WO2022082840A1 - 基于智能迷宫的大脑训练评价系统及其评价方法 - Google Patents

基于智能迷宫的大脑训练评价系统及其评价方法 Download PDF

Info

Publication number
WO2022082840A1
WO2022082840A1 PCT/CN2020/124836 CN2020124836W WO2022082840A1 WO 2022082840 A1 WO2022082840 A1 WO 2022082840A1 CN 2020124836 W CN2020124836 W CN 2020124836W WO 2022082840 A1 WO2022082840 A1 WO 2022082840A1
Authority
WO
WIPO (PCT)
Prior art keywords
evaluation
maze
level
score
brain wave
Prior art date
Application number
PCT/CN2020/124836
Other languages
English (en)
French (fr)
Inventor
李华京
陈禹伸
Original Assignee
垒途智能教科技术研究院江苏有限公司
英智医疗科技南京有限公司
贝利尔科技发展南京有限公司
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 垒途智能教科技术研究院江苏有限公司, 英智医疗科技南京有限公司, 贝利尔科技发展南京有限公司 filed Critical 垒途智能教科技术研究院江苏有限公司
Publication of WO2022082840A1 publication Critical patent/WO2022082840A1/zh

Links

Images

Classifications

    • AHUMAN NECESSITIES
    • A61MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61BDIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00Measuring for diagnostic purposes; Identification of persons
    • A61B5/16Devices for psychotechnics; Testing reaction times ; Devices for evaluating the psychological state
    • A61B5/165Evaluating the state of mind, e.g. depression, anxiety
    • AHUMAN NECESSITIES
    • A61MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61BDIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00Measuring for diagnostic purposes; Identification of persons
    • A61B5/72Signal processing specially adapted for physiological signals or for diagnostic purposes
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks

