Walking posture recognition system and recognition equipment

Abstract

The invention relates to a walking posture recognition system and recognition equipment. The walking posture recognition system comprises a data acquisition module, a gait event detection module, a fractal dimension calculation module, a score calculation module and an intelligent feedback module. The data acquisition module is used for respectively collecting front-back, left-right and vertical linear accelerations and angular velocities of left and right feet of a user, and outputting a data stream; the gait event detection module receives the data stream, determines timestamps of a plurality of gait events through the data stream, and outputs a data structure based on the determined timestamps; the fractal dimension calculation module is used for receiving and dividing the user gait into four dimensions based on the data structure content, and calculating the score of each dimension; the score calculation module sets a weight coefficient for each of the four dimensions based on the fractal dimension calculation module. The score calculation module calculates and outputs a total score based on respective scores of the four dimensions and the weight coefficients. The invention aims to solve the problem of multi-dimensional, quantifiable and real-time gait posture scoring through portable equipment.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of wearable devices and motion analysis, in particular to a walking posture recognition system and a recognition device. BACKGROUND

[0002] The health status of a user can be clearly judged from the state of gait, but traditional gait analysis relies on an optical motion capture system and a force platform, and multiple infrared cameras and sensors need to be laid out in a laboratory environment, a single set of equipment costs more than one million yuan and cannot be used outdoors, resulting in a serious limitation of the evaluation scene.

[0003] Although consumer-grade wearable devices are portable, they can only detect basic parameters such as step count and step frequency because they are worn on the wrist far from the foot movement chain, and cannot obtain foot-ground interaction dynamics characteristics, resulting in the absence of evaluation of core quality indicators such as gait symmetry and efficiency. In the field of rehabilitation medicine, clinical gait abnormalities (such as hemiplegic gait) are evaluated by therapists through visual assessment, which has strong subjectivity and lacks quantitative standards.

[0004] In summary, the existing technology has the core defects of incompatibility of portability, multi-dimensional quantification, and real-time performance. SUMMARY

[0005] The main purpose of the present application is to provide a walking posture recognition system, which aims to solve the problem of multi-dimensional, quantifiable, and real-time gait posture scoring through a portable device.

[0006] To achieve the above-mentioned purpose, the walking posture recognition system proposed by the present application comprises a data acquisition module, a gait event detection module, a multi-dimensional calculation module, a scoring calculation module, and an intelligent feedback module.

[0007] The data acquisition module is used to collect the linear acceleration and angular velocity of the front and back, left and right, and vertical directions of the left and right feet of the user respectively, and output data streams;

[0008] The gait event detection module is configured with a processing unit based on the peak detection method, the processing unit receives the data stream and determines the timestamps of multiple gait events such as the initial landing time, toe-off time, support phase, swing phase, and gait cycle of the left foot and / or right foot of the user through the data stream, and outputs a data structure based on the determined timestamps;

[0009] The multi-dimensional calculation module is used to receive and divide the gait of the user into four dimensions of gait symmetry, dynamic stability, gait efficiency and fluency, and body posture and foot contact based on the content of the data structure, and calculate the scores of each dimension respectively;

[0010] The score calculation module is provided with one weight coefficient for each of the four dimensions based on the dimension calculation module, and the sum of the four weight coefficients is constrained to be 1; the score calculation module calculates and outputs the total score based on the respective scores of the four dimensions and the weight coefficients;

[0011] The intelligent feedback module receives the total score and feeds back the improvement suggestions to the user in real time according to the total score.

[0012] In an embodiment of the present application, the data acquisition module further comprises:

[0013] A microprocessor is configured to receive the data stream and perform preliminary filtering and unit conversion on the data stream;

[0014] A data buffer unit is connected to the microprocessor and configured to temporarily store the data stream.

[0015] In an embodiment of the present application, the data buffer unit is provided with a remote connection module for remotely transmitting the data stream to an external device, thereby improving the subsequent data processing accuracy.

