当所述均匀分布随机数rand小于所述响度A t时，更新所述脉冲发射率R t为R t+1，更新所述响度A t值为A t+1，并返回步骤四继续判断所述均匀分布随机数rand与所述初始化脉冲发射率R t的大小关系。当所述均匀分布随机数rand大于所述响度A t时，返回步骤三，重新生成一个均匀分布随机数rand。
本申请较佳实施例利用下述函数更新所述初始化脉冲发射率R t为R t+1：
R t+1＝R t(1-e -γt)
利用下述函数更新所述响度A t值为A t+1：
A t+1＝αA t
其中，γ脉冲发射率的增强系数，α为音量的衰减系数，e为无限循环无理数。
步骤七、当所述均匀分布随机数rand大于所述初始化脉冲发射率R t时，判断所述预测位置
Figure PCTCN2019120802-appb-000053
与所述终点位置
Figure PCTCN2019120802-appb-000054
是否在同一栅格内。
步骤八、当所述预测位置
Figure PCTCN2019120802-appb-000055
与所述终点位置
Figure PCTCN2019120802-appb-000056
不在同一栅格内，则更新当前位置为所述预测位置
Figure PCTCN2019120802-appb-000057
当前脉冲频率为f t+1，当前初始化速度为v t+1，并返回步骤二。
步骤九、当所述预测位置
Figure PCTCN2019120802-appb-000058
与所述终点位置
Figure PCTCN2019120802-appb-000059
在同一栅格内，所述路径规划层结合各时间的预测位置完成最优路径并输出。
本申请较佳实施例，如t时间的位置为
Figure PCTCN2019120802-appb-000060
t+1时间的位置
Figure PCTCN2019120802-appb-000061
t+n时间的位置
Figure PCTCN2019120802-appb-000062
结合所有时间的位置，可确定所述路径最优规划并输出。
可选地，在其他实施例中，智能化路径规划程序还可以被分割为一个或者多个模块，一个或者多个模块被存储于存储器11中，并由一个或多个处理器(本实施例为处理器12)所执行以完成本申请，本申请所称的模块是指能够完成特定功能的一系列计算机程序指令段，用于描述智能化路径规划程序在智能化路径规划装置中的执行过程。
例如，参照图3所示，为本申请智能化路径规划方法装置一实施例中的智能化路径规划方法程序的程序模块示意图，该实施例中，所述智能化路径规划方法程序可以被分割为地图图片接收模块10、方向与路径求解模块20、脉冲判断并输出路径模块30、响度判断模块40示例性地：
所述地图图片接收模块10用于：接收已标记出障碍物的地图图片和初始化速度v t，其中t表示时间，并基于栅格法对所述已标记出障碍物的地图图片进行栅格划分得到栅格地图，并在所述栅格地图内设置路径的初始位置
Figure PCTCN2019120802-appb-000063
与终点位置
Figure PCTCN2019120802-appb-000064
并随机产生脉冲频率f t、脉冲发射率R t和响度A t的值。
所述方向与路径求解模块20用于：将所述时间t更新至t+1，并基于梯度下降算法求解所述初始栅格内的最优方向解x *，结合所述最优方向解x *预测所述初始位置
Figure PCTCN2019120802-appb-000065
在t+1时间的预测位置为
Figure PCTCN2019120802-appb-000066
所述脉冲频率f t为f t+1，所述初始化速度v t为v t+1。
所述脉冲判断并输出路径模块30用于：利用反馈位移寄存发生器产生均匀分布随机数rand，判断所述均匀分布随机数rand与所述初始化脉冲发射率R t和响度A t的大小关系，当所述均匀分布随机数rand小于所述初始化脉冲发射率R t时，执行响度判断模块40，当所述均匀分布随机数rand大于所述初始化脉冲发射率R t时，判断所述预测位置
Figure PCTCN2019120802-appb-000067
与所述终点位置
Figure PCTCN2019120802-appb-000068
