Exoskeleton gait adaptive adjustment method based on terrain pre-perception

Abstract

This invention relates to the field of exoskeleton control technology and discloses an exoskeleton gait adaptive adjustment method based on terrain pre-sensing. The method includes: acquiring real-time data from inertial measurement units and distance sensors at the foot and lower leg, including acceleration, angular velocity, and foot-to-ground distance; fusing and generating a sequence of foot posture angle changes, a rate of change in foot-to-ground distance, hip joint angles, and knee joint angles; calculating the tilt angle and trend of the ground ahead based on the posture angle change sequence and the rate of change in distance, and predicting the slope change within future gait cycles; determining the current gait phase based on the hip and knee joint angles; calculating the joint torque adjustment required to maintain center of gravity stability based on the tilt angle, slope change, and gait phase; and generating and executing control commands at different times within the next gait cycle to drive joint motors to output auxiliary force. This invention can predict the slope ahead before the foot touches the ground and adjust the assist strategy in advance, maintaining center of gravity stability and reducing the risk of falls.

Description

Technical Field

[0001] This invention relates to the field of exoskeleton control technology, and in particular to an exoskeleton gait adaptive adjustment method based on terrain pre-sensing. Background Technology

[0002] In the field of power transmission line construction and equipment inspection and maintenance in mountainous areas, exoskeleton-assisted walking systems are gradually becoming an important technological tool. Mountainous environments are characterized by complex terrain and frequent changes in slope, requiring construction and maintenance personnel to walk with heavy loads on alternating steep and gentle slopes without paved surfaces. Traditional exoskeleton systems can provide relatively reliable assistance and support on flat surfaces or standardized staircases. However, when applied to power transmission line operations in mountainous areas, the core challenge is that the terrain slope is difficult to predict in advance, causing the output of auxiliary force to lag behind actual terrain changes. Users may experience insufficient support when going uphill and instability when going downhill, increasing the risk of falls and physical exertion.

[0003] To address this issue, some existing technologies attempt to detect the current foot-ground contact state in real time using plantar pressure sensors or inertial measurement units, thereby adjusting joint output torque upon detecting changes in slope. For example, some exoskeleton systems analyze changes in pressure distribution at the moment of foot contact to determine whether the user is on an uphill or downhill slope and adjust hip assist strategies accordingly. However, these methods are essentially reactive, meaning they can only respond after the user actually enters the slope and contacts the ground, failing to anticipate changes in slope. When a user suddenly enters an uphill section from flat ground, the system's delayed response leads to disjointed gait and untimely weight adjustment, increasing the burden on the knee and ankle joints. Furthermore, existing methods often fail to fully utilize terrain information perceived by the foot during the swing phase, resulting in a lack of foresight in judging slope changes and hindering true gait-terrain coordination. Summary of the Invention

[0004] Therefore, the technical problem to be solved by the present invention is to overcome the shortcomings of the existing exoskeleton system in mountainous and other non-flat terrains, which cannot detect changes in the slope in advance, resulting in delayed response of auxiliary force, easy loss of center of gravity and increased risk of fall when the user goes up or down slopes. The present invention provides an exoskeleton gait adaptive adjustment method based on terrain pre-sensing. It can predict the slope and its trend in advance by integrating the posture angle change sequence, distance change rate and hip and knee joint angle information during the foot swing phase, and calculate and output the joint torque adjustment amount that matches the terrain in real time in combination with the current gait phase. This enables the replanning of the assistance strategy before the foot touches the ground, thereby effectively maintaining the user's center of gravity stability, improving gait continuity and reducing the risk of fall in complex terrain.

[0005] To address the aforementioned technical problems, this invention provides an exoskeleton gait adaptive adjustment method based on terrain pre-sensing, comprising the following steps: The system acquires real-time data streams from inertial measurement units and distance sensors mounted on the exoskeleton's feet and lower legs, including acceleration, angular velocity, and the relative distance between the feet and the ground. Multi-source sensor data in the real-time data stream are fused to generate a fused sequence of foot posture angle changes, the rate of change of distance between the foot and the ground, and the hip and knee joint angles. Based on the sequence of foot posture angle changes and the rate of distance change, the tilt angle of the ground in front and its changing trend are calculated, and the slope changes in the next few gait cycles are predicted based on the changing trend; the user's current gait phase is determined based on the hip joint angle and knee joint angle. Based on the tilt angle, predicted slope changes, and current gait phase, the gait planning algorithm calculates the torque adjustments required to maintain the user's center of gravity stability at each joint. Based on the torque adjustment of each joint, a real-time control command sequence for the drive motors of the hip and knee joints of the exoskeleton is generated, including the output torque values ​​of each joint at different times in the next gait cycle. Execute a real-time control command sequence to drive the corresponding joint motor to output auxiliary force.