[0016] In an embodiment of the present application, the gait symmetry dimension is based on comparing the gait states of the left and right feet of the user, and the dimension is divided into four indexes, and the dimension score is calculated through the four indexes:

[0017] Swing_Time_Diff=(abs(T_left_swing-T_right_swing) / max(T_left_swing,T_right_swing))*100%;

[0018] Stance_Time_Diff=(abs(T_left_stance-T_right_stance) / max(T_left_stance,T_right_stance))*100%;

[0019] Step_Length_Diff=(abs(StepLen_left-StepLen_right) / max(StepLen_left,StepLen_right))*100%;

[0020] Step_Length_Diff=(abs(StepLen_left-StepLen_right) / max(StepLen_left,StepLen_right))*100%;

[0021] The initial ground angle difference between the left and right feet of the user is calculated by the linear acceleration and angular velocity at the initial landing time in the data stream:

[0022] Strike_Angle_Diff = abs(StrikeAngle_left - StrikeAngle_right) ;

[0023] Symmetry Dimension Score Calculation: Sym_Score = 100-

[0024] (W1 * Swing_Time_Diff + W2 * Stance_Time_Diff + W3 * Step_Length_Diff + W4 * Stike_Angle_Diff) ;

[0025] T_left_swing is the left swing phase, T_right_swing is the right swing phase, T_left_stance is the left stance phase, T_right_stance is the right stance phase, StepLen_left is the left step length, StepLen_right is the right step length, StrikeAngle_left is the initial ground angle difference of the left foot, StrikeAngle_right is the initial ground angle difference of the right foot.

[0026] W1, W2, W3, W4 are the weight coefficients of the four indicators in the evaluation of the dimension score, the total is 1, and the symmetry dimension score range is [0, 100], the greater the difference, the lower the score.

[0027] In an embodiment of the present application, the score calculation formula of the dynamic stability dimension is:

[0028] Stab_Score = 100 - (W5 * Norm(ML_Accel_RMS) + W6 * Norm(AP_Accel_RMS) + W7 * Norm(Gyro_Z_Max) + W8 * Norm(Stance_Time_Var)) ;

[0029] W5 + W6 + W7 + W8 = 1;

[0030] ML_Accel_RMS is the root mean square value of the linear acceleration in the left-right direction during the stance phase, which is used to reflect the amplitude of body left-right shaking;

[0031] AP_Accel_RMS is the root mean square value of the linear acceleration in the front-back direction during the stance phase, which is used to reflect the amplitude of body front-back shaking;

[0032] Gyro_Z_Max is the maximum absolute value of the angular velocity in the vertical direction during the stance phase, which is used to reflect the severity of the ankle inversion / eversion;

[0033] Stance_Time_Var is the coefficient of variation of the time of successive stance phases, calculated as (standard deviation / average value*100%) and used to reflect the stability of gait rhythm;

[0034] W5, W6, W7, W8 are the weight coefficients of the four indicators of the dynamic stability dimension, and the sum is 1. The symmetry dimension score ranges from [0, 100].

[0035] In an embodiment of the present application, the sub-dimension calculation module is provided with a data input module for inputting the height value of the user. The calculation formula of the gait efficiency and fluency dimension is:

[0036] Eff_Score = (W9*Norm_Cadence+W10*Norm(Stride_Regularity)+W11*Norm(Swing_Accel_Peak_Ratio))-(W12*Norm(Heel_Strike_Sharpness));

[0037] Norm_Cadence is the number of gait cycles per minute, which is normalized by the height value of the user;

[0038] Stride_Regularity is the autocorrelation coefficient of the stride or gait cycle calculated from the data stream;

[0039] Swing_Accel_Peak_Ratio is the peak value of acceleration in the swing phase / initial landing acceleration peak value, which is used to reflect the balance between propulsion and buffering;

[0040] Heel_Strike_Sharpness is the rising slope of vertical acceleration at initial landing, which is used to represent the stiffness of foot landing;

[0041] W9, W10, W11, W12 are the weight coefficients of the four indicators of the gait efficiency and fluency dimension, and the sum is 1. The symmetry dimension score ranges from [0, 100].

[0042] In an embodiment of the present application, the sub-dimension calculation module is further provided with a posture collection unit for collecting body posture data. The calculation formula of the body posture and foot contact dimension is:

[0043] Post_Score = 100-(W13*abs(Avg_Torso_Angle-Target_Angle)+W14*abs(Foot_Progression_Angle)+W15*Strike_Type_Penalty);

[0044] Avg_Torso_Angle is obtained by combining body posture data with gait cycle;

[0045] Target_Angle is the ideal torso angle;

[0046] Foot_Progression_Angle is the average angle between foot progression direction and straight line progression direction in the support phase;

[0047] Strike_Type_Penalty is the initial landing time based on three linear acceleration and angular velocity to determine the foot landing position;

[0048] W13, W14, W15 are weight coefficients in calculating scores in body posture and foot contact dimension.