Definitions

  • the invention relates to an evaluation system and method for game competition ability, in particular to a maze level evaluation system and an evaluation method, belonging to the field of intelligent learning evaluation.
  • Picomouse As a walking machine for competition, Picomouse especially tests the designer's chip, communication and algorithm design ability in the maze walking competition.
  • the prior art improves the anti-jamming (CN102841618A), dynamic level (CN103605363A), turning motion (CN103472831A), sprint time (CN103529846A), and motion accuracy (CN206039293U) of the pico-mouse by improving the function of the components of the pico-mouse.
  • maze walking has no international competition for manual walking, but prefers games and intelligence tests, and the level evaluation depends solely on the length of time.
  • Using brain wave control equipment to play games to control game screen objects (CN201949631U), or identify emotions in games (CN110866537A) has become a non-contact control technology for game control.
  • the existing technology has not paid attention to the development prospects of two aspects.
  • second, the existing On the one hand, the technology has a single evaluation method for excellent players, and on the other hand, it is impossible to obtain a model of the skills of the excellent players through a reasonable model, so as to optimize the artificial intelligence ability of Weiwei Mouse.
  • the patent CN107468260A uses a single classification model to identify the animal's mental state, which is obviously not suitable for the state of the human maze competition.
  • the competition is a comprehensive intelligence and psychological comprehensive ability test.
  • the competition requires not only time, but also whether the walking route during the competition is smooth, the anxiety and excitement during the competition, and the ability to overcome anxiety and reflect the comprehensive quality of choice. Therefore, it is necessary to detect these factors affecting the game in real time.
  • the single-pass machine evaluation is often not objective enough for the level evaluation due to the inaccuracy of the algorithm. Therefore, there is a need for a device that can detect the walking state and psychological state of the players in real time, so that a machine evaluation method can be obtained through a reasonable algorithm.
  • human experience evaluation is also an indispensable link, and human evaluation can give expert-level judgment based on the experience of the game.
  • the prior art does not comprehensively consider the combination of machine evaluation and manual evaluation.
  • a maze level evaluation system includes: a maze, a brain wave testing device, a brain wave analysis module, a level evaluation module, a camera for photographing the maze, and a display for on-site display and recording of the actual maze walking with the recording system, as well as the control module.
  • the control module is electrically connected with the brain wave testing device, the brain wave analysis module, the level evaluation module, the camera, and the display and recording system, respectively, and is used to control these devices and modules. Evaluation score.
  • the brainwave testing device continuously detects the tester's brainwave signal, de-noises the signal and extracts the characteristic signal value through the brainwave analysis module, so as to determine the residence time at a certain point in the walking route according to the competition.
  • the end is walking smoothly, it is stipulated that staying at a certain position for more than 3 seconds during the maze walking process is considered to be unsmooth, and within a certain period of time (such as 5-15 seconds) multiple (more than 2 times) walking occurs
  • the average value of the change value of the brain wave characteristic signal value is reduced as the anxiety characteristic signal to determine whether it is in an anxiety state, and after more than two anxiety characteristic signals appear, it is detected whether the normal characteristic signal value is restored to determine whether it has overcome. anxiety.
  • the average value of the elevated change value higher than the normal characteristic signal value is regarded as the excitation characteristic signal.
  • the average value of the decrease (or increase) change value of the brainwave characteristic signal value refers to the arithmetic mean of the sum of the difference (defined as a positive value) between the characteristic signal value of anxiety (or excitement) and the normal characteristic signal value for n times.
  • the scores of smooth walking p1 (positive value), unsmooth walking p2 (negative value), anxiety p3 (negative value), overcoming anxiety p4 (positive value), game excitement p5 (positive value), and game time score P6 is used as the six evaluation indicators. Five brain wave characteristic signals are sampled every 2-5 seconds, and the obtained five types of brain wave characteristic signal values, and the game time score is obtained at the end of the game.
  • all the sampling values and the score of the game time are processed by the entropy weight method through the level evaluation module to obtain the score F' , to obtain machine evaluation results.
  • the entropy weight method includes index normalization processing, entropy value calculation, information entropy redundancy, index weight, and calculation score.
  • the display and recording system will display the live video captured by the camera, respectively, and multiple on-site game experts will watch the game through the display screen in the display and recording system, thereby obtaining multiple manual evaluation scores.