[0006] In one embodiment of the present invention, the fusion processing of multi-source sensor data in a real-time data stream includes: After Kalman filtering is applied to the acceleration and angular velocity to eliminate the gravitational acceleration component and zero-bias drift, the pitch angle change sequence of the foot in three-dimensional space is obtained by quaternion attitude calculation as the foot attitude angle change sequence. The rate of change of distance between the foot and the ground is obtained by performing a first-order difference operation on the relative distance; The angular velocities collected by the inertial measurement units installed at the hip and knee joints are integrated, and kinematic calculations are performed in conjunction with the geometric constraints of the exoskeleton mechanical structure to obtain the hip and knee joint angles.

[0007] In one embodiment of the present invention, the step of calculating the tilt angle of the ground in front and its changing trend based on the foot posture angle change sequence and the distance change rate specifically includes: During the foot swing phase, the foot pitch angle is recorded at multiple consecutive sampling times. The curve of pitch angle changing with time is fitted. The first derivative of the curve at the last sampling time before the foot touches the ground is taken as the tangent of the tilt angle. The tilt angle is obtained by arctangent calculation. The distance change rate is compared with a preset foot swing speed threshold. When the distance change rate jumps from a negative value to a positive value and the absolute value exceeds the threshold, it is determined that the ground in front has changed from downhill to uphill. Conversely, it is determined that the ground has changed from uphill to downhill, thus determining the trend of change.

[0008] In one embodiment of the present invention, the step of predicting the slope change over several future gait cycles based on the changing trend specifically includes: The tilt angle calculated in the current gait cycle is combined with the historical tilt angle values ​​of the last three gait cycles to form a sequence. The sign of the first difference value of the sequence is calculated. If the sign is positive twice in a row, the slope in the next gait cycle is predicted to continue to increase by a preset step length value. If the sign is negative twice in a row, the slope in the next gait cycle is predicted to continue to decrease by a preset step length value. If the sign changes alternately, the slope is predicted to remain unchanged at the current value.

[0009] In one embodiment of the present invention, the preset step length value is dynamically determined based on the maximum value of the distance change rate within the current gait cycle: the maximum value is multiplied by a fixed time coefficient and used as the step length value, and the step length value is limited to between 0.5 degrees and 3 degrees. When the predicted slope will increase, the predicted slope value in the next gait cycle is equal to the current tilt angle plus the step length; when the predicted slope will decrease, the predicted slope value in the next gait cycle is equal to the current tilt angle minus the step length.

[0010] In one embodiment of the present invention, the step of determining the user's current gait phase based on the hip joint angle and the knee joint angle specifically includes: When the hip angle begins to decrease from its maximum flexion position and the knee angle increases from its extended position to more than 30 degrees, it is determined that the current phase is the early swing phase. When the hip joint angle decreases to below zero degrees and the knee joint angle increases to its peak and then begins to decrease, it is determined that the current phase is in the middle of the swing phase. When the hip joint angle continues to extend backward and the knee joint angle decreases to near zero degrees, it is determined that the current phase is the late swing phase. When the hip joint angle changes from extension to flexion and the knee joint angle remains less than 10 degrees, it is determined that the current position is the support phase.

[0011] In one embodiment of the present invention, the step of calculating the torque adjustment of each joint required to maintain the stability of the user's center of gravity using a gait planning algorithm specifically includes: Establish a center of gravity reference coordinate system with the user's hip joint as the origin, calculate the target height of the center of gravity in the direction perpendicular to the slope based on the tilt angle, and calculate the target forward movement speed of the center of gravity along the slope based on the predicted slope change. The height and velocity deviations are obtained by subtracting the measured center of gravity height and forward velocity from the target height and forward velocity at the current gait phase. The height correction torque is obtained by multiplying the height deviation by a preset height scaling factor, and the speed correction torque is obtained by multiplying the speed deviation by a preset speed scaling factor. The sum of the height correction torque and the speed correction torque is used as the torque adjustment amount for the hip joint. In one embodiment of the present invention, if the gait phase is in the support phase, the hip joint torque adjustment is set to: ΔT hip =Kp1·ΔH+Kv1·ΔV; Where ΔH is the height deviation, ΔV is the speed deviation, Kp1 is the height proportional gain with a value range of 80–120 N·m / m, Kv1 is the speed proportional gain with a value range of 40–60 N·m / (m / s); and the knee joint torque adjustment ΔT knee For ΔT hip Half of them and in the same direction; If the gait phase is in the swing phase, then the hip joint torque adjustment is set to: ΔT hip =Kp2·ΔH; Where Kp2 is less than Kp1 and its value ranges from 20 to 40 N·m / m; the knee joint torque adjustment ΔT knee Set to zero; If the gait phase is in the late swing phase or the early stance phase transition stage, then the hip joint torque adjustment is set as follows: ΔT hip =Kp3·ΔH+Kd·(dΔH / dt); Where Kp3 ranges from 50 to 80 N·m / m, Kd is the differential gain and ranges from 10 to 20 N·m / (m / s); the knee joint torque adjustment ΔT knee Set with ΔT hip They are equal and have the same direction.

