Terrain recognition method based on lower limb inertia information

Abstract

The invention relates to a terrain recognition method based on lower limb inertia information, belongs to the technical field of terrain recognition, and solves the problems that a terrain recognition method in the prior art is high in cost, low in recognition precision and not suitable for complex terrains. Constructing a self-oscillation sine curve based on the hip joint angle to obtain a first terrain recognition result, performing energy peak detection based on acceleration data output by an inertial measurement unit of the back waist, and obtaining a second terrain recognition result based on a detection result; acquiring amplitudes and offsets of a hip joint angle curve, a knee joint angle curve and a front inclination angle curve, and comparing the amplitudes and the offsets with amplitude ranges and offset ranges of an upper body front inclination angle, a hip joint angle and a knee joint angle corresponding to each terrain stored in a self-adaptive threshold knowledge base to obtain a third terrain recognition result; and obtaining a final terrain recognition result based on the first, second and third recognition results. The lower limb assisted exoskeleton terrain recognition method is low in cost and high in recognition precision.

Description

Technical Field

[0001] The present invention relates to the technical field of terrain recognition, and in particular, to a terrain recognition method based on lower limb inertial information. Background Art

[0002] A lower limb assistive exoskeleton is a mechanical device worn on the human lower limbs. It can real-time sense the terrain type (such as flat ground, slope, stairs, etc.) when the user walks through sensors and algorithms, and dynamically adjust the assistive strategy of the exoskeleton based on the recognition result. Therefore, terrain recognition is of great significance for lower limb assistive exoskeletons. It enables the exoskeleton to adjust the assistive mode according to different terrain conditions, such as flat ground, slope, stairs, etc., so as to optimize the human movement efficiency, reduce energy consumption, and improve the safety and comfort of movement.

[0003] In the prior art, the terrain recognition methods for lower limb assistive exoskeletons mainly include: (1) a terrain recognition method for lower limb assistive exoskeletons based on visual sensors, which collects the front ground image or three-dimensional point cloud data by mounting a camera or a depth camera on the exoskeleton, and uses computer vision or deep learning models to identify the terrain category. However, this method depends on light and environment, requires high computing resources, and increases the load and the complexity of wearing; (2) a method based on inertial measurement units (IMUs) and pressure sensors, which is based on the IMU data of both thighs and calves and the sole, and the sole pressure sensor data. Through gait cycle division, filtering and normalization processing, and through a neural network model or threshold classification method to achieve terrain recognition. However, this method has low recognition accuracy and limited recognition ability for complex terrains.

[0004] Therefore, there is a need to provide a terrain recognition method for lower limb assistive exoskeletons with low cost, high recognition accuracy, and applicable to complex terrains. Summary of the Invention

[0005] In view of the above analysis, an embodiment of the present invention aims to provide a terrain recognition method based on lower limb inertial information to solve the problems of high cost, low recognition accuracy, and inapplicability to complex terrains of existing terrain recognition methods.

[0006] An embodiment of the present invention provides a terrain recognition method based on lower limb inertial information, including: Arranging inertial measurement units on the left and right thighs, left and right calves, and the lower back of the human body to be measured; Calculating the hip joint angle based on the output data of the inertial measurement units on the left and right thighs and the lower back, constructing a self-oscillating sine curve based on the hip joint angle, and obtaining a first terrain recognition result based on the self-oscillating sine curve; Performing energy spike detection based on the acceleration data output by the inertial measurement unit on the lower back, and obtaining a second terrain recognition result based on the detection result; Calculate the hip joint angle, knee joint angle, and anterior tilt angle of the human body to be measured at each moment during the current movement based on the output data of the inertial measurement units on the left and right thighs, left and right calves, and the lower back, thereby constructing a hip joint angle curve, a knee joint angle curve, and an anterior tilt angle curve, obtaining the amplitude and offset corresponding to each of the above curves, and comparing them with the amplitude range and offset range of the upper body anterior tilt angle, hip joint angle, and knee joint angle corresponding to each terrain stored in the adaptive threshold knowledge base to obtain the third terrain recognition result; Obtain the final terrain recognition result based on the first terrain recognition result, the second terrain recognition result, and the third terrain recognition result.

[0007] Based on a further improvement of the above method, the hip joint angle includes: a first hip joint angle calculated based on the output data of the inertial measurement units on the left thigh and the lower back; a second hip joint angle calculated based on the output data of the inertial measurement units on the right thigh and the lower back; the self-oscillating sine curve includes: a first self-oscillating sine curve constructed based on the first hip joint angle, and a second self-oscillating sine curve constructed based on the second hip joint angle.

[0008] Based on a further improvement of the above method, constructing the self-oscillating sine curve based on the hip joint angle includes: S1: Initialize the amplitude, frequency, phase, and offset of the self-oscillating sine curve; S2: Obtain the actual hip joint angle at the current moment, and calculate the tracking error based on the actual hip joint angle and the simulated hip joint angle output by the self-oscillating sine curve; S3: If the tracking error is less than the first target threshold, use the current self-oscillating sine curve as the self-oscillating sine curve corresponding to the human body to be measured; if the tracking error is greater than or equal to the first target threshold, update the amplitude, frequency, phase, and offset of the self-oscillating sine curve at the current moment based on the tracking error, and use the updated values as the amplitude, frequency, phase, and offset of the self-oscillating sine curve at the next moment, and return to step S2.

[0009] Based on a further improvement of the above method, obtaining the first terrain recognition result based on the self-oscillating sine curve includes: Compare the amplitude, frequency, phase, and offset in the first self-oscillating sine curve with a preset first threshold classification table to obtain a first initial terrain recognition result, and compare the amplitude, frequency, phase, and offset in the second self-oscillating sine curve with a preset first threshold classification table to obtain a second initial terrain recognition result; If the first initial terrain recognition result is consistent with the second initial terrain recognition result, then use the first initial terrain recognition result or the second initial terrain recognition result as the first terrain recognition result; if they are inconsistent, then use the irregular terrain as the first terrain recognition result; Among them, the first threshold classification table is used to store the amplitude range, frequency range, phase range, and bias range corresponding to each type of terrain.