[0049] In an embodiment of the present application, the calculation formula of the score calculation module is:

[0050] Gait_Score = (W_sym*Sym_Score + W_stab*Stab_Score + W_eff*Eff_Score + W_post*Post_Score);

[0051] W_sym, W_stab, W_eff, W_post are total score weight coefficients of the four dimensions respectively, and W_sym + W_stab + W_eff + W_post = 1.

[0052] In an embodiment of the present application, the score calculation module is further provided with a plurality of scene adjustment modules, and the total score weight coefficient values of the four dimensions of each scene adjustment module are different.

[0053] The present application further provides a walking posture recognition device comprising the walking posture recognition system, the walking posture recognition device comprising a shoe body, a three-axis accelerometer and a three-axis gyroscope arranged in the shoe body, and a microprocessor connected to the three-axis accelerometer and / or the three-axis gyroscope at the bottom of the shoe body;

[0054] The microprocessor is connected with a data buffer and is configured with a wireless transmission element for connecting with a wireless device.

[0055] By adopting the above technical scheme, the present application has the following advantages:

[0056] The application is a walking posture recognition system which can solve multi-dimension, quantifiable and real-time gait posture scoring. The system has simple structure and can be set in a portable device such as a user's shoe, which can effectively reduce the cost and improve the accuracy of recognition. In order to achieve this purpose, the system includes data acquisition module, gait event detection module, dimension calculation module, scoring calculation module and intelligent feedback module from structure and function. The module mainly set in the shoe sole is the data acquisition module, which is generally various sensor structures. Because the application is mainly used to test and recognize walking posture, the data acquisition module in the shoe sole can more accurately collect the linear acceleration (g) of the user's shoe sole in the front and back (AP), left and right (ML) and vertical (V) three directions and the angular velocity (degree / second or radian / second) of the shoe sole around the front and back (AP), left and right (ML) and vertical (V) three axes. These data will be packaged into data stream after being preliminarily processed in the data acquisition module and sent to the gait event detection module.

[0057] The processing unit is provided in the gait event detection module, which mainly aims to decompose the walking posture into various gait events including initial landing time, toe-off time, support phase, swing phase and gait cycle according to the time node and peak value of data waveguide, wherein the data of left and right feet are completely separated so that the subsequent test can be more accurate. The decomposed data structure will be transmitted to the dimension calculation module.

[0058] Unlike the traditional method of directly observing and collecting the user's gait with high-precision instruments and making judgments, the system of the application has smaller size and simpler structure. In order to ensure the accuracy of testing and recognition, the data structure needs to be decomposed into different independent dimensions. In the application, the gait is quantified into four dimensions through the dimension calculation module under the condition of data stream and data structure state. The four dimensions include gait symmetry dimension for evaluating the balance of left and right feet in the movement mode (especially important for rehabilitation), dynamic stability dimension for evaluating the ability of the body to resist shaking and maintain balance (key to preventing falls), gait efficiency and fluency dimension for evaluating whether the energy transfer is smooth and whether the walking is natural and labor-saving, and body posture and foot contact dimension for evaluating the overall posture of the trunk and the foot contact mode. The scores of the four dimensions are calculated respectively and transmitted to the scoring calculation module.

[0059] Because the gait itself is divided into multiple dimensions, combining the scores of multiple dimensions can quantitatively and real-timely judge whether the current walking posture state is good or not, and whether there is a point to be improved. In the application, the score calculation module configures a weight for each dimension according to different rehabilitation applications or sports performance applications, the sum of the four weights is 1, and when the scores of each dimension enter the score calculation module, the state of the walking posture can be quickly calculated by combining the weights of each dimension. The total score is a number between [0, 100], and the larger the number is, the better the gait is. This can make the user more intuitively see the current state, which is beneficial to improve the rehabilitation or sports effect.

[0060] The intelligent feedback module is used for analyzing the total score, which can more systematically help the user adjust the gait and give feedback, which is beneficial to improve the user's confidence in rehabilitation or sports. BRIEF DESCRIPTION OF DRAWINGS

[0061] In order to more clearly illustrate the technical solutions of the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description only some embodiments of the present application, and for those skilled in the art, other drawings can be obtained without creative labor based on the drawings shown.

[0062] Figure 1 The structure diagram of the walking posture recognition system of the present application;

[0063] Figure 2 The structure diagram of the walking posture recognition device of the present application.