是否在同一栅格内，当所述预测位置
Figure PCTCN2019120802-appb-000069
与所述终点位置
Figure PCTCN2019120802-appb-000070
不在同一栅格内，更新当前位置为所述预测位置
Figure PCTCN2019120802-appb-000071
当前脉冲频率为f t+1，当前初始化速度为v t+1，并重复方向与路径求解模块20和脉冲判断模块30，当所述预测位置
Figure PCTCN2019120802-appb-000072
与所述终点位 置
Figure PCTCN2019120802-appb-000073
在同一栅格内，所述路径规划层结合各时间的预测位置完成最优路径并输出。
所述响度判断模块40用于：当所述均匀分布随机数rand小于所述响度A t时，更新所述初始化脉冲发射率R t为R t+1，所述响度A t值为A t+1，并重复执行方向与路径求解模块20，当所述均匀分布随机数rand的值介于所述初始化脉冲发射率R t和响度A t之间，则重复执行脉冲判断模块30。
上述地图图片接收模块10、方向与路径求解模块20、脉冲判断并输出路径模块30、响度判断模块40等程序模块被执行时所实现的功能或操作步骤与上述实施例大体相同，在此不再赘述。
此外，本申请实施例还提出一种计算机可读存储介质，所述计算机可读存储介质上存储有智能化路径规划程序，所述智能化路径规划程序可被一个或多个处理器执行，以实现如下操作：
步骤A：图片处理层接收已标记出障碍物的地图图片和初始化速度v t，其中t表示时间，并基于栅格法对所述已标记出障碍物的地图图片进行栅格划分得到栅格地图，并在所述栅格地图内设置路径的初始位置
Figure PCTCN2019120802-appb-000074
与终点位置
Figure PCTCN2019120802-appb-000075
并随机产生脉冲频率f t、脉冲发射率R t和响度A t的值，将所述栅格地图、所述初始化速度v t、所述脉冲频率f t、脉冲发射率R t和响度A t的值输入至路径规划层；
步骤B：所述路径规划层将所述时间t更新至t+1，并基于梯度下降算法求解所述初始栅格内的最优方向解x *，结合所述最优方向解x *预测所述初始位置
Figure PCTCN2019120802-appb-000076
在t+1时间的预测位置为
Figure PCTCN2019120802-appb-000077
所述脉冲频率f t为f t+1，所述初始化速度v t为v t+1；
步骤C：所述路径规划层利用反馈位移寄存发生器产生均匀分布随机数rand，判断所述均匀分布随机数rand与所述初始化脉冲发射率R t和响度A t的大小关系；
步骤D：当所述均匀分布随机数rand小于所述初始化脉冲发射率R t时，执行步骤E：当所述均匀分布随机数rand小于所述响度A t时，更新所述初始化脉冲发射率R t为R t+1，所述响度A t值为A t+1，并重复步骤D，当所述均匀分布随机数rand的值介于所述初始化脉冲发射率R t和响度A t之间，则重复步骤C与步骤D；
当所述均匀分布随机数rand大于所述初始化脉冲发射率R t时，判断所述预测位置
Figure PCTCN2019120802-appb-000078
与所述终点位置
Figure PCTCN2019120802-appb-000079
是否在同一栅格内，当所述预测位置
Figure PCTCN2019120802-appb-000080
与所述终点位置
Figure PCTCN2019120802-appb-000081
不在同一栅格内，更新当前位置为所述预测位置
Figure PCTCN2019120802-appb-000082
当前脉冲频率为f t+1及当前初始化速度为v t+1，并重复步骤B和步骤C，当所述预测位置
Figure PCTCN2019120802-appb-000083
与所述终点位置
Figure PCTCN2019120802-appb-000084
在同一栅格内，所述路径规划层结合各时间的预测位置完成最优路径并输出。