[0010] Based on a further improvement of the above method, updating the amplitude, frequency, phase, and bias of the self-oscillating sine curve at the current moment based on the tracking error includes: ; Among them, is the updated phase, is the updated frequency, is the updated amplitude, is the updated bias, is the lateral learning parameter, is the longitudinal learning parameter, is the tracking error.

[0011] Based on a further improvement of the above method, obtaining the final terrain recognition result based on the first terrain recognition result, the second terrain recognition result, and the third terrain recognition result includes: If the first terrain recognition result, the second terrain recognition result, and the third terrain recognition result are the same, then use the first terrain recognition result, the second terrain recognition result, or the third terrain recognition result as the final terrain recognition result; If the first terrain recognition result, the second terrain recognition result, and the third terrain recognition result are different, then construct a recognition result set with the terrain types that appear in each terrain recognition result, and calculate the probability of each terrain type in the recognition result set based on the weights of the first terrain recognition result, the second terrain recognition result, and the third terrain recognition result, and use the terrain type with the highest probability as the final terrain recognition result; The terrain recognition result includes: flat ground type, going up stairs type, going uphill type, going down stairs type, going downhill type, irregular terrain type.

[0012] Based on a further improvement of the above method, calculating the probability of each terrain type in the recognition result set includes: , Among them, c is the terrain type in the recognition result set, is the probability value when the terrain type is c, , , is the result value for terrain type c; when the terrain type c is consistent with the first terrain recognition result, , otherwise, ; when the terrain type c is consistent with the second terrain recognition result, , otherwise, ; when the terrain type c is consistent with the third terrain recognition result, , otherwise, ; , , are the weights corresponding to the first terrain recognition result, the second terrain recognition result, and the third terrain recognition result.

[0013] Based on a further improvement of the above method, performing energy spike detection based on the acceleration data output by the inertial measurement unit on the lower back, and obtaining a second terrain recognition result based on the detection result, including: Obtain the acceleration data, use the time period when the acceleration value in the acceleration data exceeds a preset detection threshold as the spike detection interval, and perform energy spike detection based on the spike detection interval to determine the touchdown moment; According to the output data of the inertial measurement unit at the touchdown moment and the limb lengths of the left and right thighs and the left and right calves, construct a human lower limb model at the touchdown moment, determine the centroid coordinates of the left and right feet based on the human lower limb model, and obtain a second terrain recognition result based on the centroid coordinates of the left and right feet.

[0014] Based on a further improvement of the above method, the adaptive threshold knowledge base is used to store the first offset range of the upper body forward tilt angle, the second amplitude range and the second offset range of the hip joint angle, and the third amplitude range and the third offset range of the knee joint angle corresponding to different terrains; The first offset range, the second amplitude range and the second offset range of the hip joint angle, and the third amplitude range and the third offset range of the knee joint angle are obtained through the following method: For any terrain, obtain the upper body forward tilt angle, hip joint angle, and knee joint angle of different people at different times, construct an upper body forward tilt angle curve, a hip joint angle curve, and a knee joint angle curve based on each moment and its corresponding upper body forward tilt angle, hip joint angle, and knee joint angle, and extract the amplitudes and offsets corresponding to the upper body forward tilt angle curve, the hip joint angle curve, and the knee joint angle curve; Obtain the first offset range based on the offsets corresponding to the upper body forward tilt angle curves of all people; Obtain the second amplitude range and the second offset range based on the amplitudes and offsets corresponding to the hip joint angle curves of all people; The third amplitude range and the third offset range are obtained based on the amplitudes and offsets corresponding to the knee joint angle curves of all people.

[0015] Based on a further improvement of the above method, the comparison of the amplitudes and offset ranges of the upper body forward tilt angle, hip joint angle, and knee joint angle stored in the adaptive threshold knowledge base to obtain the third terrain recognition result includes: For any terrain stored in the adaptive threshold knowledge base, if the offset of the forward tilt angle is within the first offset range, and the amplitude and offset of the first hip joint angle are respectively within the second amplitude range and the second offset range, or the amplitude and offset of the second hip joint angle are respectively within the second amplitude range and the second offset range, and the amplitude and offset of the first knee joint angle are respectively within the third amplitude range and the third offset range, or the amplitude and offset of the second knee joint angle are respectively within the third amplitude range and the third offset range, then this terrain is used as the third terrain recognition result. If no conforming terrain exists, an irregular terrain is used as the third terrain recognition result.

[0016] Compared with the prior art, the present invention can at least achieve one of the following beneficial effects: The present invention provides a terrain recognition method based on lower limb inertial information. The final terrain recognition result is determined based on the recognition results of the terrain recognition algorithm based on an adaptive oscillator, the terrain recognition algorithm based on ground contact detection, and the terrain recognition algorithm based on joint angle feature information, reducing the error of obtaining the recognition result by only using a single recognition method and improving the accuracy and reliability of the terrain recognition result. In addition, the method proposed by the present invention can achieve terrain recognition only with the data obtained by five inertial measurement units. Compared with the prior art's terrain recognition method for lower limb assistive exoskeletons based on visual sensors, its hardware cost is lower. Moreover, all the inertial measurement units are arranged on the human body, making this method not affected by the external environment. The high-precision motion capture ability and efficient dynamic response ability enable this method to better apply to complex terrain recognition.

[0017] In the present invention, the above technical solutions can also be combined with each other to achieve more preferred combination schemes. Other features and advantages of the present invention will be described in the subsequent specification, and some advantages can be made obvious from the specification or understood by implementing the present invention. The objectives and other advantages of the present invention can be realized and obtained through the content specifically pointed out in the specification and the drawings. Description of the Drawings

[0018] The drawings are only for the purpose of showing specific embodiments and are not considered to be a limitation of the present invention. Throughout the drawings, the same reference signs represent the same components; Figure 1 This is an example diagram of a terrain recognition method based on lower limb inertial information in an embodiment of the present invention; Figure 2 This is an example diagram of energy peak detection in an embodiment of the present invention; Figure 3 This is an example diagram of a human lower limb model in an embodiment of the present invention. DETAILED DESCRIPTION

[0019] The preferred embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, and are not used to limit the scope of the present invention.