[0064] Explanation of reference numerals:

[0065] 1, data acquisition module; 2, gait event detection module; 3, dimension calculation module; 4, score calculation module; 5, intelligent feedback module.

[0066] The implementation of the object of the present application, the functional characteristics and the advantages will be further described with reference to the embodiments and the drawings. DETAILED DESCRIPTION

[0067] In order to make the object, technical scheme and advantages of the present application more clear, the present application will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application.

[0068] The following will be combined with the drawings Figure 1 The present application will be further described.

[0069] To achieve the above object, the application provides a walking posture recognition system, which comprises a data acquisition module 1, a gait event detection module 2, a dimension calculation module 3, a score calculation module 4 and an intelligent feedback module 5.

[0070] The data acquisition module 1 is used for collecting the linear acceleration and angular velocity of the left and right feet of a user in the front-back, left-right and vertical directions respectively and outputting data streams.

[0071] The gait event detection module 2 is configured with a processing unit based on the peak value detection method, the processing unit receives the data streams and determines the timestamps of the initial landing time, toe-off time, support phase, swing phase and gait cycle of the left foot and / or right foot of the user through the data streams, and simultaneously outputs a data structure based on the determined timestamps.

[0072] The dimension calculation module 3 is used for receiving and dividing the gait of the user into four dimensions, i.e. gait symmetry, dynamic stability, gait efficiency and fluency and body posture and foot contact, based on the content of the data structure, and calculating the scores of the respective dimensions.

[0073] The score calculation module 4 is provided with a weight coefficient for each of the four dimensions based on the dimension calculation module 3, the sum of the four weight coefficients is constrained to be 1, and the score calculation module 4 calculates and outputs a total score based on the respective scores of the four dimensions and the weight coefficients.

[0074] The intelligent feedback module 5 receives the total score and feeds back improvement suggestions to the user in real time according to the total score.

[0075] The walking posture recognition system can solve the problems of multi-dimension, quantification and real-time gait posture scoring, has a simple structure, can be preferably arranged in a portable device such as a shoe of a user, can effectively reduce the cost and improve the recognition accuracy, in order to achieve the object, the system comprises the data acquisition module 1, the gait event detection module 2, the dimension calculation module 3, the score calculation module 4 and the intelligent feedback module 5 from the structure and function, the module arranged in the shoe sole is the data acquisition module 1, which is generally various sensor structures, because the application is mainly used for testing and recognizing the walking posture, the data acquisition module 1 arranged in the shoe sole can more accurately collect the linear acceleration (g) of the shoe sole in the front-back (AP), left-right (ML) and vertical (V) directions and the angular velocity (degree / second or radian / second) of the shoe sole around the front-back (AP), left-right (ML) and vertical (V) axes, the data will be packaged into data streams after being preliminarily processed in the data acquisition module 1 and sent to the gait event detection module 2.

[0076] The gait event detection module 2 is provided with a processing unit, and the main purpose is to decompose the collected data into various gait events according to time nodes and data waveguide peaks, including initial landing time IC (the time when the foot first contacts the ground (usually accompanied by a significant impact peak of vertical acceleration)), toe-off time TO (the time when the toes leave the ground (usually accompanied by an increase in forward acceleration or a specific angular velocity change)), support phase (from IC to TO, the time period when the foot contacts the ground), swing phase (from TO to the next IC, the stage of the foot leaving the ground and swinging forward), gait cycle (from one IC to the next IC of the same side foot), and the like, wherein the data of the left and right feet are completely separated, so that subsequent tests can be more accurate, and the decomposed data structure is transmitted to the multi-dimension calculation module 3.

[0077] Unlike the traditional method of directly observing and collecting the user's gait with high-precision instruments and making judgments, the system of the present application is small in size and simple in structure. In order to ensure the accuracy of the test and identification, the data structure needs to be decomposed into different independent dimensions. In the present application, the gait is quantitatively processed under the condition of data flow and data structure by the multi-dimension calculation module 3. In the present application, the quantifiable gait is divided into four dimensions, including the gait symmetry dimension for evaluating the balance of the left and right feet in the movement mode (especially important for rehabilitation), the dynamic stability dimension for evaluating the body's ability to resist shaking and maintain balance during walking (key to fall prevention), the gait efficiency and fluency dimension for evaluating whether the energy transfer is smooth and whether the walking is natural and energy-saving, and the body posture and foot contact dimension for evaluating the overall posture of the torso and the foot contact mode. The scores of the four dimensions are calculated respectively, and the scores of the four dimensions are transmitted to the scoring calculation module 4.