本申请还提供一种智能化路径规划设备，包括：存储器和至少一个处理器，所述存储器中存储有指令，所述存储器和所述至少一个处理器通过线路互联； 所述至少一个处理器调用所述存储器中的所述指令，以使得所述智能化路径规划设备执行上述智能化路径规划方法中的步骤。
本申请还提供一种计算机可读存储介质，该计算机可读存储介质可以为非易失性计算机可读存储介质，也可以为易失性计算机可读存储介质。计算机可读存储介质存储有计算机指令，当所述计算机指令在计算机上运行时，使得计算机执行如下步骤：
步骤A：图片处理层接收已标记出障碍物的地图图片和初始化速度v t，其中t表示时间，并基于栅格法对所述已标记出障碍物的地图图片进行栅格划分得到栅格地图，并在所述栅格地图内设置路径的初始位置
Figure PCTCN2019120802-appb-000085
与终点位置
Figure PCTCN2019120802-appb-000086
并随机产生脉冲频率f t、脉冲发射率R t和响度A t的值，将所述栅格地图、所述初始化速度v t、所述脉冲频率f t、脉冲发射率R t和响度A t的值输入至路径规划层；
步骤B：所述路径规划层将所述时间t更新至t+1，并基于梯度下降算法求解所述初始栅格内的最优方向解x *，结合所述最优方向解x *预测所述初始位置
Figure PCTCN2019120802-appb-000087
在t+1时间的预测位置为
Figure PCTCN2019120802-appb-000088
所述脉冲频率f t为f t+1，所述初始化速度v t为v t+1；
步骤C：所述路径规划层利用反馈位移寄存发生器产生均匀分布随机数rand，判断所述均匀分布随机数rand与所述初始化脉冲发射率R t和响度A t的大小关系；
步骤D：当所述均匀分布随机数rand小于所述初始化脉冲发射率R t时，执行步骤E：当所述均匀分布随机数rand小于所述响度A t时，更新所述初始化脉冲发射率R t为R t+1，所述响度A t值为A t+1，并重复步骤D，当所述均匀分布随机数rand的值介于所述初始化脉冲发射率R t和响度A t之间，则重复步骤C与步骤D；
当所述均匀分布随机数rand大于所述初始化脉冲发射率R t时，判断所述预测位置
Figure PCTCN2019120802-appb-000089
与所述终点位置
Figure PCTCN2019120802-appb-000090
是否在同一栅格内，当所述预测位置
Figure PCTCN2019120802-appb-000091
与所述终点位置
Figure PCTCN2019120802-appb-000092
不在同一栅格内，更新当前位置为所述预测位置
Figure PCTCN2019120802-appb-000093
当前脉冲频率为f t+1及当前初始化速度为v t+1，并重复步骤B和步骤C，当所述预测位置
Figure PCTCN2019120802-appb-000094
与所述终点位置
Figure PCTCN2019120802-appb-000095
在同一栅格内，所述路径规划层结合各时间的预测位置完成最优路径并输出。
所属领域的技术人员可以清楚地了解到，为描述的方便和简洁，上述描述的系统，装置和单元的具体工作过程，可以参考前述方法实施例中的对应过程，在此不再赘述。
在本申请所提供的几个实施例中，应该理解到，所揭露的系统，装置和方法，可以通过其它的方式实现。例如，以上所描述的装置实施例仅仅是示意性的，例如，所述单元的划分，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式，例如多个单元或组件可以结合或者可以集成到另一个系统，或一些特征可以忽略，或不执行。另一点，所显示或讨论的相互之间的耦合或直 接耦合或通信连接可以是通过一些接口，装置或单元的间接耦合或通信连接，可以是电性，机械或其它的形式。
以上所述，以上实施例仅用以说明本申请的技术方案，而非对其限制；尽管参照前述实施例对本申请进行了详细的说明，本领域的普通技术人员应当理解：其依然可以对前述各实施例所记载的技术方案进行修改，或者对其中部分技术特征进行等同替换；而这些修改或者替换，并不使相应技术方案的本质脱离本申请各实施例技术方案的精神和范围。