[0020] A specific embodiment of the present invention discloses a terrain recognition method based on lower limb inertial information, such as Figure 1 Shown, including: S1: Inertial measurement units are placed on the left and right thighs, left and right calves, and lower back of the human body to be measured.

[0021] Among them, the layout position and angle of the inertial measurement unit (IMU) directly determine the validity of sensor data and the accuracy of terrain classification.

[0022] For both thighs, the IMU can be placed on the front of the thigh, which makes it easier to hide and fix the sensor cable and improves wearing stability. To effectively reduce the impact of muscle contraction or soft tissue movement on the sensor and ensure measurement stability, it can be placed on the midline of the front thigh (near the rectus femoris), about 15-20 cm away from the hip joint (or 1 / 3 of the thigh length).

[0023] For the left and right calves, the IMU can be placed on the front of the calf. In order to reduce the sensor shaking caused by the swing of the calf during walking, avoid the tendon activity area, and avoid the sensor displacement caused by the violent contraction of the tibialis anterior muscle, it can be placed on the midline of the front of the calf (near the tibialis anterior muscle), about 10-15 cm away from the knee joint (or 1 / 2 of the calf length).

[0024] For the lower back, because it provides a global motion reference (such as the direction of gravitational acceleration), it can also be used to calibrate the coordinate system of the lower limb sensors. It is usually placed at the L3 or L4 position of the lumbar spine, near the midline of the spine. It can effectively reduce the coupling interference of trunk rotation on lower limb movement, so that the output data of the IMU can represent the movement of the pelvis.

[0025] After placing the inertial measurement units on the left and right thighs, left and right calves, and lower back of the human body, the IMU installation angle needs to be calibrated to ensure the accuracy of kinematic calculations. For example, the calibration includes: (1) The skeletal axes are aligned, that is, the coordinate systems of the thigh and calf IMUs need to be strictly aligned with the anatomical axes of the femur and tibia (e.g., the Y-axis is along the long axis of the bone), ensuring that the measured angular velocity / acceleration reflects the actual limb movement.

[0026] The calibration methods include: static calibration. When the user stands upright and still, the sensor coordinate system is aligned using the direction of gravity (the Z-axis is vertically downward); dynamic verification: verify whether the rotation of the coordinate system matches the joint movement plane through walking or leg-lifting movements.

[0027] (2) Establish a multi-IMU synchronous reference system. Use the lower back IMU as the origin of the global reference system, and other IMUs are transformed to this coordinate system through the initial attitude to ensure data spatial consistency.

[0028] Optionally, for the front of the thigh, the x-axis of the IMU is along the long axis of the thigh (from the hip joint to the knee joint), the y-axis points to the side of the human body, and the z-axis is perpendicular to the first two axes. For the front of the calf, the x-axis of the IMU is along the long axis of the calf (from the knee joint to the ankle joint), the y-axis points to the side of the human body, and the z-axis is perpendicular to the first two axes. For the lower back, the x-axis of the IMU is roughly along the front-back direction of the human body (from the front side of the pelvis to the back side), the y-axis is along the left-right direction of the human body, and the z-axis is along the vertical direction of the human body.

[0029] S2: Calculate the hip joint angle based on the output data of the inertial measurement units of the left and right thighs and the lower back, construct a self-oscillating sine curve based on the hip joint angle, and obtain a first terrain recognition result based on the self-oscillating sine curve.

[0030] The output data refers to the three-axis acceleration, three-axis angular velocity output by the IMU sensor, and the three-axis magnetic field intensity data in the case of including a magnetometer. Before calculating the hip joint angle, it is necessary to verify whether the collected IMU data at each position is synchronized in time, so as to facilitate subsequent accurate correlation analysis and fusion calculation of data at different positions. Exemplarily, data synchronization can be achieved by setting a unified timestamp in the data acquisition system or using an external synchronization signal.

[0031] The calculating the hip joint angle based on the output data of the inertial measurement units of the left and right thighs and the lower back includes: A1: Filter the three output data to remove noise and unnecessary high-frequency interference. For example, using a low-pass filter can remove high-frequency noise and retain the low-frequency signals related to human movement; A2: Transform the local coordinate systems of the three IMUs into a unified reference coordinate system. Usually, the IMU at the lower back is taken as the origin of the global reference system, and mathematical tools such as coordinate transformation matrices or quaternions are used to align the coordinate systems, ensuring that data at different positions can be compared and calculated in the same reference system; A3: Adopt data fusion algorithms to estimate the posture of each part where the IMU is located. Common data fusion algorithms include Extended Kalman Filter (EKF), complementary filtering, etc. These algorithms can integrate the data of accelerometers, gyroscopes, and magnetometers (if any) to obtain more accurate and stable human body part posture information. Represent the estimated posture information in a suitable form. Usually, quaternions or Euler angles can be used. Quaternions have advantages such as stable calculation and no singularities when representing three-dimensional rotations, and are suitable for subsequent posture operations and fusions; while Euler angles more intuitively reflect the angular changes of human joints, facilitating the final hip joint angle calculation and interpretation; A4: Use the IMU posture data of the left and right thighs and the lower back to calculate the posture difference of the thighs relative to the lower back (pelvis), that is, calculate the relative rotation matrix of the thigh posture relative to the trunk posture, and convert the relative rotation matrix into Euler angles (i.e., flexion and extension angles), corresponding to the movement angles of the hip joint.

[0032] Optionally, after obtaining the hip joint angle, error correction and compensation can also be performed to reduce the angle calculation deviation caused by factors such as sensor errors, inaccurate model assumptions, and cumulative errors during the data processing process, thereby improving the accuracy and reliability of the hip joint angle calculation.