[0012] In one embodiment of the present invention, a real-time control command sequence for the drive motors of the hip and knee joints of the exoskeleton is generated based on the torque adjustment of each joint, including: Divide a gait cycle into N control beats on an average time basis; Multiply the torque adjustment of each joint by an envelope curve that varies with time to obtain the instantaneous torque value corresponding to each control beat. The envelope curve takes the value of 100% of the torque adjustment in the early and middle stages of the swing phase, linearly decreases to 0% in the later stage of the swing phase, linearly increases to 80% of the torque adjustment in the early stage of the support phase, and linearly decreases to 0% in the later stage of the support phase. The instantaneous torque value of each control cycle is added to the basic assist torque value of the corresponding joint to obtain the final output torque value of the control cycle, and then arranged in chronological order to form a control command sequence.

[0013] In one embodiment of the present invention, executing a real-time control command sequence to drive the corresponding joint motor to output auxiliary force includes: At the start of the next gait cycle, current commands are sent to the drive motors of the hip and knee joints according to the final output torque value of the first control beat in the control command sequence. At the end of each control cycle, the actual angle of the current joint is read. If the deviation between the actual angle and the expected angle exceeds the preset threshold of 5°, the final output torque value of the control cycle is increased by 20% and used as the instruction value for the next control cycle. Otherwise, the output continues according to the next control cycle value in the original sequence. Repeat the above process until all control cycles in the control command sequence have been executed.

[0014] The technical solution of the present invention has the following advantages compared with the prior art: The terrain-preception-based exoskeleton gait adaptive adjustment method described in this invention fuses sensor data from the exoskeleton's feet and lower legs to anticipate the tilt angle and its changing trends of the ground ahead, and predict future slope conditions. Combined with the user's real-time gait phase, the method proactively calculates and adjusts the auxiliary torques of each joint of the exoskeleton, thereby achieving adaptive gait adjustment. This solves the problems of traditional exoskeleton systems, such as lagging auxiliary force output, lack of gait smoothness, and instability in mountainous terrain, reducing the risk of falls and energy consumption, and improving the adaptability and safety of exoskeletons in mountainous power transmission line operation scenarios. Attached Figure Description

[0015] To make the content of this invention easier to understand, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings, wherein: Figure 1 This is a flowchart of the steps of the exoskeleton gait adaptive adjustment method based on terrain pre-sensing of the present invention; Figure 2 This is a flowchart of the steps for fusing multi-source sensor data according to the present invention; Figure 3 This is a flowchart illustrating the steps of calculating the tilt angle of the ground in front and its changing trend in this invention. Figure 4 This is a flowchart illustrating the steps of the present invention to predict slope changes over several future gait cycles based on changing trends. Figure 5This is a flowchart of the steps in this invention to determine the current gait phase of a user; Figure 6 This is a flowchart illustrating the steps of calculating the torque adjustments required to maintain the user's center of gravity in this invention. Figure 7 This is a flowchart illustrating the steps of generating a real-time control command sequence and driving the corresponding joint motor to output auxiliary force according to the present invention. Detailed Implementation

[0016] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, so that those skilled in the art can better understand and implement the present invention. However, the embodiments described are not intended to limit the present invention.

[0017] Reference Figure 1 As shown, this invention proposes an exoskeleton gait adaptive adjustment method based on terrain pre-awareness. First, inertial measurement units and distance sensors installed on the exoskeleton's foot and lower leg collect acceleration, angular velocity, and the relative distance between the foot and the ground in real time. These sensors begin operating during the foot swing, acquiring information on the foot's posture and distance changes relative to the ground before the foot even contacts the ground. Based on this, the multi-source sensor data is fused to generate a sequence of foot posture angle changes, a rate of change in the distance between the foot and the ground, and hip and knee joint angles. The posture angle change sequence and distance change rate reflect the foot's tilting tendency and approach velocity relative to the ground during the swing phase, while the hip and knee joint angles characterize the user's current gait phase.