Claims (20)

  1. 一种智能化路径规划方法，包括：
    步骤A：图片处理层接收已标记出障碍物的地图图片和初始化速度v t，其中t表示时间，并基于栅格法对所述已标记出障碍物的地图图片进行栅格划分得到栅格地图，并在所述栅格地图内设置路径的初始位置
    Figure PCTCN2019120802-appb-100001
    与终点位置
    Figure PCTCN2019120802-appb-100002
    并随机产生脉冲频率f t、脉冲发射率R t和响度A t的值，将所述栅格地图、所述初始化速度v t、所述脉冲频率f t、脉冲发射率R t和响度A t的值输入至路径规划层；
    步骤B：所述路径规划层将所述时间t更新至t+1，并基于梯度下降算法求解所述初始栅格内的最优方向解x *，结合所述最优方向解x *预测所述初始位置
    Figure PCTCN2019120802-appb-100003
    在t+1时间的预测位置为
    Figure PCTCN2019120802-appb-100004
    所述脉冲频率f t为f t+1，所述初始化速度v t为v t+1；
    步骤C：所述路径规划层利用反馈位移寄存发生器产生均匀分布随机数rand，判断所述均匀分布随机数rand与所述初始化脉冲发射率R t和响度A t的大小关系；
    步骤D：当所述均匀分布随机数rand小于所述初始化脉冲发射率R t时，执行步骤E：当所述均匀分布随机数rand小于所述响度A t时，更新所述初始化脉冲发射率R t为R t+1，所述响度A t值为A t+1，并重复步骤D，当所述均匀分布随机数rand的值介于所述初始化脉冲发射率R t和响度A t之间，则重复步骤C与步骤D；
    当所述均匀分布随机数rand大于所述初始化脉冲发射率R t时，判断所述预测位置
    Figure PCTCN2019120802-appb-100005
    与所述终点位置
    Figure PCTCN2019120802-appb-100006
    是否在同一栅格内，当所述预测位置
    Figure PCTCN2019120802-appb-100007
    与所述终点位置
    Figure PCTCN2019120802-appb-100008
    不在同一栅格内，更新当前位置为所述预测位置
    Figure PCTCN2019120802-appb-100009
    当前脉冲频率为f t+1及当前初始化速度为v t+1，并重复步骤B和步骤C，当所述预测位置
    Figure PCTCN2019120802-appb-100010
    与所述终点位置
    Figure PCTCN2019120802-appb-100011
    在同一栅格内，所述路径规划层结合各时间的预测位置完成最优路径并输出。
  2. 根据权利要求1所述的智能化路径规划方法，所述基于栅格法对所述已标记出障碍物的地图图片进行栅格划分得到栅格地图，包括：
    所述图片处理层对所述已标记出障碍物的地图图片进行单元分割，得到多个单元，其中，所述多个单元为大小固定且相同的方块；
    对所述多个单元进行图像预处理操作，所述图像预处理操作包括膨胀、腐蚀与二值化；
    对所述图像预处理操作完成的多个单元使用摩尔领域跟踪算法提取障碍物的轮廓，得到所述栅格地图。
  3. 根据权利要求1所述的智能化路径规划方法，所述智能化路径规划方法包括：
    预测的所述初始位置
    Figure PCTCN2019120802-appb-100012
    的t+1时间位置为：
    Figure PCTCN2019120802-appb-100013
    预测的所述脉冲频率为：
    f t+1＝f min+(f max-f min)β；
    预测的所述初始化速度为：
    Figure PCTCN2019120802-appb-100014
    其中，β为[0,1]内产生的随机数，f min，f max分别表示脉冲频率的最小值与最大值。
  4. 根据权利要求1-3中任一项所述的智能化路径规划方法，所述智能化路径规划方法包括：
    所述梯度下降算法包含损失函数与迭代函数；
    所述损失函数为：
    Figure PCTCN2019120802-appb-100015
    所述迭代函数为：
    Figure PCTCN2019120802-appb-100016
    其中，θ为所述梯度下降算法的迭代参数，T表示矩阵的转置，δ为迭代函数的噪声参数。
  5. 根据权利要求1所述的智能化路径规划方法，所述智能化路径规划方法包括：
    更新的所述初始化脉冲发射率为：
    R t+1＝R t(1-e -γt)；
    更新的所述响度为：
    A t+1＝αA t；
    其中，γ脉冲发射率的增强系数，α为音量的衰减系数，e为无限循环无理数。
  6. 一种智能化路径规划装置，所述智能化路径规划装置包括：
    地图图片接收模块，用于接收已标记出障碍物的地图图片和初始化速度v t，其中t表示时间，并基于栅格法对所述已标记出障碍物的地图图片进行栅格划分得到栅格地图，并在所述栅格地图内设置路径的初始位置