[0033] Constructing the self-oscillating sine curve based on the hip joint angle includes: N1: Initialize the amplitude, frequency, phase, and bias of the self-oscillating sine curve.

[0034] The mathematical form of the self-oscillating sine curve is: , where is the oscillator amplitude, is the oscillator frequency, is the oscillator phase, is the oscillator bias, is the simulated hip joint angle.

[0035] When initializing each parameter of the self-oscillating sine curve, it can be set randomly or according to experience or historical templates. Each piece of data in the historical template includes factors such as the height, weight, age, and gender of the human body, as well as the values of the amplitude, frequency, phase, and offset of the corresponding self-oscillating sine curve. The user can perform initialization settings according to the values of each parameter of the template that is roughly the same as the factors such as the height, weight, age, and gender of the human body to be measured and stored.

[0036] N2: Obtain the actual hip joint angle at the current moment, and calculate the tracking error based on the actual hip joint angle and the simulated hip joint angle output by the self-oscillating sine curve.

[0037] Among them, the calculation of the tracking error based on the actual hip joint angle and the simulated hip joint angle output by the self-oscillating sine curve can be expressed as: , Among them, is the actual hip joint angle.

[0038] N3: If the tracking error is less than the first target threshold, then use the current self-oscillating sine curve as the self-oscillating sine curve corresponding to the human body to be measured; if the tracking error is greater than or equal to the first target threshold, then update the amplitude, frequency, phase, and offset of the self-oscillating sine curve at the current moment based on the tracking error, and use the updated values as the amplitude, frequency, phase, and offset of the self-oscillating sine curve at the next moment, and return to step N2.

[0039] Among them, the update of the amplitude, frequency, phase, and offset of the self-oscillating sine curve at the current moment based on the tracking error includes: ; Among them, is the updated phase, is the updated frequency, is the updated amplitude, is the updated offset, is the lateral learning parameter, is the longitudinal learning parameter, is the tracking error.

[0040] Exemplarily, the first target threshold can be set according to actual needs or experience. Preferably, the first target threshold is 0.01.

[0041] As can be seen from step S1, the terrain recognition method proposed by the present invention arranges inertial measurement units on the left and right thighs, left and right calves, and the lower back of the human body, a total of 5 IMUs. When calculating the hip joint angle, it is necessary to consider the hip joint angle between the left thigh and the lower back and the hip joint angle between the right thigh and the lower back. Therefore, the actual hip joint angle includes: a first hip joint angle calculated based on the output data of the inertial measurement units of the left thigh and the lower back; a second hip joint angle calculated based on the output data of the inertial measurement units of the right thigh and the lower back. Since two hip joint angles are obtained, there are also two self-oscillating sine curves, that is, a first self-oscillating sine curve constructed based on the first hip joint angle, and a second self-oscillating sine curve constructed based on the second hip joint angle. Since two self-oscillating sine curves are obtained, there are also two terrain recognition results, that is, a first initial terrain recognition result corresponding to the left leg and a second initial terrain recognition result corresponding to the right leg.

[0042] The obtaining of the first terrain recognition result based on the self-oscillating sine curve includes: Comparing the amplitude, frequency, phase, and offset in the first self-oscillating sine curve with a preset first threshold classification table to obtain a first initial terrain recognition result, and comparing the amplitude, frequency, phase, and offset in the second self-oscillating sine curve with the preset first threshold classification table to obtain a second initial terrain recognition result; If the first initial terrain recognition result is consistent with the second initial terrain recognition result, then use the first initial terrain recognition result or the second initial terrain recognition result as the first terrain recognition result. If they are inconsistent, then use the irregular terrain as the first terrain recognition result.

[0043] Among them, the first threshold classification table is used to store the amplitude range, frequency range, phase range, and offset range corresponding to each type of terrain. The first threshold classification table is shown in Table 1.

[0044] Comparing the amplitude, frequency, phase, and offset in the first self-oscillating sine curve or the amplitude, frequency, phase, and offset in the second self-oscillating sine curve with the preset first threshold classification table to obtain the initial terrain recognition result means: M1: Judge whether the phase of the current self-oscillating sine curve conforms to the phase range specified for each terrain. If it conforms, it indicates that the adaptive oscillator is operating normally, and execute step M2; if it does not conform, then use the irregular terrain as the initial terrain recognition result, and reduce the weight of the adaptive oscillator recognition in the subsequent step S5, and at the same time output an error message to indicate that the current adaptive oscillator fails to be enabled normally, and jump to M5.

[0045] M2: Determine whether the frequency of the current self-oscillating sine curve conforms to the frequency range specified for each terrain. If the current frequency is within the specified frequency range, maintain the amplitude range and offset range specified for each terrain. If the current frequency is not within the specified frequency range, reduce the amplitude range and offset range specified for each terrain by 10%.

[0046] Exemplarily, set the frequency range to 0 - 0.7. If the current frequency is less than or equal to 0.7, the amplitude range and offset range specified for each terrain remain unchanged. If the current frequency is greater than 0.7, reduce the amplitude range and offset range specified for each terrain by 10%.

[0047] M3: Determine whether the amplitude of the current self-oscillating sine curve conforms to the amplitude range of any terrain type. If it conforms, execute step M4. If it does not conform, take the irregular terrain as the initial terrain recognition result and jump to M5.

[0048] M4: Determine whether the offset value of the current self-oscillating sine curve conforms to the offset range specified for this terrain (i.e., the terrain type determined in step M3). If it conforms, execute step M5. If it does not conform, take the irregular terrain as the initial terrain recognition result and jump to M5.

[0049] M5: Output the initial terrain recognition result.