  • Experts input their final evaluation scores F1, F2...Fn (n is a non-zero natural number) through the input device of the control module, and obtain the comprehensive evaluation score through calculation by the level evaluation module.
  • the present invention provides an evaluation method for a brain training evaluation system based on an intelligent maze, which is characterized by comprising the following steps:
  • the S1 control module controls the opening of the brain wave test device, the brain wave analysis module, the level evaluation module, the camera, and the display and recording system, so that it is in a working state;
  • S2 connects the tester's head with the sensor patch in the radio wave testing device, and conducts the test, sampling five brainwave characteristic signals every 2-5 seconds; at this time, the level evaluation module starts timing, and the brainwave analysis module accepts ⁇ Noise reduction and processing of the brain wave characteristic signal, and transmit the processed brain wave characteristic signal to the level evaluation module; at the same time, the camera starts to record the maze video, and the referee watches the game through the display screen in the display and recording system;
  • the level evaluation module processes the entropy weight method to obtain the score F', and obtains the machine evaluation result, wherein the score of each index is the arithmetic mean of all the sampling scores, and F' is the arithmetic mean of the scores of the six indices
  • the fastest among them is evaluated manually as the best.
  • the present invention also provides a method capable of optimizing the ability of the pico-rat maze, which is characterized in that the deep learning of various mazes is carried out through the six index values of the top 100 outstanding people in the world, and the location of the fork in the walking Choose the path with the highest score, so as to continuously accumulate the skills of maze walking, and optimize the maze walking ability of the pico mouse.
  • the deep learning includes at least one of convolutional neural network, self-encoding, sparse encoding, and deep belief network.
  • Fig. 1 is a schematic diagram of the composition of the brain training evaluation system based on the intelligent maze of the present invention.
  • 1. the competition platform 2. the maze, the brain wave test device 3, the brain wave analysis module 4, the level evaluation module 5, the camera 6 for shooting the maze, and the display and recording of the live display and recording of the maze walking System 7, and Control Module 8.
  • the brain training evaluation system and its evaluation method based on the intelligent maze will be described below with reference to FIG. 1 .
  • Figure 1 is a brain training evaluation system based on an intelligent maze, which is set on a competition platform 1, including a maze 2, a brain wave testing device 3, a brain wave analysis module 4, a level evaluation module 5, and a camera 6 for shooting the maze,
  • the display and recording system 7 and the control module 8 are used to display and record the actual situation of the maze walking on the spot.
  • the control module 8 is electrically connected with the brain wave testing device 3, the brain wave analysis module 4, the level evaluation module 5, the camera 6, and the display and recording system 7, respectively, for controlling these devices and modules, and the control module 8 also has an input A device (not shown) for entering scores including expert evaluations.
  • This embodiment is based on the evaluation method of the intelligent maze-based brain training evaluation system of Embodiment 1, and is characterized in that it includes the following steps:
  • the S1 control module 8 controls and opens the brain wave testing device 3, the brain wave analysis module 4, the level evaluation module 5, the camera 6, and the display and recording system 7, so as to be in a working state;
  • S2 connects the tester's head with the sensor patch in the radio wave testing device 3, and conducts the test, sampling five brain wave characteristic signals every 2-5 seconds; at this time, the level evaluation module 5 starts timing, and the brain wave analysis The module 4 accepts and denoises the brainwave characteristic signal, and transmits the processed brainwave characteristic signal to the level evaluation module 5; at the same time, the camera 6 starts to record the video of the maze 2, and the referee passes the display in the display and recording system 7. watch the game on the screen;
  • the score of the game time is obtained, and all the sampling values and the score of the game time are processed by the entropy weight method in the level evaluation module 5 to obtain the score F', and the machine evaluation result is obtained, wherein the score of each index is all the sampling scores.
  • F' is the arithmetic mean of the scores of the six indicators; the referee inputs the manual evaluation scores F1, F2...Fn through the input device of the display and recording system 7, where n is a non-zero natural number, and
  • the calculation method for the entropy weight method is as follows:
  • x ij represents the value of the jth evaluation index of the ith sampling sample
  • This embodiment provides a method capable of optimizing the ability of the pico-mouse maze, which is characterized in that the deep learning of a variety of different mazes is carried out through the six index values of the top 100 outstanding people in the world, and the location of the fork in the walking Choose the path with the highest score, so as to continuously accumulate the skills of maze walking, and optimize the maze walking ability of the pico mouse.
  • the deep learning includes at least one of convolutional neural network, self-encoding, sparse encoding, and deep belief network.