    Figure PCTCN2019120802-appb-100017
    与终点位置
    Figure PCTCN2019120802-appb-100018
    并随机产生脉冲频率f t、脉冲发射率R t和响度A t的值。
    方向与路径求解模块，用于将所述时间t更新至t+1，并基于梯度下降算法求解所述初始栅格内的最优方向解x *，结合所述最优方向解x *预测所述初始位置
    Figure PCTCN2019120802-appb-100019
    在t+1时间的预测位置为
    Figure PCTCN2019120802-appb-100020
    所述脉冲频率f t为f t+1，所述初始化速度v t为v t+1。
    脉冲判断并输出路径模块，用于利用反馈位移寄存发生器产生均匀分布随机数rand，判断所述均匀分布随机数rand与所述初始化脉冲发射率R t和响度A t的大小关系，当所述均匀分布随机数rand小于所述初始化脉冲发射率R t时， 执行响度判断模块40，当所述均匀分布随机数rand大于所述初始化脉冲发射率R t时，判断所述预测位置
    Figure PCTCN2019120802-appb-100021
    与所述终点位置
    Figure PCTCN2019120802-appb-100022
    是否在同一栅格内，当所述预测位置
    Figure PCTCN2019120802-appb-100023
    与所述终点位置
    Figure PCTCN2019120802-appb-100024
    不在同一栅格内，更新当前位置为所述预测位置
    Figure PCTCN2019120802-appb-100025
    当前脉冲频率为f t+1，当前初始化速度为v t+1，并重复方向与路径求解模块20和脉冲判断模块30，当所述预测位置
    Figure PCTCN2019120802-appb-100026
    与所述终点位置
    Figure PCTCN2019120802-appb-100027
    在同一栅格内，所述路径规划层结合各时间的预测位置完成最优路径并输出。
    响度判断模块，用于当所述均匀分布随机数rand小于所述响度A t时，更新所述初始化脉冲发射率R t为R t+1，所述响度A t值为A t+1，并重复执行方向与路径求解模块20，当所述均匀分布随机数rand的值介于所述初始化脉冲发射率R t和响度A t之间，则重复执行脉冲判断模块30。
  7. 根据权利要求6所述的智能化路径规划装置，所述地图图片接收模块具体用于：
    所述图片处理层对所述已标记出障碍物的地图图片进行单元分割，得到多个单元，其中，所述多个单元为大小固定且相同的方块；
    对所述多个单元进行图像预处理操作，所述图像预处理操作包括膨胀、腐蚀与二值化；
    对所述图像预处理操作完成的多个单元使用摩尔领域跟踪算法提取障碍物的轮廓，得到所述栅格地图。
  8. 根据权利要求6所述的智能化路径规划装置，所述智能化路径规划装置具体用于：
    预测的所述初始位置
    Figure PCTCN2019120802-appb-100028
    的t+1时间位置为：
    Figure PCTCN2019120802-appb-100029
    预测的所述脉冲频率为：
    f t+1＝f min+(f max-f min)β；
    预测的所述初始化速度为：
    Figure PCTCN2019120802-appb-100030
    其中，β为[0,1]内产生的随机数，f min，f max分别表示脉冲频率的最小值与最大值。
  9. 根据权利要求6-8中任一项所述的智能化路径规划装置，所述智能化路径规划装置具体用于：
    所述损失函数为：
    Figure PCTCN2019120802-appb-100031
    所述迭代函数为：
    Figure PCTCN2019120802-appb-100032
    其中，θ为所述梯度下降算法的迭代参数，T表示矩阵的转置，δ为迭代函数的噪声参数。
  10. 根据权利要求1所述的智能化路径规划装置，所述智能化路径规划装置具体用于：
    更新的所述初始化脉冲发射率为：
    R t+1＝R t(1-e -γt)；
    更新的所述响度为：
    A t+1＝αA t；
    其中，γ脉冲发射率的增强系数，α为音量的衰减系数，e为无限循环无理数。
  11. 一种智能化路径规划设备，包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机可读指令，所述处理器执行所述计算机可读指令时，实现如下步骤：
    步骤A：图片处理层接收已标记出障碍物的地图图片和初始化速度v t，其中t表示时间，并基于栅格法对所述已标记出障碍物的地图图片进行栅格划分得到栅格地图，并在所述栅格地图内设置路径的初始位置
    Figure PCTCN2019120802-appb-100033
    与终点位置
    Figure PCTCN2019120802-appb-100034