[0050] Table 1 ; S3: Perform energy spike detection based on the acceleration data output by the inertial measurement unit of the lower back, and obtain the second terrain recognition result based on the detection result, including: Obtain the acceleration data, take the time period when the acceleration value in the acceleration data exceeds the preset detection threshold as the spike detection interval, and perform energy spike detection based on the spike detection interval to determine the touchdown moment; According to the output data of the inertial measurement unit at the touchdown moment and the limb lengths of the left and right thighs and left and right calves, construct a human lower limb model at the touchdown moment, determine the left and right foot centroid coordinates based on the human lower limb model, and obtain the second terrain recognition result based on the left and right foot centroid coordinates.

[0051] During walking, at the moment when the foot touches the ground, the lower back IMU node will generate a relatively obvious acceleration spike. By detecting this spike, the moment when the foot touches the ground during walking can be detected. After detecting the moment when the foot touches the ground, establish a lower limb posture model through the lower limb joint angles, obtain the spatial positions of the two feet at the touchdown moment through the lower limb posture model, and identify the current terrain.

[0052] In touchdown detection, false detection caused by glitches is avoided by combining threshold detection and energy spike detection. First, a preset detection threshold is defined. When the absolute value of the acceleration exceeds this threshold, spike detection starts, and when the absolute value of the acceleration drops below this threshold, spike detection ends. Integrate the acceleration during spike detection. If it exceeds a certain energy, it means that this spike can be used as a single-foot touchdown estimation point. If it does not exceed, it means that this spike may be caused by glitches and is not used as a single-foot touchdown estimation point. The energy spike detection formula is as follows: , where, is the acceleration data, and t1~t2 are the lower and upper limits of the time interval above the preset acceleration detection threshold.

[0053] Figure 2 An example diagram of energy spike detection is given. Exemplarily, the preset detection threshold can be set to 0.36.

[0054] After calculating the acceleration integral value of the current interval based on the energy spike detection formula, compare it with the preset energy threshold. If the current acceleration integral value is greater than the preset energy threshold, use the point with the maximum acceleration value within the current interval as the touchdown point, and the moment corresponding to this maximum acceleration value is the touchdown moment. Among them, the preset energy threshold can be set based on experience or can be set artificially by considering factors such as the height and weight of the current human body to be measured.

[0055] After detecting the touchdown moment, judge the current motion mode of the human body and thus judge the current terrain by analyzing information such as the touchdown height of the heel and the touchdown distance of the heel through the currently constructed human lower limb model. The current terrain can be judged through the centroid coordinates of the two feet at touchdown. For example, when on flat ground, the vertical coordinates of the centroids of the two feet should be approximately the same. When going upstairs, the horizontal and vertical coordinates of the centroid of the front foot are both higher, and the horizontal and vertical coordinates of the centroid of the rear foot are both lower. When going downstairs, the horizontal coordinate of the front foot is high and the vertical coordinate is low, and vice versa for the rear foot.

[0056] The human lower limb model is as Figure 3 shown, is the centroid of the left foot, is the centroid of the right foot, and are the angles between the left and right lower legs and the vertical direction respectively, and are the angles between the left and right thighs and the vertical direction respectively, obtained through inertial sensors, and represent the lengths of the left and right lower legs, and represent the lengths of the left and right thighs, and then represent and the coordinates of two points. Assuming that the waist is the origin of the coordinate system, the coordinates of the center of mass of the two feet are: , .

[0057] S4: Calculate the hip joint angle, knee joint angle and anterior tilt angle of the human body to be measured at each moment during the current movement based on the output data of the inertial measurement units of the left and right thighs, left and right calves and the lower back, so as to construct a hip joint angle curve, a knee joint angle curve and an anterior tilt angle curve, obtain the amplitudes and offsets corresponding to the above curves, and compare them with the amplitude ranges and offset ranges of the anterior tilt angle, hip joint angle and knee joint angle of the upper body corresponding to each terrain stored in the adaptive threshold knowledge base to obtain the third terrain recognition result.

[0058] The limitation on the output data in this step is the same as that in step S2.

[0059] It is particularly worth noting that the terrain recognition method proposed by the present invention is a method that runs in real time in a lower limb assisted exoskeleton mechanical device. Therefore, in this step, the hip joint angle curve, knee joint angle curve and anterior tilt angle curve actually constructed are all curves formed with the end moment of the previous sampling period of the current moment as the starting moment. Furthermore, the amplitudes and offsets extracted are also the amplitudes and offsets corresponding to each curve with the end moment of the previous sampling period as the starting moment. That is, the terrain type at the current moment is judged (predicted) according to the characteristics of the curve formed by the current sampling period, and a rapid response is made according to the recognized terrain type. Exemplarily, assuming that the current moment is t, the end moment of the previous sampling period of the current moment is obtained and denoted as t-n, where n is the sampling time window, generally selected to include a complete cycle of one gait of the left and right legs, about 1 second. Using the hip joint angle, knee joint angle and anterior tilt angle at each moment of t-n, t-n+1, t-n+2, t-n+3,..., t-1, construct a hip joint angle curve, a knee joint angle curve and an anterior tilt angle curve, and extract the amplitudes and offsets corresponding to each curve.

[0060] The knee joint angle is calculated in the following way: B1: Filter the four output data to remove noise and unnecessary high-frequency interference. For example, using a low-pass filter can remove high-frequency noise and retain low-frequency signals related to human movement.

[0061] B2: Transform the local coordinate systems of the four IMUs to a unified reference coordinate system. Usually, the IMU at the lower back is taken as the origin of the global reference system, and mathematical tools such as coordinate transformation matrices or quaternions are used to align the coordinate systems, ensuring that data at different positions can be compared and calculated in the same reference system.

[0062] B3: Adopt data fusion algorithms to estimate the posture of each body part where the IMU is located. Common data fusion algorithms include Extended Kalman Filter (EKF), complementary filtering, etc. These algorithms can integrate data from accelerometers, gyroscopes, and magnetometers (if available) to obtain more accurate and stable human body part posture information. Represent the estimated posture information in a suitable form, usually quaternions or Euler angles can be used.