Landscapes

  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Molecular Biology (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • General Health & Medical Sciences (AREA)
  • Psychiatry (AREA)
  • Public Health (AREA)
  • Veterinary Medicine (AREA)
  • Artificial Intelligence (AREA)
  • Animal Behavior & Ethology (AREA)
  • Surgery (AREA)
  • Medical Informatics (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Pathology (AREA)
  • Theoretical Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Data Mining & Analysis (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Signal Processing (AREA)
  • Software Systems (AREA)
  • Mathematical Physics (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Computing Systems (AREA)
  • Physiology (AREA)
  • Computational Linguistics (AREA)
  • Child & Adolescent Psychology (AREA)
  • Developmental Disabilities (AREA)
  • Educational Technology (AREA)
  • Hospice & Palliative Care (AREA)
  • Psychology (AREA)
  • Social Psychology (AREA)
  • Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
  • Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)

Abstract

一种迷宫(2)水平评估系统,包括迷宫(2),脑电波测试装置(3),脑电波分析模块(4),水平评价模块(5),控制模块(8),用于拍摄迷宫(2)的摄像仪(6),以及用于现场显示和记录迷宫(2)行走实况的显示与记录系统(7)。通过脑电波分析模块(4)分析出比赛者在比赛时的行走顺畅以及不顺畅,出现比赛焦虑,比赛兴奋,以及焦虑克服后相应五类脑电波特征值,通过熵权法计算得到得分,实现对比赛者迷宫(2)水平的机器评价。而摄像仪(6)拍摄到的迷宫(2)行走实况供测评人员根据经验判断,结合机器评价得到最终的综合评价。还包括一种基于评估系统的迷宫(2)水平评价方法以及利用优秀测试者的脑电波特征通过深度学习提高微微鼠迷宫能力。利用熵权法重组机器评价,获得最优机器评价,结合人工评价,能够客观准确科学地分析比赛者迷宫(2)水平。