    并随机产生脉冲频率f t、脉冲发射率R t和响度A t的值，将所述栅格地图、所述初始化速度v t、所述脉冲频率f t、脉冲发射率R t和响度A t的值输入至路径规划层；
    步骤B：所述路径规划层将所述时间t更新至t+1，并基于梯度下降算法求解所述初始栅格内的最优方向解x *，结合所述最优方向解x *预测所述初始位置
    Figure PCTCN2019120802-appb-100035
    在t+1时间的预测位置为
    Figure PCTCN2019120802-appb-100036
    所述脉冲频率f t为f t+1，所述初始化速度v t为v t+1；
    步骤C：所述路径规划层利用反馈位移寄存发生器产生均匀分布随机数rand，判断所述均匀分布随机数rand与所述初始化脉冲发射率R t和响度A t的大小关系；
    步骤D：当所述均匀分布随机数rand小于所述初始化脉冲发射率R t时，执行步骤E：当所述均匀分布随机数rand小于所述响度A t时，更新所述初始化脉冲发射率R t为R t+1，所述响度A t值为A t+1，并重复步骤D，当所述均匀分布随机数rand的值介于所述初始化脉冲发射率R t和响度A t之间，则重复步骤C与步骤D；
    当所述均匀分布随机数rand大于所述初始化脉冲发射率R t时，判断所述预测位置
    Figure PCTCN2019120802-appb-100037
    与所述终点位置
    Figure PCTCN2019120802-appb-100038
    是否在同一栅格内，当所述预测位置
    Figure PCTCN2019120802-appb-100039
    与所述终点位置
    Figure PCTCN2019120802-appb-100040
    不在同一栅格内，更新当前位置为所述预测位置
    Figure PCTCN2019120802-appb-100041
    当前脉冲频率为f t+1及当前初始化速度为v t+1，并重复步骤B和步骤C，当所述预测位置
    Figure PCTCN2019120802-appb-100042
    与所述终点位置
    Figure PCTCN2019120802-appb-100043
    在同一栅格内，所述路径规划层结合各时间的预测位置完成最优路径并输出。
  12. 根据权利要求11所述的智能化路径规划设备，所述处理器执行所述计算机可读指令实现所述基于栅格法对所述已标记出障碍物的地图图片进行栅格划分得到栅格地图时，包括以下步骤：
    所述图片处理层对所述已标记出障碍物的地图图片进行单元分割，得到多个单元，其中，所述多个单元为大小固定且相同的方块；
    对所述多个单元进行图像预处理操作，所述图像预处理操作包括膨胀、腐蚀与二值化；
    对所述图像预处理操作完成的多个单元使用摩尔领域跟踪算法提取障碍物的轮廓，得到所述栅格地图。
  13. 根据权利要求11所述的智能化路径规划设备，所述处理器执行所述计算机可读指令实现所述智能化路径规划方法的步骤时，包括以下步骤：
    预测的所述初始位置
    Figure PCTCN2019120802-appb-100044
    的t+1时间位置为：
    Figure PCTCN2019120802-appb-100045
    预测的所述脉冲频率为：
    f t+1＝f min+(f max-f min)β；
    预测的所述初始化速度为：
    Figure PCTCN2019120802-appb-100046
    其中，β为[0,1]内产生的随机数，f min，f max分别表示脉冲频率的最小值与最大值。
  14. 根据权利要求11-13中任一项所述的智能化路径规划设备，所述处理器执行所述计算机可读指令实现所述智能化路径规划方法的步骤时，包括以下步骤：
    所述梯度下降算法包含损失函数与迭代函数；
    所述损失函数为：
    Figure PCTCN2019120802-appb-100047
    所述迭代函数为：
    Figure PCTCN2019120802-appb-100048
    其中，θ为所述梯度下降算法的迭代参数，T表示矩阵的转置，δ为迭代函数的噪声参数。
  15. 根据权利要求11所述的智能化路径规划设备，所述处理器执行所述计算机可读指令实现所述智能化路径规划方法的步骤时，包括以下步骤：
    更新的所述初始化脉冲发射率为：
    R t+1＝R t(1-e -γt)；
    更新的所述响度为：
    A t+1＝αA t；
    其中，γ脉冲发射率的增强系数，α为音量的衰减系数，e为无限循环无理数。
  16. 一种计算机可读存储介质，所述计算机可读存储介质中存储计算机指令，当所述计算机指令在计算机上运行时，使得计算机执行如下步骤：
    步骤A：图片处理层接收已标记出障碍物的地图图片和初始化速度v t，其中t表示时间，并基于栅格法对所述已标记出障碍物的地图图片进行栅格划分得到栅格地图，并在所述栅格地图内设置路径的初始位置
    Figure PCTCN2019120802-appb-100049
    与终点位置
    Figure PCTCN2019120802-appb-100050