[0018] Using the fused data, the tilt angle and its trend of change of the ground ahead are further calculated based on the posture angle change sequence and the distance change rate. The core principle of this calculation process is that when the foot swings forward in the air, there is a geometrical mapping relationship between the trend of the foot's posture angle change and the slope of the ground ahead. The distance change rate reflects the rate of change of the foot's height relative to the ground. Combining these two factors allows for the prediction of the steepness and direction of change of the slope ahead before the foot touches the ground. Simultaneously, the system determines whether the user is currently in the support phase or the swing phase based on the hip and knee joint angles, thus determining the most appropriate time for intervention and adjustment. Based on this, the system calculates the amount of torque adjustment required to maintain the user's center of gravity stability using a gait planning algorithm, based on the predicted tilt angle, the predicted slope changes over several future gait cycles, and the current gait phase. This process is equivalent to completing the slope identification and assist strategy replanning before the foot touches the ground, achieving a fundamental shift from reactive response to proactive prediction.

[0019] Finally, based on the calculated joint torque adjustment, a real-time control command sequence is generated for the exoskeleton's hip and knee joint drive motors. This command sequence precisely includes the output torque values ​​of each joint at different times in the next gait cycle, and drives the corresponding joint motors to output auxiliary force. Because the control commands are generated before the feet touch the ground, when the user actually enters the ramp, the exoskeleton can synchronously provide support force that matches the slope and direction, avoiding gait interruption or center of gravity imbalance caused by response lag.

[0020] Based on the above technical solution, this method can achieve the following beneficial effects: First, by introducing a terrain pre-sensing mechanism into the exoskeleton control process, the system can identify the slope ahead during the foot swing phase, solving the lag problem caused by the existing technology's response only after ground contact, and significantly improving gait continuity in complex terrain. Second, by integrating multi-dimensional information such as posture angle change sequence, distance change rate, and hip and knee joint angles, it can not only determine the angle of the slope to be entered, but also predict the slope change trend in the next few gait cycles, making the exoskeleton control strategy forward-looking and continuous, avoiding the shock and maladaptation caused by sudden slope changes. Third, by combining the prediction results with the real-time gait phase, the calculation of torque adjustment is more targeted—stable support needs to be provided during the support phase, and the foot lift-off height needs to be controlled during the swing phase to avoid collisions. This time-segmented fine control reduces the user's physical exertion and the risk of falls.

[0021] Overall, this invention achieves adaptive adjustment of exoskeleton gait in complex mountainous terrain by using only low-cost sensors in the feet and lower legs without relying on high-precision maps or visual sensors. It is particularly suitable for assistive walking needs in typical non-flat and unstructured environments such as power transmission line construction and inspection in mountainous areas.

[0022] In the implementation process, directly fusing raw sensor data often fails to effectively eliminate interference such as sensor noise, drift, and gravity components. This may result in insufficient accuracy and reliability of key parameters such as foot posture angle, distance change rate, and joint angle, thereby affecting the accuracy and stability of subsequent terrain perception and gait adjustment.

[0023] Reference Figure 2As shown, this application further proposes a specific method for fusing multi-source sensor data in a real-time data stream, including: applying Kalman filtering to acceleration and angular velocity, which can employ nonlinear filtering algorithms such as Extended Kalman Filter (EKF) or Unscented Kalman Filter (UKF). The filter's state vector typically includes the foot's attitude, angular velocity deviation, and possible acceleration deviation. By using the measurements from the accelerometer and gyroscope as the observation vector input, the Kalman filter can effectively fuse the data from these two sensors, separating the gravitational acceleration component in dynamic motion and estimating and compensating for the gyroscope's zero-bias drift, thereby obtaining a more accurate and stable estimation of the foot's motion state. Based on this, a quaternion attitude calculation algorithm, such as one based on complementary filtering or a more complex fusion algorithm, is used to convert the filtered angular velocity and acceleration information into the foot's attitude in three-dimensional space, and the pitch angle change sequence is extracted from it.

[0024] =The first-order difference operation on relative distance aims to obtain the rate of change of distance between the foot and the ground over time. This operation is usually achieved by calculating the difference in relative distance between two consecutive sampling times and then dividing by the sampling time interval. For example, if the relative distance at time t is D(t) and the relative distance at time t-1 is D(t-1), then the rate of change of distance can be approximated as (D(t) - D(t-1)) / (t - (t-1)). To further improve the smoothness and noise resistance of the rate of change of distance, the original relative distance data can be low-pass filtered before performing the difference operation, or the difference result can be smoothed.