[0063] B4: Use the IMU posture data of the left and right thighs and the left and right calves to calculate the posture difference between the thighs and the calves, that is, calculate the relative rotation matrix of the thigh posture relative to the calf posture, and convert the relative rotation matrix into Euler angles (i.e., flexion and extension angles), which correspond to the movement angles of the knee joint.

[0064] Optionally, after obtaining the knee joint angle, error correction and compensation can also be performed to reduce the angle calculation deviation caused by factors such as sensor errors, inaccurate model assumptions, and cumulative errors in the data processing process, thereby improving the accuracy and reliability of the hip joint angle calculation.

[0065] The anteversion angle is calculated in the following way: C1: Filter the output data of the lower back IMU to remove noise and unnecessary high-frequency interference. For example, using a low-pass filter can remove high-frequency noise and retain low-frequency signals related to human movement; C2: Take the lower back IMU as the origin of the global reference system, and use mathematical tools such as coordinate transformation matrices or quaternions to align the coordinate systems, ensuring that data at different positions can be compared and calculated in the same reference system; C3: Adopt data fusion algorithms to estimate the posture of each body part where the IMU is located. Common data fusion algorithms include Extended Kalman Filter (EKF), complementary filtering, etc. These algorithms can integrate data from accelerometers, gyroscopes, and magnetometers (if available) to obtain more accurate and stable human body part posture information. Represent the estimated posture information in a suitable form, usually quaternions or Euler angles can be used; C4: Convert the IMU posture data into Euler angles, and extract the rotation angle in the sagittal plane as the anteversion angle.

[0066] Optionally, after obtaining the hip inclination angle, error correction and compensation can be performed to reduce the angular calculation deviation caused by factors such as sensor errors, inaccurate model assumptions, and cumulative errors during data processing, thereby improving the accuracy and reliability of hip joint angle calculation.

[0067] The adaptive threshold knowledge base is used to store the first offset range of the upper body hip inclination angle, the second amplitude range and the second offset range of the hip joint angle, and the third amplitude range and the third offset range of the knee joint angle corresponding to different terrains; The first offset range, the second amplitude range and the second offset range of the hip joint angle, and the third amplitude range and the third offset range of the knee joint angle are obtained in the following manner: T1: For any terrain, obtain the upper body hip inclination angle, hip joint angle and knee joint angle of different people at different times, construct the upper body hip inclination angle curve, hip joint angle curve and knee joint angle curve based on each time and its corresponding upper body hip inclination angle, hip joint angle and knee joint angle, and extract the amplitudes and offsets corresponding to the upper body hip inclination angle curve, hip joint angle curve and knee joint angle curve.

[0068] In this step, the constructed hip joint angle curve, knee joint angle curve and hip inclination angle curve are actually curves formed with the end time of the previous sampling period of the current time as the starting time. Furthermore, the extracted amplitudes and offsets are also the amplitudes and offsets corresponding to each curve with the end time of the previous sampling period as the starting time.

[0069] T2: Obtain the first offset range based on the offsets corresponding to the upper body hip inclination angle curves of all people.

[0070] T3: Obtain the second amplitude range and the second offset range based on the amplitudes and offsets corresponding to the hip joint angle curves of all people.

[0071] T4: Obtain the third amplitude range and the third offset range based on the amplitudes and offsets corresponding to the knee joint angle curves of all people.

[0072] Exemplarily, obtaining the amplitude range and the offset range based on the amplitudes and offsets corresponding to the angle curves of all people includes: Compare the amplitudes of all individuals to find the minimum and maximum values. The amplitude range is the interval from the minimum amplitude to the maximum amplitude, and this range reflects the difference in the fluctuation amplitude of the angle curves of different individuals. Compare the offsets of all individuals to find the minimum and maximum values. The offset range is the interval from the minimum offset to the maximum offset, and this range reflects the difference in the offset of the angle curves of different individuals relative to the reference value. The offset range is the interval from the minimum offset to the maximum offset. Exemplarily, the data stored in the adaptive threshold knowledge base is shown in Table 2, where "amp" represents amplitude and "bias" represents offset.

[0073] Table 2 ; Comparing it with the amplitude range and offset range of the upper body forward inclination angle, hip joint angle, and knee joint angle stored in the adaptive threshold knowledge base to obtain the third terrain recognition result includes: For any terrain stored in the adaptive threshold knowledge base, if the offset of the forward inclination angle is within the first offset range, and the amplitude and offset of the first hip joint angle are respectively within the second amplitude range and the second offset range, or the amplitude and offset of the second hip joint angle are respectively within the second amplitude range and the second offset range, and the amplitude and offset of the first knee joint angle are respectively within the third amplitude range and the third offset range, or the amplitude and offset of the second knee joint angle are respectively within the third amplitude range and the third offset range, then this terrain is taken as the third terrain recognition result. If there is no matching terrain, the irregular terrain is taken as the third terrain recognition result.

[0074] As can be seen from step S1, the terrain recognition method proposed by the present invention arranges inertial measurement units on the left and right thighs, left and right calves, and the lower back of the human body, a total of 5 IMUs. When calculating the hip joint angle, it is necessary to consider the hip joint angle between the left thigh and the lower back and the hip joint angle between the right thigh and the lower back. Therefore, the hip joint angle includes: the first hip joint angle calculated based on the output data of the inertial measurement units of the left thigh and the lower back; the second hip joint angle calculated based on the output data of the inertial measurement units of the right thigh and the lower back. Similarly, when calculating the knee joint angle, it is necessary to consider the knee joint angle between the left thigh and the left calf and the knee joint angle between the right thigh and the right calf, that is, the first knee joint angle and the second knee joint angle are obtained. Therefore, a total of two hip joint angle curves, two knee joint angle curves, and one forward inclination angle curve are constructed in step S4.