Description

基于智能迷宫的大脑训练评价系统及其评价方法 技术领域
本发明涉及一种游戏比赛能力的评价系统及其方法,尤其涉及到迷宫水平评价系统及其评价方法,属于智能学习评价领域。
背景技术
微微鼠作为竞赛用行走机器,在迷宫行走比赛中特别考验设计者的芯片以及通讯和算法设计的能力。现有技术通过改良微微鼠的部件功能,来改善其在迷宫型走时的抗干扰(CN102841618A),动态水平(CN103605363A),转弯运动(CN103472831A)、冲刺时间(CN103529846A)、运动精确性(CN206039293U)。迷宫行走传统上没有人工行走的国际性比赛,而偏向于游戏以及智力测试方向,并且水平评价单一依靠时间长短。利用脑电波控制设备进行游戏,来控制游戏画面对象(CN201949631U),或识别游戏中的情感(CN110866537A),成为游戏控制的一个非接触控制技术。
综上,现有技术并未关注到两个方面的开发前景,第一,没有注意到实现微微鼠的迷宫能力通过学习,尤其是通过优秀人工玩家的行走技巧学习以提高;第二,现有技术对于一方面优秀选手的评价方法单一,另一方面也不能通过合理的模型得到优秀者技巧的模型,从而优化微微鼠行走的人工智能能力。专利CN107468260A使用单一的分类模型识别动物心里状态,显然不适合人的迷宫比赛时的状态,CN107538492A只提供了利用脑电波的方法控制了机器人的运动,并不涉及比赛水平的评价。
而比赛是综合性智力和心理的综合能力测试,比赛不仅需要时间、还需要通过比赛时行走路线是否顺畅、比赛时的焦虑、兴奋、以及克服焦虑能力,反应选择的综合素质。因此需要实时检测这些影响比赛的因素。而单一通过机器评价往往由于算法不精确也容易对水平评估不够客观。因此 需要一种能够实时检测比赛选手行走状态以及心理状态的设备,从而能够通过合理的算法得到机器评价方法。
此外,人的经验评价也是不可缺少的一个环节,人的评价能够根据比赛的经验而给予专家级别的判断。现有技术并未综合考虑机器评价与人工评价的综合。
发明内容
基于上述现有技术的不足,本发明提供一种能够综合合理的机器算法和基于智能迷宫的大脑训练评价系统。解决了现有技术使用单一的时间长短评价,以及忽略了及其评价与人工评价的综合考量的技术问题。根据本发明的一种迷宫水平评价系统,其包括:迷宫,脑电波测试装置,脑电波分析模块,水平评价模块,用于拍摄迷宫的摄像仪,以及用于现场显示和记录迷宫行走实况的显示与记录系统,以及控制模块。所述控制模块分别与脑电波测试装置,脑电波分析模块,水平评价模块,摄像仪,显示与记录系统电连接,用于控制这些装置与模块,控制模块还具备输入设备,用于输入包括专家评价的得分。
当测试者进行比赛时,脑电波测试装置不断检测测试者的脑电波信号,通过脑电波分析模块对信号进行去噪以及特征信号值的提取,从而根据比赛是行走路线中在某一点的滞留时间来看端是否属于行走顺畅,规定在迷宫行走过程中在某一位置点滞留超过3秒属于行走不顺畅,而在一定时间内(比如5-15秒)内出现多次(2次以上)行走不顺畅是则根据脑电波特征信号值降低变化值平均值作为焦虑特征信号,判断是否处于焦虑状态,并且在出现2次以上焦虑特征信号后检测是否恢正常特征信号值,来判定是否已经克服了焦虑。而将比正常特征信号值高的升高变化值平均值作为兴奋特征信号
其中,所述脑电波特征信号值降低(或升高)变化值平均值是指在n次焦虑(或兴奋)特征信号值与正常特征信号值得差值(以正值定义)之和的算术平均值,差值越大则相应会增大所获得分的绝对值。从而将行走 顺畅p1(为正值)、行走不顺畅p2(为负值)、焦虑p3(为负值)、克服焦虑p4(为正值)、比赛兴奋p5(为正值)、比赛时间得分p6的作为六个评价指标,每隔2-5秒对五个脑电波特征信号采样一次,获取到的五类脑电波特征信号值,以及比赛结束时获得比赛时间得分。最后将所有采样值以及比赛时间得分(每一次获得的采样,其对应有相同的比赛时间得分指标,也即比赛时间得分是不变的),通过水平评价模块进行熵权法处理得到得分F',获得机器评价结果。其中所述熵权法包括指标归一化处理、熵值计算、信息熵冗余度、指标权重、以及计算得分。并将世界最快纪录时间定义为比赛时间得分1,根据比赛时间超出世界最快纪录时间的百分比来w相应计算比赛时间得分f=1/(1+w)。当至少一名选手用时平或超过世界最快纪录时间,则即人工评价其中最快者为最优秀,而无需参与机器评价。
当比赛结束,显示与记录系统将摄像仪拍摄的实况的分别有多位现场观看比赛专家通过显示与记录系统中的显示屏观看,从而得到多个人工评价得分。当得分有异议,还能通过比赛录像回看重新评价。专家通过控制模块的输入设备将各自的最终评价得分F1,F2...Fn(n为不为零的自然数)输入,经由水平评价模块计算得到综合评价得分。所述综合评价得分F=aF'+bF1+cF2+...+xFn,其中a,b,c,...,x为正实数权重,a不小于0.5,且a+b+c+...+x=1;a,b,c,d...x根据裁判的等级水平给到相应的权重,等级越高权重越大。
本发明提供一种基于智能迷宫的大脑训练评价系统的评价方法,其特征在于,包括如下步骤:
S1控制模块控制打开脑电波测试装置,脑电波分析模块,水平评价模块,摄像仪,显示与记录系统,使处于工作状态;
S2将测试者头部与电波测试装置中传感贴片连接,并进行测试,每隔2-5秒对五个脑电波特征信号采样一次;此时水平评价模块开始计时,脑电波分析模块接受、降噪处理脑电波特征信号,将处理后的脑电波特征信号传递给水平评价模块;与此同时,摄像仪开始拍摄迷宫录像,裁判通过显 示与记录系统中的显示屏观看比赛;