[0025] Inertial measurement units (IMUs) installed at the hip and knee joints can acquire angular velocity information of the joints in real time. Angular displacement of the joints can be obtained by integrating this angular velocity data over time. However, simple integration is prone to cumulative errors and drift. Therefore, this application incorporates the geometric constraints of the exoskeleton's mechanical structure for kinematic calculation. This means constructing a kinematic model of the exoskeleton using geometric information such as the known lengths of each link, the connection method of the joints, and the range of joint motion. By using the angular displacement obtained from IMU integration as input and combining it with the kinematic model, integration errors can be corrected, and the absolute angles of the hip and knee joints can be accurately calculated. For example, the relative angle between the thigh and tibia segments, measured by the IMU, can be calculated through vector operations or rotation matrix transformations; similarly, the hip joint angle can be calculated by the relationship between the thigh segment's posture and the posture of the torso or ground reference frame.

[0026] In some embodiments described above in this application, an exoskeleton gait adaptive adjustment method based on terrain pre-sensing is proposed. Its core lies in calculating the tilt angle and its changing trend of the ground ahead based on the foot posture angle change sequence and distance change rate, referring to... Figure 3 As shown, the specific steps include: during the foot swing phase, recording the foot pitch angle at multiple consecutive sampling moments, fitting a curve showing the pitch angle changing with time, using the first derivative of the curve at the last sampling moment before the foot touches the ground as the tangent of the tilt angle, and obtaining the tilt angle through arctangent calculation; comparing the distance change rate with a preset foot swing speed threshold, and determining the change trend when the distance change rate jumps from negative to positive and the absolute value exceeds the threshold, indicating that the ground in front has changed from downhill to uphill, and vice versa.

[0027] Specifically, this technical solution aims to accurately acquire the actual tilt angle of the ground in front of the exoskeleton foot at the instant it is about to contact the ground. During the foot swing phase, the foot is in the air, and its posture changes reflect the terrain features of the ground it is about to contact. By continuously recording the foot's pitch angle (acquired by an inertial measurement unit installed on the foot), a time-series data can be generated. Curve fitting of this time-series data (e.g., using polynomial fitting, spline interpolation, etc.) can effectively smooth sensor noise and obtain a continuous function of the foot's pitch angle changing with time. At the last sampling moment before the foot touches the ground, the first derivative of this fitted curve is calculated. This derivative represents the angular velocity of the foot at the instant it is about to touch the ground, and its tangent value directly reflects the degree of tilt of the foot relative to the horizontal plane. Therefore, the tilt angle of the ground in front can be obtained by arctangent calculation. This method avoids the lag that may be caused by measuring after the foot touches the ground, improving the real-time performance and accuracy of tilt angle measurement.

[0028] Simultaneously, this technical solution is used to dynamically identify the slope change trend of the terrain ahead, that is, to determine whether the current terrain is transitioning from downhill to uphill or from uphill to downhill. The distance change rate (measured by a distance sensor and obtained through first-order difference calculation) reflects the speed at which the distance between the foot and the ground changes over time. When the foot is swinging, if the distance change rate jumps from a negative value (foot moving closer to the ground) to a positive value (foot moving away from the ground), and the absolute value of this jump exceeds a preset foot swing speed threshold, it indicates that the terrain beneath the foot is changing from a downward trend to an upward trend, that is, from downhill to uphill. Conversely, if the distance change rate jumps from a positive value to a negative value and the absolute value exceeds the threshold, it is determined to be a transition from uphill to downhill. This preset foot swing speed threshold is used to distinguish between normal gait swing and actual terrain changes, avoiding misjudgment. In this way, the system can perceive the dynamic changes of the terrain in real time and accurately, providing forward-looking information for subsequent gait adjustments.

[0029] Furthermore, referring to Figure 4 As shown, this application further proposes a method for predicting future slope changes. This application constructs a sequence by combining the tilt angle calculated within the current gait cycle with historical tilt angle values ​​from the last three gait cycles. This sequence is designed to provide a time window, allowing the system to observe continuous slope changes over multiple gait cycles, rather than relying solely on instantaneous measurements. By collecting and maintaining a historical sequence containing tilt angles from the current and the past three gait cycles, sufficient data can be provided for subsequent trend analysis, thereby more accurately capturing the dynamic characteristics of terrain changes.