[0078] Because the gait itself is divided into multiple dimensions, the combination of the scores of the multiple dimensions can quantitatively and real-time judge whether the current walking posture is good or not, and whether there is room for improvement. In the present application, the scoring calculation module 4 configures a weight for each dimension according to different rehabilitation applications or movement performance applications. The combined value of the four weights is 1. When the scores of each dimension enter the scoring calculation module 4, the state of the walking posture can be quickly calculated by combining the weights of each dimension. The total score is a number between [0, 100], and the larger the number, the better the gait. This allows the user to more intuitively see the current state, which is beneficial to improving the rehabilitation or exercise effect.

[0079] The intelligent feedback module 5 is used for analysis according to the total score, which can more systematically help the user adjust the gait and give feedback, which is beneficial to improving the user's confidence in rehabilitation or exercise.

[0080] The data acquisition module 1 further comprises a microprocessor and a data buffer unit in structure, the microprocessor is used for receiving a data stream, and performing preliminary filtering (such as removing noise) and unit conversion on the data stream, and can help the gait event detection module 2 to detect and analyze gait events;

[0081] The data buffer unit is connected to the microprocessor and is used for temporarily storing the data stream to prevent loss during data transmission.

[0082] The data buffer unit is configured with a remote connection module, which is used for remotely transmitting the data stream to external devices (such as a mobile phone APP, a cloud or a rehabilitation device), improving the subsequent data processing accuracy, and ensuring that the device can quickly identify and judge the walking posture.

[0083] The gait symmetry dimension is mainly based on the input preprocessed linear acceleration and angular velocity for subsequent calculation, and based on the comparison of the gait states of the left and right feet, the dimension is divided into four indexes, and the dimension score is calculated through the four indexes:

[0084] Swing_Time_Diff=(abs(T_left_swing-T_right_swing) / max(T_left_swing,T_right_swing))*100%;

[0085] Stance_Time_Diff=(abs(T_left_stance-T_right_stance) / max(T_left_stance,T_right_stance))*100%;

[0086] Step_Length_Diff=(abs(StepLen_left-StepLen_right) / max(StepLen_left,StepLen_right))*100%;

[0087] Step_Length_Diff=(abs(StepLen_left-StepLen_right) / max(StepLen_left,StepLen_right))*100%;

[0088] Strike_Angle_Diff=abs(StrikeAngle_left-StrikeAngle_right);

[0089] Strike_Angle_Diff=abs(StrikeAngle_left-StrikeAngle_right);

[0090] Sym_Score=100-

[0091] (W1*Swing_Time_Diff+W2*Stance_Time_Diff+W3*Step_Length_Diff+W4*Strike_Angle_Diff);

[0092] T_left_swing is the left swing phase, T_right_swing is the right swing phase, T_left_stance is the left stance phase, T_right_stance is the right stance phase, StepLen_left is the left step length, StepLen_right is the right step length, StrikeAngle_left is the initial ground angle difference of the left foot, and StrikeAngle_right is the initial ground angle difference of the right foot.

[0093] W1, W2, W3, and W4 are weight coefficients of the four indicators in evaluating the score of the dimension, the total sum is 1, and the greater the data difference between the left and right feet, the lower the score.

[0094] The score calculation formula of the dynamic stability dimension is:

[0095] Stab_Score = 100-(W5*Norm(ML_Accel_RMS)+W6*Norm(AP_Accel_RMS)+W7*Norm(Gyro_Z_Max)+W8*Norm(Stance_Time_Var));

[0096] W5+W6+W7+W8 = 1;

[0097] ML_Accel_RMS is the root mean square value of the linear acceleration in the left and right directions during the stance phase, which is used to reflect the amplitude of body sway left and right;

[0098] AP_Accel_RMS is the root mean square value of the linear acceleration in the front and back directions during the stance phase, which is used to reflect the amplitude of body sway front and back;

[0099] Gyro_Z_Max is the maximum absolute value of the angular velocity in the vertical direction during the stance phase, which is used to reflect the severity of ankle inversion / eversion;

[0100] Stance_Time_Var is the coefficient of variation of the time of several consecutive stance phases, the calculation formula is: (standard deviation / average value*100%), which is used to reflect the stability of gait rhythm;

[0101] W5, W6, W7, W8 are four indicators of dynamic stability dimension, the weight coefficients in calculating the dimension score, the sum is 1 (for example, W5, W6, W7, W8 are 0.4, 0.3, 0.2, 0.1 respectively), the greater the index value of the four indicators in this dimension (the greater the swing), the more the score is deducted, and the symmetry dimension score range is [0, 100].