    并随机产生脉冲频率f t、脉冲发射率R t和响度A t的值，将所述栅格地图、所述初始化速度v t、所述脉冲频率f t、脉冲发射率R t和响度A t的值输入至路径规划层；
    步骤B：所述路径规划层将所述时间t更新至t+1，并基于梯度下降算法求解所述初始栅格内的最优方向解x *，结合所述最优方向解x *预测所述初始位置
    Figure PCTCN2019120802-appb-100051
    在t+1时间的预测位置为
    Figure PCTCN2019120802-appb-100052
    所述脉冲频率f t为f t+1，所述初始化速度v t为v t+1；
    步骤C：所述路径规划层利用反馈位移寄存发生器产生均匀分布随机数rand，判断所述均匀分布随机数rand与所述初始化脉冲发射率R t和响度A t的大小关系；
    步骤D：当所述均匀分布随机数rand小于所述初始化脉冲发射率R t时，执行步骤E：当所述均匀分布随机数rand小于所述响度A t时，更新所述初始化脉冲发射率R t为R t+1，所述响度A t值为A t+1，并重复步骤D，当所述均匀分布随机数rand的值介于所述初始化脉冲发射率R t和响度A t之间，则重复步骤C与步骤D；
    当所述均匀分布随机数rand大于所述初始化脉冲发射率R t时，判断所述预测位置
    Figure PCTCN2019120802-appb-100053
    与所述终点位置
    Figure PCTCN2019120802-appb-100054
    是否在同一栅格内，当所述预测位置
    Figure PCTCN2019120802-appb-100055
    与所述终点位置
    Figure PCTCN2019120802-appb-100056
    不在同一栅格内，更新当前位置为所述预测位置
    Figure PCTCN2019120802-appb-100057
    当前脉冲频率为f t+1及当前初始化速度为v t+1，并重复步骤B和步骤C，当所述预测位置
    Figure PCTCN2019120802-appb-100058
    与所述终点位置
    Figure PCTCN2019120802-appb-100059
    在同一栅格内，所述路径规划层结合各时间的预测位置完成最优路径并输出。
  17. 根据权利要求16所述的计算机可读存储介质，当所述计算机指令在计算机上运行所述基于栅格法对所述已标记出障碍物的地图图片进行栅格划分得到栅格地图时，使得计算机执行如下步骤：
    所述图片处理层对所述已标记出障碍物的地图图片进行单元分割，得到多个单元，其中，所述多个单元为大小固定且相同的方块；
    对所述多个单元进行图像预处理操作，所述图像预处理操作包括膨胀、腐蚀与二值化；
    对所述图像预处理操作完成的多个单元使用摩尔领域跟踪算法提取障碍物的轮廓，得到所述栅格地图。
  18. 根据权利要求16所述的计算机可读存储介质，当所述计算机指令在计算机上运行所述智能化路径规划方法的步骤时，使得计算机执行如下步骤：预测的所述初始位置
    Figure PCTCN2019120802-appb-100060
    的t+1时间位置为：
    Figure PCTCN2019120802-appb-100061
    预测的所述脉冲频率为：
    f t+1＝f min+(f max-f min)β；
    预测的所述初始化速度为：
    Figure PCTCN2019120802-appb-100062
    其中，β为[0,1]内产生的随机数，f min，f max分别表示脉冲频率的最小值与最大值。
  19. 根据权利要求16-18中任一项所述的计算机可读存储介质，当所述计算机指令在计算机上运行所述智能化路径规划方法的步骤时，使得计算机执行如下步骤：
    所述梯度下降算法包含损失函数与迭代函数；
    所述损失函数为：
    Figure PCTCN2019120802-appb-100063
    所述迭代函数为：
    Figure PCTCN2019120802-appb-100064
    其中，θ为所述梯度下降算法的迭代参数，T表示矩阵的转置，δ为迭代函数的噪声参数。
  20. 根据权利要求16所述的计算机可读存储介质，当所述计算机指令在计算机上运行所述智能化路径规划方法的步骤时，使得计算机执行如下步骤：
    更新的所述初始化脉冲发射率为：
    R t+1＝R t(1-e -γt)；
    更新的所述响度为：
    A t+1＝αA t；
    其中，γ脉冲发射率的增强系数，α为音量的衰减系数，e为无限循环无理数。
PCT/CN2019/120802 2019-05-21 2019-11-26 智能化路径规划方法、装置、设备及存储介质 WO2020233052A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910421933.0 2019-05-21
CN201910421933.0A CN110274608A (zh) 2019-05-21 2019-05-21 智能化路径规划方法、装置及计算机可读存储介质