[0075] When comparing, it is possible that the results corresponding to the first hip joint angle and the second hip joint angle are different, and the results corresponding to the first knee joint angle and the second knee joint angle are different. Therefore, if the anteversion angle meets the requirement, and either the first hip joint angle or the second hip joint angle meets the requirement, and either the first knee joint angle or the second knee joint angle meets the requirement, then this terrain is regarded as the third terrain recognition result.

[0076] S5: Obtain the final terrain recognition result based on the first terrain recognition result, the second terrain recognition result, and the third terrain recognition result, including: If the first terrain recognition result, the second terrain recognition result, and the third terrain recognition result are the same, then take the first terrain recognition result, the second terrain recognition result, or the third terrain recognition result as the final terrain recognition result; If the first terrain recognition result, the second terrain recognition result, and the third terrain recognition result are different, then construct a recognition result set with the terrain types that appear in each terrain recognition result, and calculate the probability of each terrain type in the recognition result set based on the weights of the first terrain recognition result, the second terrain recognition result, and the third terrain recognition result, and take the terrain type with the highest probability as the final terrain recognition result; The terrain recognition result includes: flat ground type, going up stairs type, going uphill type, going down stairs type, going downhill type, irregular terrain type.

[0077] The calculation of the probability of each terrain type in the recognition result set includes: , where c is the terrain type in the recognition result set, is the probability value when the terrain type is c, , , is the result value when the terrain type is c; when the terrain type c is consistent with the first terrain recognition result, otherwise, ; when the terrain type c is consistent with the second terrain recognition result, otherwise, ; when the terrain type c is consistent with the third terrain recognition result, otherwise, ; , , are the weights corresponding to the first terrain recognition result, the second terrain recognition result, and the third terrain recognition result, and the sum of the three weights is 1.

[0078] Exemplarily, if the current first terrain recognition result is flat ground, the second terrain recognition result is going upstairs, and the third terrain recognition result is going upstairs, then the first, second, and third terrain recognition results are different, and a recognition result set {flat ground, going upstairs} is constructed. Based on the weights of the first, second, and third terrain recognition results, the probability of each terrain type in the recognition result set is calculated, and the probability that the final terrain recognition result is flat ground is: , Then the probability that the final terrain recognition result is going upstairs is: , Compare and in size, and take the larger one as the final terrain recognition result.

[0079] For the weights of each terrain recognition result, they can be set to equal weights or can be set according to experience.

[0080] The method proposed by the present invention also supports the adaptive adjustment of weights to improve the reliability of the decision classification system. The main adjustment principles are as follows: 1. When there are errors or large fluctuations in the data used by this method, the weight of this method is reduced. For example, when the adaptive oscillator does not converge, the weight is reduced. 2. For the method based on data feature judgment, when the features of certain specific terrains are similar, the weight of this method is reduced. For example, in the terrain recognition algorithm based on joint angle feature information, the features of the knee joint are similar when going upstairs and downstairs, and there may be confusion, then the weight of this method is reduced.

[0081] Compared with the prior art, a terrain recognition method based on lower limb inertial information provided in this embodiment determines the final terrain recognition result based on the recognition results of the terrain recognition algorithm based on the adaptive oscillator, the terrain recognition algorithm based on touchdown detection, and the terrain recognition algorithm based on joint angle feature information, reducing the error of obtaining the recognition result by only using a single recognition method, and improving the accuracy and reliability of the terrain recognition result. In addition, the method proposed by the present invention can achieve terrain recognition only with the data obtained by five inertial measurement units. Compared with the prior art terrain recognition method for lower limb assistive exoskeletons using visual sensors, its hardware cost is lower, and all the inertial measurement units are arranged on the human body, making this method not affected by the external environment. The high-precision motion capture ability and efficient dynamic response ability enable this method to better apply to complex terrain recognition.

[0082] Those skilled in the art can understand that all or part of the processes of implementing the methods in the above embodiments can be completed by instructing relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium. Among them, the computer-readable storage medium is a disk, an optical disc, a read-only memory or a random access memory, etc.

[0083] As described above, the above are only the preferred specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. A terrain recognition method based on lower limb inertial information, characterized in that Comprising: Arranging inertial measurement units on the left and right thighs, left and right calves, and the lower back of the human body to be measured; Calculating the hip joint angle based on the output data of the inertial measurement units on the left and right thighs and the lower back, constructing a self-oscillating sine curve based on the hip joint angle, and obtaining a first terrain recognition result based on the self-oscillating sine curve; Performing energy spike detection based on the acceleration data output by the inertial measurement unit on the lower back, and obtaining a second terrain recognition result based on the detection result; Calculating the hip joint angle, knee joint angle, and forward tilt angle of the human body to be measured at each moment during the current movement based on the output data of the inertial measurement units on the left and right thighs, left and right calves, and the lower back, thereby constructing a hip joint angle curve, a knee joint angle curve, and a forward tilt angle curve, obtaining the amplitudes and offsets corresponding to the above curves, and comparing them with the amplitude ranges and offset ranges of the upper body forward tilt angle, hip joint angle, and knee joint angle corresponding to each terrain stored in the adaptive threshold knowledge base to obtain a third terrain recognition result; Obtaining a final terrain recognition result based on the first terrain recognition result, the second terrain recognition result, and the third terrain recognition result.

2. The method for terrain recognition based on lower limb inertial information according to claim 1, wherein The hip joint angle includes: a first hip joint angle calculated based on the output data of the inertial measurement units on the left thigh and the lower back; a second hip joint angle calculated based on the output data of the inertial measurement units on the right thigh and the lower back. The self-oscillating sine curve includes: a first self-oscillating sine curve constructed based on the first hip joint angle, and a second self-oscillating sine curve constructed based on the second hip joint angle.