[0102] The sub-dimension calculation module 3 is provided with a data input module, and the data input module is used for inputting the height value of the user. The calculation formula of the gait efficiency and fluency dimension is:

[0103] Eff_Score = (W9*Norm_Cadence + W10*Norm(Stride_Regularity) + W11*Norm(Swing_Accel_Peak_Ratio)) - (W12*Norm(Heel_Strike_Sharpness));

[0104] Norm_Cadence is the number of gait cycles per minute, and the height value of the user is normalized. Cadence (steps / min) is the step frequency.

[0105] Stride_Regularity is the autocorrelation coefficient (lag 1 step) of the step length or gait cycle calculated according to the data stream. The closer the value is to 1, the more regular the pace is, and the more fluent it is.

[0106] Swing_Accel_Peak_Ratio is the peak value of acceleration in the swing phase / initial landing acceleration peak value, which is used to reflect the balance of propulsion (swing phase in the middle of the swing) and buffering force (landing). Moderate ratio represents high efficiency.

[0107] Heel_Strike_Sharpness is the rising slope of vertical acceleration or impact peak value (which needs to be filtered) when landing. The value is too large, indicating that the landing is hard, the buffering is poor, the efficiency is low, and it is easy to be injured.

[0108] W9, W10, W11, W12 are four indicators of gait efficiency and fluency dimension, the weight coefficients in calculating the dimension score, and the symmetry dimension score range is [0, 100].

[0109] The sum of the weights W9, W10, W11, W12 is not necessarily 1. Heel_Strike_Sharpness is subtracted as a penalty term. All indicators need to be normalized,

[0110] The dimensionality calculation module 3 is also provided with a posture collection unit for collecting body posture data, because the quantity flow and data structure are transmitted to the remote wireless device, and these data cannot completely determine the posture of the user, therefore, in order to further improve the recognition accuracy, the posture collection unit can manually input the height of the user on the mobile phone, and accurately determine through the mobile phone IMU, the calculation formula of the body posture and the foot contact dimension is:

[0111] Post_Score=100-(W13*abs(Avg_Torso_Angle-Target_Angle)+W14*abs(Foot_Progression_Angle)+W15*Strike_Type_Penalty);

[0112] Avg_Torso_Angle is obtained by combining the body posture data with the gait cycle;

[0113] Target_Angle is the ideal torso angle;

[0114] Foot_Progression_Angle is the average angle between the foot progression direction and the straight line progression direction in the support phase (negative for inside eight, positive for outside eight), and the inside / outside eight is penalized;

[0115] Strike_Type_Penalty is used to quantify the landing mode (determine who lands first among the heel, midfoot, and forefoot). It can be classified by analyzing the acceleration characteristics (impact peak, rise time) at IC or combining the gyroscope (angle change). According to the application scenario (such as the rehabilitation goal is set to midfoot landing), different modes are given scores or penalty values (Strike_Type_Penalty);

[0116] W13, W14, W15 are weight coefficients in the calculation of scores in the body posture and foot contact dimension, the greater the deviation from the target value (angle difference, inside / outside eight degree, non-ideal landing mode), the more points are deducted. The weight W13, W14, W15 allocates the deduction proportion.

[0117] The calculation formula of the score calculation module 4 is:

[0118] Gait_Score=(W_sym*Sym_Score+W_stab*Stab_Score+W_eff*Eff_Score+W_post*Post_Score);

[0119] W_sym, W_stab, W_eff, W_post are respectively the total score weight coefficients of the four dimensions, W_sym+W_stab+W_eff+W_post=1.

[0120] The score calculation module 4 is further provided with a plurality of scene adjustment modules, and the total score weight coefficient values of the four dimensions of each scene adjustment module are different.

[0121] Through the scene adjustment module, the present recognition system has configurability, and the weight can be adjusted according to the application scene.

[0122] When applied in the rehabilitation aspect, the application mainly focuses on symmetry and stability, therefore, the weight can be W_sym=0.40, W_stab=0.30, W_eff=0.20, W_post=0.10.

[0123] When applied in the sports performance, the application mainly focuses on efficiency and posture, therefore, the weight can be W_sym=0.20, W_stab=0.20, W_eff=0.35, W_post=0.25.