S3.将测得的行走顺畅正值p1、行走不顺畅负值p2为、焦虑负值p3、克服焦虑正值p4、比赛兴奋正值p5、比赛时间得分正值p6的作为六个评价指标,每隔3秒对五个脑电波特征信号采样一次,获取到的五类脑电波特征信号值;根据比赛时间超出世界最快纪录时间的百分比来w相应计算比赛时间得分f=1/(1+w);
S4.比赛结束时,水平评价模块进行熵权法处理得到得分F',获得机器评价结果,其中每个指标的得分为所有采样得分的算术平均值,F'为六个指标的得分的算术平均值;裁判通过显示与记录系统的输入设备输入人工评价得分F1,F2...Fn,其中n为不为零的自然数,并由水平评价模块计算综合评价得分F=aF'+bF1+cF2+...+xFn,其中a,b,c,...,x为正实数权重,0.5≤a<0.65,且a+b+c+...+x=1;a,b,c,d...x根据裁判的等级水平给到相应的权重,等级越高权重越大,规定只设置一名特级裁判给0.2权重,一名一级裁判给0.15权重,其他裁判中的每一位给剩余的权重的按照人头的平均数z;设一共y名其他裁判,则z=(0.65-a)/y;
其中,根据比赛时间超出世界最快纪录时间的百分比来w相应计算比赛时间得分f=1/(1+w)。当至少一名选手用时平或超过世界最快纪录时间,则即人工评价其中最快者为最优秀。
本发明还提供了一种能够优化微微鼠迷宫能力的方法,其特征在于,通过世界排名前一百优秀者的六个指标值来进行多种不同迷宫的深度学习,而在行走中岔路位置点选择最高得分的路径进行,从而不断积累迷宫行走的技巧,优化微微鼠的迷宫行走能力。其中所述深度学习包括卷积神经网络、自编码、稀疏编码,以及深度置信网络中的至少一种。
附图说明
图1.本发明基于智能迷宫的大脑训练评价系统组成示意图。
其中,1.赛台、2.迷宫、脑电波测试装置3,脑电波分析模块4,水平评价模块5,用于拍摄迷宫的摄像仪6,用于现场显示和记录迷宫行走实况的显 示与记录系统7,以及控制模块8。
具体实施方式
以下结合附图1对基于智能迷宫的大脑训练评价系统及其评价方法进行说明。
实施例1
如图1是基于智能迷宫的大脑训练评价系统,设置于赛台1上,包括迷宫2、脑电波测试装置3,脑电波分析模块4,水平评价模块5,用于拍摄迷宫的摄像仪6,用于现场显示和记录迷宫行走实况的显示与记录系统7,以及控制模块8。所述控制模块8分别与脑电波测试装置3,脑电波分析模块4,水平评价模块5,摄像仪6,显示与记录系统7电连接,用于控制这些装置与模块,控制模块8还具备输入设备(图未示),用于输入包括专家评价的得分。
实施例2
本实施例基于实施例1的基于智能迷宫的大脑训练评价系统的评价方法,其特征在于,包括如下步骤:
S1控制模块8控制打开脑电波测试装置3,脑电波分析模块4,水平评价模块5,摄像仪6,显示与记录系统7,使处于工作状态;
S2将测试者头部与电波测试装置3中传感贴片连接,并进行测试,每隔2-5秒对五个脑电波特征信号采样一次;此时水平评价模块5开始计时,脑电波分析模块4接受、降噪处理脑电波特征信号,将处理后的脑电波特征信号传递给水平评价模块5;与此同时,摄像仪6开始拍摄迷宫2录像,裁判通过显示与记录系统7中的显示屏观看比赛;
S3.将测得的行走顺畅正值p1、行走不顺畅负值p2为、焦虑负值p3、克服焦虑正值p4、比赛兴奋正值p5、比赛时间得分正值p6的作为六个评价指标,每隔3秒对五个脑电波特征信号采样一次,获取到的五类脑电波特征信号值;根据比赛时间超出世界最快纪录时间的百分比来w相应计算 比赛时间得分f=1/(1+w);
S4.比赛结束时,获得比赛时间得分,将所有采样值以及比赛时间得分在水平评价模块5中进行熵权法处理得到得分F',获得机器评价结果,其中每个指标的得分为所有采样得分的算术平均值,F'为六个指标的得分的算术平均值;裁判通过显示与记录系统7的输入设备输入人工评价得分F1,F2...Fn,其中n为不为零的自然数,并由水平评价模块5计算综合评价得分F=aF'+bF1+cF2+dF3,其中a=0.6;设置一名特级裁判相应的b=0.2,一名一级裁判相应的c=0.15,其他裁判两名,每一位给剩余的权重为0.025;其中对于每一次获得的采样,其对应有相同的比赛时间得分指标,也即比赛时间得分是不变的。
对于熵权法计算方法如下:
(1)假设有m个待评价采样样本,n项评价指标,形成原始指标数据矩阵:
Figure PCTCN2020124836-appb-000001
,其中x ij表示第i个采样样本第j项评价指标的数值;
(2)对各指标进行标准化处理如下:
p1、p4、p5、p6采用正向指标:
Figure PCTCN2020124836-appb-000002
p2、p3采用负向指标:
Figure PCTCN2020124836-appb-000003
(3)计算权重
计算第j个指标中,第i个采样值(已标准化)的权重
Figure PCTCN2020124836-appb-000004
(4)计算第j个指标熵:
Figure PCTCN2020124836-appb-000005
其中
Figure PCTCN2020124836-appb-000006
因为0≤ej≤1,于是
Figure PCTCN2020124836-appb-000007
(5)权重计算:
Figure PCTCN2020124836-appb-000008
其中d j=1-e j
(6)计算得到第i个
Figure PCTCN2020124836-appb-000009
采样样本的得分值:
实施例3
本实施例提供了一种能够优化微微鼠迷宫能力的方法,其特征在于,通过世界排名前一百优秀者的六个指标值来进行多种不同迷宫的深度学习,而在行走中岔路位置点选择最高得分的路径进行,从而不断积累迷宫行走的技巧,优化微微鼠的迷宫行走能力。其中所述深度学习包括卷积神经网络、自编码、稀疏编码,以及深度置信网络中的至少一种。