3. The terrain recognition method based on lower limb inertial information according to claim 2, wherein The constructing a self-oscillating sine curve based on the hip joint angle includes: S1: Initializing the amplitude, frequency, phase, and offset of the self-oscillating sine curve; S2: Obtaining the actual hip joint angle at the current moment, and calculating a tracking error based on the actual hip joint angle and the simulated hip joint angle output by the self-oscillating sine curve; S3: If the tracking error is less than a first target threshold, then taking the current self-oscillating sine curve as the self-oscillating sine curve corresponding to the human body to be measured; if the tracking error is greater than or equal to the first target threshold, then updating the amplitude, frequency, phase, and offset of the self-oscillating sine curve at the current moment based on the tracking error, and taking the updated values as the amplitude, frequency, phase, and offset of the self-oscillating sine curve at the next moment, and returning to step S2.

4. The terrain recognition method based on lower limb inertial information according to claim 3, characterized in that, The obtaining a first terrain recognition result based on the self-oscillating sine curve includes: Comparing the amplitude, frequency, phase, and offset in the first self-oscillating sine curve with a preset first threshold classification table to obtain a first initial terrain recognition result, and comparing the amplitude, frequency, phase, and offset in the second self-oscillating sine curve with a preset first threshold classification table to obtain a second initial terrain recognition result; If the first initial terrain recognition result is consistent with the second initial terrain recognition result, then use the first initial terrain recognition result or the second initial terrain recognition result as the first terrain recognition result; if they are inconsistent, then use the irregular terrain as the first terrain recognition result; Among them, the first threshold classification table is used to store the amplitude range, frequency range, phase range, and bias range corresponding to each type of terrain.

5. The terrain recognition method based on lower limb inertial information according to claim 3, wherein Updating the amplitude, frequency, phase, and bias of the self-oscillating sine curve at the current moment based on the tracking error includes: ； wherein, is the updated phase, is the updated frequency, is the updated amplitude, is the updated offset, is the lateral learning parameter, is the longitudinal learning parameter, is the tracking error.

6. The terrain recognition method based on lower limb inertial information according to claim 1, wherein, Obtaining the final terrain recognition result based on the first terrain recognition result, the second terrain recognition result, and the third terrain recognition result includes: If the first terrain recognition result, the second terrain recognition result, and the third terrain recognition result are the same, then use the first terrain recognition result, the second terrain recognition result, or the third terrain recognition result as the final terrain recognition result; If the first terrain recognition result, the second terrain recognition result, and the third terrain recognition result are different, then construct a recognition result set with the terrain types that appear in each terrain recognition result, and calculate the probability of each terrain type in the recognition result set based on the weights of the first terrain recognition result, the second terrain recognition result, and the third terrain recognition result, and use the terrain type with the highest probability as the final terrain recognition result; The terrain recognition result includes: flat ground type, going up stairs type, going uphill type, going down stairs type, going downhill type, irregular terrain type.

7. The terrain recognition method based on lower limb inertial information according to claim 6, characterized in that, Calculating the probability of each terrain type in the recognition result set includes: ， where c is the terrain type in the set of recognition results, is the probability value when the terrain type is c, , , are the result values when the terrain type is c; when the terrain type c is consistent with the first terrain recognition result, , otherwise, ; when the terrain type c is consistent with the second terrain recognition result, , otherwise, ; when the terrain type c is consistent with the third terrain recognition result, , otherwise, ; , , are the weights corresponding to the first terrain recognition result, the second terrain recognition result, and the third terrain recognition result.

8. The terrain recognition method based on lower limb inertial information according to claim 1, characterized in that Performing energy spike detection based on the acceleration data output by the inertial measurement unit of the lower back, and obtaining the second terrain recognition result based on the detection result, including: Obtain the acceleration data, use the time period when the acceleration value in the acceleration data exceeds the preset detection threshold as the spike detection interval, and perform energy spike detection based on the spike detection interval to determine the touchdown moment; According to the output data of the inertial measurement unit at the touchdown moment and the limb lengths of the left and right thighs and the left and right calves, construct a human lower limb model at the touchdown moment, determine the centroid coordinates of the left and right feet based on the human lower limb model, and obtain the second terrain recognition result based on the centroid coordinates of the left and right feet.

9. The terrain recognition method based on lower limb inertial information according to claim 1, characterized in that The adaptive threshold knowledge base is used to store the first bias range of the upper body forward tilt angle, the second amplitude range and the second bias range of the hip joint angle, and the third amplitude range and the third bias range of the knee joint angle corresponding to different terrains; The first bias range, the second amplitude range and the second bias range of the hip joint angle, and the third amplitude range and the third bias range of the knee joint angle are obtained through the following methods: For any terrain, obtain the upper body forward inclination angles, hip joint angles, and knee joint angles of different populations at different times. Based on each time and its corresponding upper body forward inclination angle, hip joint angle, and knee joint angle, construct an upper body forward inclination angle curve, a hip joint angle curve, and a knee joint angle curve, and extract the amplitudes and offsets corresponding to the upper body forward inclination angle curve, the hip joint angle curve, and the knee joint angle curve; Obtain a first offset range based on the offsets corresponding to the upper body forward inclination angle curves of all people; Obtain a second amplitude range and a second offset range based on the amplitudes and offsets corresponding to the hip joint angle curves of all people; Obtain a third amplitude range and a third offset range based on the amplitudes and offsets corresponding to the knee joint angle curves of all people.

10. The terrain recognition method based on lower limb inertial information according to claim 9, characterized in that, Comparing it with the amplitude ranges and offset ranges of the upper body forward inclination angle, hip joint angle, and knee joint angle stored in the adaptive threshold knowledge base to obtain a third terrain recognition result, including: For any terrain stored in the adaptive threshold knowledge base, if the offset of the forward inclination angle is within the first offset range, and the amplitude and offset of the first hip joint angle are respectively within the second amplitude range and the second offset range, or the amplitude and offset of the second hip joint angle are respectively within the second amplitude range and the second offset range, and the amplitude and offset of the first knee joint angle are respectively within the third amplitude range and the third offset range, or the amplitude and offset of the second knee joint angle are respectively within the third amplitude range and the third offset range, then regard this terrain as the third terrain recognition result. If there is no matching terrain, regard the irregular terrain as the third terrain recognition result.

Images