[0030] Based on this, this application calculates the sign of the first-order difference value of the sequence. The first-order difference value reflects the change in tilt angle between adjacent gait cycles, and its sign directly indicates whether the slope is increasing, decreasing, or remaining constant. For example, if the current tilt angle is greater than the tilt angle of the previous gait cycle, the difference value is positive, indicating that the slope is increasing. By analyzing the signs of these difference values, the system can identify the specific direction of the slope change.

[0031] Furthermore, this application predicts future slope based on the continuity of the difference value's sign. If the sign is positive twice consecutively, it means the slope has been continuously increasing in the last two gait cycles, indicating the user is continuously going uphill. In this case, the system predicts the slope will continue to increase by a preset step size in the next gait cycle. This preset step size represents the typical magnitude of slope change, and its setting should ensure the reasonableness and stability of the prediction. If the sign is negative twice consecutively, it indicates the slope has been continuously decreasing in the last two gait cycles, indicating the user is continuously going downhill. In this case, the system predicts the slope will continue to decrease by a preset step size in the next gait cycle. This prediction mechanism based on continuous trends can effectively cope with continuous terrain changes. If the sign changes alternately, such as one positive and one negative, or a zero value appears, it indicates that the slope change does not have a clear continuous trend and may be in flat or undulating terrain. In this case, the system predicts the slope will remain unchanged to avoid over-prediction or misjudgment and ensure the robustness of the prediction.

[0032] Specifically, in its implementation, if a fixed step size is used for slope prediction, the prediction result may deviate significantly from the actual terrain change, especially when the terrain changes are complex or fast. This will affect the accuracy and timeliness of the exoskeleton gait adjustment.

[0033] To address the aforementioned issues, this application proposes that the preset step length value be dynamically determined based on the maximum value of the distance change rate within the current gait cycle: the maximum value is multiplied by a fixed time coefficient to obtain the step length value, which is limited to between 0.5 degrees and 3 degrees; when the predicted slope will increase, the predicted slope value within the next gait cycle is equal to the current tilt angle plus the step length value; when the predicted slope will decrease, the predicted slope value within the next gait cycle is equal to the current tilt angle minus the step length value.

[0034] Specifically, the preset step length value is no longer a fixed value, but is dynamically determined based on the maximum value of the distance change rate within the current gait cycle. The distance change rate reflects the relative vertical velocity of the foot to the ground, and its maximum value effectively indicates the severity of terrain changes within the current gait cycle. For example, when the foot rapidly approaches or moves away from the ground, the absolute value of the distance change rate will be larger, which usually means a more significant slope change. The system continuously monitors and records the maximum absolute value of the distance change rate within each gait cycle, using this as the basis for dynamically adjusting the step length value. The maximum value of the distance change rate obtained above is multiplied by a fixed time coefficient to obtain the step length value. This time coefficient is an empirical value that can be calibrated through extensive experimental data or simulation analysis, aiming to convert the physical quantity of the distance change rate into a suitable slope change. For example, if the maximum distance change rate is 0.1 m / s and the time coefficient is 20 degrees / (m / s), the calculated step length value is 2 degrees. The selection of this coefficient needs to comprehensively consider the kinematic characteristics of the exoskeleton, the user's gait habits, and the requirements for the exoskeleton's response sensitivity. To ensure the stability and safety of the predictions, the calculated step size is limited to between 0.5 and 3 degrees. This means that even in extreme cases, where sensor noise or abnormal data causes the calculated step size to be too large or too small, the final value will still be constrained within this preset reasonable range. This limitation effectively prevents the exoskeleton from making excessive or insufficient adjustments due to large prediction deviations, thereby enhancing the robustness of the system. When the predicted slope is expected to increase according to the aforementioned method, the predicted slope value in the next gait cycle equals the current tilt angle plus the dynamically determined step size. Conversely, when the predicted slope is expected to decrease, the predicted slope value in the next gait cycle equals the current tilt angle minus the dynamically determined step size. This approach allows slope prediction to more accurately reflect the actual trend and magnitude of terrain changes.

[0035] In some embodiments described above in this application, the current gait phase of a user is determined based on the hip and knee joint angles. However, in practical applications, accurately identifying the user's inconsistent gait phases in complex terrain to ensure the accuracy and timeliness of subsequent torque adjustments and control commands is a key technical problem that needs to be solved.