Publications (1)

Publication Number Publication Date
WO2020233052A1 true WO2020233052A1 (zh) 2020-11-26

Family

ID=67959432

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/120802 WO2020233052A1 (zh) 2019-05-21 2019-11-26 智能化路径规划方法、装置、设备及存储介质

Country Status (2)

Country Link
CN (1) CN110274608A (zh)
WO (1) WO2020233052A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110274608A (zh) * 2019-05-21 2019-09-24 深圳壹账通智能科技有限公司 智能化路径规划方法、装置及计算机可读存储介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8185239B2 (en) * 2008-11-13 2012-05-22 MSI Computer (Shenzhen) Co, Ltd. Moving route planning method and navigation method for avoiding dynamic hindrances for mobile robot device
CN106323293A (zh) * 2016-10-14 2017-01-11 淮安信息职业技术学院 基于多目标搜索的两群多向机器人路径规划方法
CN110274608A (zh) * 2019-05-21 2019-09-24 深圳壹账通智能科技有限公司 智能化路径规划方法、装置及计算机可读存储介质
CN110274607A (zh) * 2019-05-21 2019-09-24 深圳壹账通智能科技有限公司 智能化路径规划方法、装置及计算机可读存储介质

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108694077B (zh) * 2017-04-10 2022-02-01 郑州芯兰德网络科技有限公司 基于改进二进制蝙蝠算法的分布式系统任务调度方法
CN107886158A (zh) * 2017-10-30 2018-04-06 中国地质大学（武汉） 一种基于迭代局部搜索和随机惯性权重的蝙蝠优化算法
CN107886157A (zh) * 2017-10-30 2018-04-06 中国地质大学（武汉） 一种新型蝙蝠优化算法系统
CN109144102B (zh) * 2018-09-19 2021-08-20 沈阳航空航天大学 一种基于改进蝙蝠算法的无人机航路规划方法

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8185239B2 (en) * 2008-11-13 2012-05-22 MSI Computer (Shenzhen) Co, Ltd. Moving route planning method and navigation method for avoiding dynamic hindrances for mobile robot device
CN106323293A (zh) * 2016-10-14 2017-01-11 淮安信息职业技术学院 基于多目标搜索的两群多向机器人路径规划方法
CN110274608A (zh) * 2019-05-21 2019-09-24 深圳壹账通智能科技有限公司 智能化路径规划方法、装置及计算机可读存储介质
CN110274607A (zh) * 2019-05-21 2019-09-24 深圳壹账通智能科技有限公司 智能化路径规划方法、装置及计算机可读存储介质

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
余冬冬 (YU, DONGDONG): "移动机器人避障与轨迹规划 (Obstacles avoidance and trajectory planning of mobile robots)", 中国优秀硕士学位论文全文数据库(信息科技辑) (INFORMATION SCIENCE & TECHNOLOGY, CHINA MASTER’S THESES FULL-TEXT DATABASE), no. 8, 15 August 2017 (2017-08-15), ISSN: 1674-0246, DOI: 20200224215948A *
耿双乐 等 (GENG, SHUANGLE ET AL.): "逃离传统势场法局部稳定点策略 (Mobile robot path planning based on improved traditional potential field method)", 计算机与数字工程 (COMPUTER & DIGITAL ENGINEERING), vol. 47, no. 04, 20 April 2019 (2019-04-20), ISSN: 1672-9722, DOI: 9140300 *

Also Published As

Publication number Publication date
CN110274608A (zh) 2019-09-24

Similar Documents

Publication Publication Date Title
US20230394203A1 (en) Generating integrated circuit floorplans using neural networks
US20230385668A1 (en) Systems and methods for analog processing of problem graphs having arbitrary size and/or connectivity
WO2020233063A1 (zh) 智能化路径规划方法、装置及计算机可读存储介质
WO2018049801A1 (zh) 基于深度图的启发式手指检测方法
US20170053412A1 (en) Image Depth Inference from Semantic Labels
JP2014520319A (ja) リレーショナルデータベースを視覚化するマップに基づく方法
US11057118B2 (en) Indoor localization with beacon technology based on signal strength distribution and deep learning techniques
WO2024061327A1 (zh) 矢量地图生成方法、装置、电子设备和可读存储介质
US20230306662A1 (en) Systems, methods, and devices for image processing
WO2020233052A1 (zh) 智能化路径规划方法、装置、设备及存储介质
CN114611755A (zh) 基于混合元启发式算法的旅行商最优路径确定方法
CN112330697B (zh) 一种图像分割方法、装置、电子设备和可读存储介质
US10198412B2 (en) Simulated annealing to place annotations in a drawing
Sun et al. Energy-aware run-time mapping for homogeneous NoC
CN104036141A (zh) 一种基于OpenCL的红黑树加速算法
CN104182208A (zh) 利用破解规则破解密码的方法及系统
CN116518979A (zh) 一种无人机路径规划方法、系统、电子设备及介质
CN114758130B (zh) 图像处理及模型训练方法、装置、设备和存储介质
Mei et al. Geodesic‐based probability propagation for efficient optical flow
Fratarcangeli et al. A gpu-based implementation of position based dynamics for interactive deformable bodies
US11941327B2 (en) Customizable reinforcement learning of column placement in structural design
CN107894992A (zh) 一种gis点缓冲区分析方法及系统
JP7496956B2 (ja) 画像処理のためのシステム、方法、および装置
CN103593529A (zh) 一种基于数学形态学的复杂桥梁对象自动符号化方法
US20220075555A1 (en) Multi-scale convolutional kernels for adaptive grids

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19930047

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 01/03/2022)

122 Ep: pct application non-entry in european phase

Ref document number: 19930047

Country of ref document: EP

Kind code of ref document: A1