[0124] Through the above system structure, the whole recognition system can be quantified from multiple dimensions, and the current application aspect score can be accurately judged in real time, which is beneficial to the user to adjust the walking posture.

[0125] In combination with the description of the above system, Figure 2 The present application further provides a walking posture recognition device, which comprises the walking posture recognition system described above, and comprises a shoe body, a three-axis accelerometer and a three-axis gyroscope arranged in the shoe body, and a microprocessor connected to the three-axis accelerometer and / or the three-axis gyroscope at the bottom of the shoe body.

[0126] The microprocessor is connected with a data buffer and is provided with a wireless transmission element for connecting with a wireless device.

[0127] The walking posture recognition device adopts the above system and is pre-prepared in the shoe body, the main sensors for testing are the three-axis accelerometer and the three-axis gyroscope, the microprocessor and the wireless transmission element for preprocessing are used to guarantee the testing accuracy, the subsequent system will be installed in the mobile phone APP, which can effectively reduce the size of the elements in the shoes and avoid the damage of the recognition device in use.

[0128] The same or similar reference numerals in the drawings of the embodiments correspond to the same or similar components; in the description of the present application, it is understood that if the orientations or positional relationships indicated by the terms "upper", "lower", "left", "right" and the like are based on the orientations or positional relationships shown in the drawings, they are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, therefore the terms describing the positional relationships in the drawings are only used for exemplary illustration, and cannot be understood as a limitation on the present patent, for those skilled in the art, the specific meanings of the above terms can be understood according to the specific circumstances.

[0129] The above is only a preferred embodiment of the present application, and is not used to limit the present application, any modification, equivalent replacement and improvement made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A walking posture recognition system, characterized by, Comprise: Data acquisition module, the data acquisition module is used to collect the linear acceleration and angular velocity of the front and back, left and right, vertical three directions of the left and right feet of the user respectively, and output data stream; Gait event detection module, the gait event detection module is configured with processing unit based on peak detection method, processing unit receives data stream and determines the time stamp of multiple gait events such as initial landing time, toe-off time, support phase, swing phase and gait cycle of the left foot and / or right foot of the user through data stream, and outputs data structure based on the determined time stamp; Dimension calculation module, the dimension calculation module is used to receive and divide the user gait into four dimensions of gait symmetry, dynamic stability, gait efficiency and fluency, and body posture and foot contact based on the content of data structure, and calculate the score of each dimension respectively; Scoring calculation module, the scoring calculation module is provided with one weight coefficient for each of the four dimensions based on the dimension calculation module, and the sum of the four weight coefficients is constrained to be 1, the scoring calculation module calculates and outputs the total score based on the respective scores and weight coefficients of the four dimensions; Intelligent feedback module, the intelligent feedback module receives the total score, and feeds back improvement suggestions to the user in real time according to the total score.

2. The walking posture recognition system according to claim 1, wherein The data acquisition module further comprises: Microprocessor, the microprocessor is used to receive data stream, and carry out preliminary filtering and unit conversion on data stream; Data buffer unit, the data buffer unit is connected to the microprocessor for temporarily storing data stream.

3. The walking posture recognition system according to claim 2, wherein The data buffer unit is configured with a remote connection module for remotely transmitting data stream to external equipment, improving the accuracy of subsequent data processing.

4. The walking posture recognition system of claim 1, wherein, The gait symmetry dimension is divided into four indexes based on comparing the gait states of the left and right feet of the user, and the dimension score is calculated through the four indexes: Swing phase time difference percentage: Swing_Time_Diff=(abs(T_left_swing-T_right_swing) / max(T_left_swing,T_right_swing))*100%; Support phase time difference percentage: Stance_Time_Diff=(abs(T_left_stance-T_right_stance) / max(T_left_stance,T_right_stance))*100%; By twice integration of linear acceleration in data stream or gait cycle and speed estimation, the step length of the left and right feet of the user is determined, and the step length difference percentage is calculated: Step_Length_Diff=(abs(StepLen_left-StepLen_right) / max(StepLen_left,StepLen_right))*100%; The initial ground angle difference of the left and right feet of the user is calculated by the linear acceleration and angular velocity of the initial landing time in data stream: Strike_Angle_Diff=abs(StrikeAngle_left-StrikeAngle_right); Sym_Score = 100 - (W1*Swing_Time_Diff+W2*Stance_Time_Diff+W3*Step_Length_Diff+W4*Stike_Angle_Diff); T_left_swing for left swing phase, T_right_swing for right swing phase, T_left_stance for left stance phase, T_right_stance for right stance phase, StepLen_left for left step length, StepLen_right for right step length, StrikeAngle_left for left initial ground angle difference, StrikeAngle_right for right initial ground angle difference; W1, W2, W3, W4 are four index weight coefficients in the evaluation of the dimension score, the total is 1, the symmetry dimension score range is [0, 100], the greater the difference, the lower the score. The score calculation formula of the dynamic stability dimension is:

5. The walking posture recognition system according to claim 4, wherein Stab_Score = 100 - (W5*Norm(ML_Accel_RMS)+W6*Norm(AP_Accel_RMS)+W7*Norm(Gyro_Z_Max)+W8*Norm(Stance_Time_Var)); W5+W6+W7+W8 = 1; ML_Accel_RMS is the root mean square value of linear acceleration in the left and right directions during the stance phase, which is used to reflect the amplitude of body sway left and right; AP_Accel_RMS is the root mean square value of linear acceleration in the front and back directions during the stance phase, which is used to reflect the amplitude of body sway front and back; Gyro_Z_Max is the maximum absolute value of angular velocity in the vertical direction during the stance phase, which is used to reflect the severity of ankle inversion / eversion; Stance_Time_Var is the coefficient of variation of the time of several consecutive stance phases, the calculation formula is: (standard deviation / average value*100%), which is used to reflect the stability of gait rhythm; W5, W6, W7, W8 are four indexes of dynamic stability dimension, which are weight coefficients in the calculation of dimension score, the total is 1, and the symmetry dimension score range is [0, 100]. The dimension calculation module is provided with a data input module, which is used to input the height value of the user, and the calculation formula of gait efficiency and fluency dimension is:

6. The walking posture recognition system according to claim 5, wherein Eff_Score = (W9*Norm_Cadence+W10*Norm(Stride_Regularity)+W11*Norm(Swing_Accel_Peak_Ratio))-(W12*Norm(Heel_Strike_Sharpness)); Norm_Cadence is the number of gait cycles per minute, which is normalized by the height value of the user. ​ Stride_Regularity is the autocorrelation coefficient of stride length or gait cycle calculated from the data stream; Swing_Accel_Peak_Ratio is the peak acceleration in swing phase / peak acceleration at initial contact, which reflects the balance between propulsion and braking; Heel_Strike_Sharpness is the slope of vertical acceleration at initial contact, which reflects the harshness of foot strike; W9, W10, W11, W12 are the weight coefficients of four indicators of gait efficiency and fluency dimension, and the symmetry dimension score ranges from 0 to 100.

7. The walking posture recognition system according to claim 6, wherein The sub-dimension calculation module is further provided with a posture collection unit for collecting body posture data, and the calculation formula of the body posture and foot contact dimension is: Post_Score = 100-(W13*abs(Avg_Torso_Angle-Target_Angle)+W14*abs(Foot_Progression_Angle)+W15*Strike_Type_Penalty); Avg_Torso_Angle is obtained by combining body posture data with gait cycle; Target_Angle is the ideal torso angle; Foot_Progression_Angle is the average angle between the foot progression direction and the straight line direction in the support phase; Strike_Type_Penalty is the foot strike position determined based on three-dimensional linear acceleration and angular velocity at the initial contact time; W13, W14, W15 are the weight coefficients for calculating the score in the body posture and foot contact dimension.

8. The walking posture recognition system according to claim 7, wherein The calculation formula of the score calculation module is: Gait_Score = (W_sym*Sym_Score+W_stab*Stab_Score+W_eff*Eff_Score+W_post*Post_Score); W_sym, W_stab, W_eff, W_post are the total score weight coefficients of the four dimensions, and W_sym+W_stab+W_eff+W_post = 1.

9. The walking posture recognition system according to claim 8, wherein The score calculation module is further provided with a plurality of scene adjustment modules, and the total score weight coefficient values of the four dimensions of each scene adjustment module are different.

10. A walking posture recognition device characterized by comprising: The walking posture recognition system comprises a shoe body, a three-axis accelerometer and a three-axis gyroscope arranged in the shoe body, and a microprocessor connected to the three-axis accelerometer and / or the three-axis gyroscope at the bottom of the shoe body. The microprocessor is connected with a data buffer and is configured with a wireless transmission element for connecting with a wireless device.

Images