[0036] Reference Figure 5 As shown, this application further proposes a step for determining the user's current gait phase based on the hip joint angle and knee joint angle, specifically including: when the hip joint angle begins to decrease from its maximum forward flexion position and the knee joint angle increases from its extended position to more than 30 degrees, it is determined that the user is currently in the early swing phase; when the hip joint angle decreases to below zero degrees and the knee joint angle increases to its peak and then begins to decrease, it is determined that the user is currently in the middle swing phase; when the hip joint angle continues to extend backward and the knee joint angle decreases to near zero degrees, it is determined that the user is currently in the late swing phase; when the hip joint angle changes from extension to forward flexion and the knee joint angle remains less than 10 degrees, it is determined that the user is currently in the stance phase.

[0037] Hip and knee angles are key biomechanical parameters reflecting the movement state of the lower limbs. The hip angle typically refers to the angle between the thigh and the trunk, with positive values ​​for flexion and negative values ​​for extension. The knee angle typically refers to the angle between the lower leg and the thigh, with zero degrees or near zero degrees in extension and positive values ​​in flexion. The dynamic patterns of these joint angles are closely related to the various phases of the gait cycle and are a direct and important basis for determining gait phase.

[0038] When the hip angle begins to decrease from its maximum flexion position, while the knee angle increases from extension and exceeds 30 degrees, this indicates that the user's foot has left the ground and begun to swing forward. The knee begins to flex to achieve foot lift-off clearance, at which point the system determines that the user is currently in the early swing phase. Maximum hip flexion typically occurs at the end of the stance phase or the beginning of the swing phase, marking the transition from stance to swing. Knee flexion exceeding 30 degrees is a key characteristic of foot lift-off at the beginning of the swing phase.

[0039] When the hip angle further decreases to below zero degrees (i.e., extension begins), and the knee angle, after increasing to its peak, begins to decrease, this indicates that the user's thigh has swung forward and begun to swing backward, while the lower leg begins to extend under gravity, preparing for the upcoming ground contact. At this point, the system determines that the user is currently in the middle of the swing phase. Hip extension and knee extension are typical characteristics of the middle of the swing phase.

[0040] As the hip joint angle continues to extend backward while the knee joint angle continues to decrease and approaches zero degrees, this indicates that the user's foot is about to touch the ground. Continued hip extension stabilizes the torso relative to the thigh, while knee extension provides a relatively stable ground-touching posture for the foot. At this point, the system determines that the user is in the late swing phase. The knee joint approaching zero degrees (extended) is an important preparation before ground contact.

[0041] When the hip joint angle changes from extension to flexion, and the knee joint angle remains less than 10 degrees, it indicates that the user's feet have touched the ground and are bearing the body weight. The hip joint's shift from extension to flexion reflects a forward shift of the body's center of gravity, while keeping the knee joint relatively straight (less than 10 degrees) ensures the lower limb's support stiffness to effectively support body weight and propel the body forward. At this point, the system determines that the user is currently in the support phase.

[0042] Based on the tilt angle, predicted slope changes, and current gait phase, refer to Figure 6 As shown, this application further proposes a method for calculating the torque adjustment of each joint required to maintain the stability of the user's center of gravity using a gait planning algorithm. The method includes: establishing a center of gravity reference coordinate system with the user's hip joint as the origin; calculating the target height of the center of gravity in the direction perpendicular to the slope based on the tilt angle; calculating the target forward movement speed of the center of gravity along the slope based on the predicted slope change; calculating the difference between the measured center of gravity height and forward movement speed in the current gait phase and the target height and forward movement speed, respectively, to obtain the height deviation and speed deviation; multiplying the height deviation by a preset height ratio coefficient to obtain the height correction torque; multiplying the speed deviation by a preset speed ratio coefficient to obtain the speed correction torque; and using the sum of the height correction torque and the speed correction torque as the torque adjustment of the hip joint.

[0043] To accurately describe and control the user's center of gravity movement, this method first establishes a center of gravity reference coordinate system with the user's hip joint as the origin. As the pivot point for lower limb movement, the hip joint provides a relatively stable reference benchmark closely related to the user's body movements. This coordinate system is typically constructed in real-time within the control system using data from an inertial measurement unit and joint encoder integrated into the exoskeleton, combined with the exoskeleton's geometric model and the user's body shape parameters. Maintaining center of gravity stability is crucial when walking on a slope. This method calculates the target height of the center of gravity perpendicular to the slope based on the inclination angle of the ground ahead. This target height is not a fixed value but rather an ideal vertical position of the center of gravity that is dynamically adjusted according to the terrain's inclination. For example, when climbing a slope, the target center of gravity height may be slightly lowered to increase stability; while when descending, a relatively high center of gravity may be needed to maintain gait smoothness. This is typically achieved through a preset biomechanical model or empirical function, ensuring that the center of gravity is always in a safe and efficient vertical position. Furthermore, this method also calculates the target forward velocity of the center of gravity along the slope direction based on predicted slope changes. Predicted changes in slope (e.g., whether the slope will increase, decrease, or remain unchanged) provide the system with forward-looking information. When an increase in slope is predicted, the system may appropriately reduce the target's forward movement speed to allow the user more time to adapt to the terrain change; conversely, when a decrease in slope is predicted, the target's forward movement speed may be appropriately increased to maintain gait consistency. This proactive speed adjustment helps to smoothly transition to new terrain, avoiding shocks or imbalances caused by speed mismatches.