Claims (6)

  1. 基于智能迷宫的大脑训练评价系统,其特征在于,所述迷宫水平评价系统设置于赛台上,迷宫水平评价系统包括:迷宫,脑电波测试装置,脑电波分析模块,水平评价模块,用于拍摄迷宫的摄像仪,用于现场显示和记录迷宫行走实况的显示与记录系统,以及控制模块;所述控制模块分别与脑电波测试装置,脑电波分析模块,水平评价模块,摄像仪,显示与记录系统电连接,控制模块还具有输入设备,用于输入包括专家评价的得分。
  2. 根据权利要求1的基于智能迷宫的大脑训练评价系统的评价方法,其特征在于,包括如下步骤:
    S1控制模块控制打开脑电波测试装置,脑电波分析模块,水平评价模块,摄像仪,显示与记录系统,使处于工作状态;
    S2将测试者头部与电波测试装置中传感贴片连接,并进行测试,每隔2-5秒对五个脑电波特征信号采样一次;此时水平评价模块开始计时,脑电波分析模块接受、降噪处理脑电波特征信号,将处理后的电波特征信号传递给水平评价模块;与此同时,摄像仪开始拍摄迷宫录像,裁判通过显示与记录系统中的显示屏观看比赛;
    S3将测得的行走顺畅正值p1、行走不顺畅负值p2为、焦虑负值p3、克服焦虑正值p4、比赛兴奋正值p5五个电波特征信号、比赛时间得分正值p6的作为六个评价指标,每隔2-5秒对五个脑电波特征信号采样一次,获取到的五类脑电波特征信号值;
    S4.比赛结束时,获得比赛时间得分,将所有采样值以及比赛时间得分在水平评价模块中进行熵权法处理得到得分F',获得机器评价结果,其中每个指标的得分为所有采样得分的算术平均值,F'为六个指标的得分的算术平均值;裁判通过显示与记录系统的输入设备输入人工评价得分F1,F2...Fn,其中n为不为零的自然数,并由水平评价模块计算综合评价得分F=aF'+bF1+cF2+...+xFn,其中a,b,c,...,x为正实数权重,0.5≤a<1,且a+b+c+...+x=1;a,b,c,d...x根据裁判的等级水平给到相应的权重,等级越高权重越大;
    对于熵权法计算方法如下:
    (1)假设有m个待评价采样样本,n项评价指标,形成原始指标数据矩阵:,其中x ij表示第i个
    Figure PCTCN2020124836-appb-100001
    采样样本第j项评价指标的数值;
    (2)对各指标进行标准化处理如下:
    p1、p4、p5、p6采用正向指标:
    Figure PCTCN2020124836-appb-100002
    p2、p3采用负向指标:
    Figure PCTCN2020124836-appb-100003
    (3)计算权重
    计算第j个指标中,第i个采样值(已标准化)的权重
    Figure PCTCN2020124836-appb-100004
    (4)计算第j个指标熵:
    Figure PCTCN2020124836-appb-100005
    其中
    Figure PCTCN2020124836-appb-100006
    因为0≤ej≤1,于是
    Figure PCTCN2020124836-appb-100007
    (5)权重计算:
    Figure PCTCN2020124836-appb-100008
    其中d j=1-e j
    (6)计算得到第i个
    Figure PCTCN2020124836-appb-100009
    采样样本的得分值:
  3. 根据权利要求2的评价方法,其特征在于,设置一名特级裁判给0.2权重,一名一级裁判给0.15权重,其他裁判中的每一位给剩余的权重的按照人头的平均数z;设一共y名其他裁判,则z=(0.65-a)/y。
  4. 根据权利要求2或3的评价方法,其特征在于,步骤S3中根据比赛时间超出世界最快纪录时间的百分比来w相应计算比赛时间得分f=1/(1+w);当至少一名选手用时平或超过世界最快纪录时间,则即人工评价其中最快者为最优秀。
  5. 一种能够优化微微鼠迷宫能力的方法,其特征在于,通过世界排名前一百优秀者的根据权利要求2所得到的六个评价指标值来进行多种不同迷宫的深度学习,而在行走中岔路位置点选择最高得分的路径进行,从而不断积累迷宫行走的技巧,优化微微鼠的迷宫行走能力。
  6. 根据权利要求4的方法,其特征在于,所述深度学习包括卷积神经网络、自编码、稀疏编码,以及深度置信网络中的至少一种。
PCT/CN2020/124836 2020-10-22 2020-10-29 基于智能迷宫的大脑训练评价系统及其评价方法 WO2022082840A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202011136200.1 2020-10-22
CN202011136200.1A CN112315471A (zh) 2020-10-22 2020-10-22 基于智能迷宫的大脑训练评价系统及其评价方法

Publications (1)

Publication Number Publication Date
WO2022082840A1 true WO2022082840A1 (zh) 2022-04-28

Family

ID=74312131

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/124836 WO2022082840A1 (zh) 2020-10-22 2020-10-29 基于智能迷宫的大脑训练评价系统及其评价方法

Country Status (2)

Country Link
CN (1) CN112315471A (zh)
WO (1) WO2022082840A1 (zh)

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030024482A1 (en) * 2001-08-06 2003-02-06 Vijay Gondhalekar Programmable electronic maze for use in the assessment of animal behavior
CN201995441U (zh) * 2010-10-14 2011-10-05 复旦大学附属华山医院 痴呆大鼠迷宫型智力康复训练装置
CN103764021A (zh) * 2011-05-20 2014-04-30 南洋理工大学 一种用于协同神经-生理学修复和/或功能提升的系统、仪器、装置和方法
CN105230520A (zh) * 2015-07-25 2016-01-13 南昌航空大学 一种全自动鼠迷宫行为训练系统
US20170340230A1 (en) * 2014-11-17 2017-11-30 National Institute Of Advanced Industrial Science And Technology Similarity Evaluation Method Using Brain Waves, Evaluation Device, Evaluation System, And Program
CN107422841A (zh) * 2017-03-03 2017-12-01 杭州市第人民医院 一种基于非接触式情绪识别的人机交互方法
CN208001848U (zh) * 2017-12-21 2018-10-26 中国科学院深圳先进技术研究院 一种迷宫试验装置
CN111420401A (zh) * 2020-04-13 2020-07-17 宋宇 一种实景迷宫系统及其控制方法

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030024482A1 (en) * 2001-08-06 2003-02-06 Vijay Gondhalekar Programmable electronic maze for use in the assessment of animal behavior
CN201995441U (zh) * 2010-10-14 2011-10-05 复旦大学附属华山医院 痴呆大鼠迷宫型智力康复训练装置
CN103764021A (zh) * 2011-05-20 2014-04-30 南洋理工大学 一种用于协同神经-生理学修复和/或功能提升的系统、仪器、装置和方法
US20170340230A1 (en) * 2014-11-17 2017-11-30 National Institute Of Advanced Industrial Science And Technology Similarity Evaluation Method Using Brain Waves, Evaluation Device, Evaluation System, And Program
CN105230520A (zh) * 2015-07-25 2016-01-13 南昌航空大学 一种全自动鼠迷宫行为训练系统
CN107422841A (zh) * 2017-03-03 2017-12-01 杭州市第人民医院 一种基于非接触式情绪识别的人机交互方法
CN208001848U (zh) * 2017-12-21 2018-10-26 中国科学院深圳先进技术研究院 一种迷宫试验装置
CN111420401A (zh) * 2020-04-13 2020-07-17 宋宇 一种实景迷宫系统及其控制方法

Also Published As

Publication number Publication date
CN112315471A (zh) 2021-02-05

Similar Documents

Publication Publication Date Title
McRobert et al. Tracing the process of expertise in a simulated anticipation task
Gilis et al. Offside decisions by expert assistant referees in association football: Perception and recall of spatial positions in complex dynamic events.
US20070166680A1 (en) Sports skill evaluation system
CN110378232B (zh) 改进的ssd双网络的考场考生位置快速检测方法
Bloß et al. Physical load and referees’ decision-making in sports games: A scoping review
CN105069304A (zh) 一种基于机器学习的评估预测asd的方法
KR101768729B1 (ko) 카테고리에 의한 지각 테스트 데이터 분석
CN113288452B (zh) 手术质量检测方法及装置
CN107610099A (zh) 一种基于瞳孔信息的情感精神状态检测系统及方法
CN111428655A (zh) 一种基于深度学习的头皮检测方法
Hu et al. Basketball activity classification based on upper body kinematics and dynamic time warping
WO2023098516A1 (zh) 拳击运动的电子裁判系统、电子裁判方法和存储介质
Scheer et al. Judging bias induced by viewing contrived videotapes: A function of selected psychological variables
Huffman et al. Situation Awareness in sports: A scoping review
WO2022082840A1 (zh) 基于智能迷宫的大脑训练评价系统及其评价方法
CN112494034B (zh) 基于3d体态检测分析的数据处理分析系统及方法
CN111048202A (zh) 一种智能化中医诊断系统及其方法
Xia et al. VREN: volleyball rally dataset with expression notation language
CN115620050A (zh) 一种基于气候室环境下改进的YOLOv5蚜虫识别计数方法
CN115575127A (zh) 一种基于鲸鱼优化lssvm的轴承剩余寿命预测方法
CN115966003A (zh) 一种基于情绪识别的学习者线上学习效率评估系统
CN115273176A (zh) 一种基于生命体征和表情的疼痛多算法客观评估方法
Chua et al. Activity identification and classification in wheelchair rugby using fractal dimensions
CN111210404B (zh) 晶状体分割难度分类方法及装置
US20210245005A1 (en) Implementation of machine learning for skill-improvement through cloud computing and method therefor

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: 20958420

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20958420

Country of ref document: EP

Kind code of ref document: A1