[0044] To achieve precise feedback control, this method compares the measured center-of-gravity height and forward velocity in the current gait phase with the calculated target height and forward velocity, respectively, and calculates the difference between the two to obtain the height deviation and velocity deviation. The measured center-of-gravity height and forward velocity can be estimated in real time using data collected by inertial measurement units and distance sensors in the exoskeleton's foot and lower leg, as well as inertial measurement units at the hip and knee joints, combined with the exoskeleton's kinematic and dynamic models. These deviations quantify the degree of deviation between the current center-of-gravity state and the desired state, serving as the direct basis for subsequent torque adjustments. After obtaining the height and velocity deviations, this method multiplies the height deviation by a preset height scaling factor to obtain the height correction torque, and multiplies the velocity deviation by a preset velocity scaling factor to obtain the velocity correction torque. These scaling factors are gain parameters obtained through system calibration and optimization, and they determine the system's response strength to center-of-gravity deviations. The height correction torque aims to adjust the user's vertical center-of-gravity position through hip joint actuation, while the velocity correction torque is used to adjust the forward velocity of the center of gravity along the slope direction. Finally, the height correction torque and the velocity correction torque are superimposed, and their sum is used as the torque adjustment amount for the hip joint. This torque adjustment strategy, which comprehensively considers the deviations in height and velocity, can fully address the impact of terrain changes on center of gravity stability, ensuring that the exoskeleton can provide precise and coordinated auxiliary forces.

[0045] Specifically, this application employs different control strategies to calculate the torque adjustment of the hip and knee joints based on different gait phases. This phased control strategy can more precisely match the assistance needs of the exoskeleton at different gait phases.

[0046] When the gait phase is in the support phase, the user's feet are in contact with the ground and bear their weight. Maintaining center of gravity stability and providing effective support are crucial at this time. Hip joint torque adjustment ΔT hip The value is set as the sum of the proportional terms for height deviation ΔH (Kp1·ΔH) and velocity deviation ΔV (Kv1·ΔV). The height proportional gain Kp1 ranges from 80 to 120 N·m / m and is used to adjust the exoskeleton's response to vertical center of gravity height deviation. A larger Kp1 value ensures sufficient vertical support force during the stance phase to quickly correct the center of gravity height and prevent falls. The velocity proportional gain Kv1 ranges from 40 to 60 N·m / (m / s) and is used to adjust the exoskeleton's response to horizontal center of gravity forward movement velocity deviation, ensuring effective control of forward movement and maintaining gait stability during the stance phase. Knee joint torque adjustment ΔT knee Set as hip joint torque adjustment ΔT hipThe fact that the knee and hip joints work together to provide support and propulsion during the support phase indicates that the knee joint's auxiliary strength is slightly lower than that of the hip joint to accommodate its biomechanical function during the support phase.

[0047] When the gait phase is in the swing phase, the user's feet are off the ground, and the primary task is to complete the leg swing and maintain the foot-off-ground clearance. At this time, the exoskeleton should not provide excessive support to avoid hindering the natural swing. Therefore, the hip joint torque adjustment ΔT... hip The proportional term (Kp2·ΔH) is set to be related only to the height deviation ΔH. The height proportional gain Kp2 ranges from 20 to 40 N·m / m, and Kp2 is less than Kp1. A smaller Kp2 value means that during the swing phase, the exoskeleton's correction effect on the center of gravity height deviation is weakened, avoiding unnecessary vertical support forces and allowing the legs to swing more freely. Knee joint torque adjustment ΔT knee Setting it to zero further emphasizes that the knee joint does not provide active vertical support during the swing phase, but instead focuses on completing the swing motion.

[0048] When the gait phase is in the transition phase from the late swing phase to the early stance phase, this is a critical period when the foot is about to strike the ground or has just struck it, requiring a smooth transition and preparation to withstand impact. At this time, the hip joint torque adjustment ΔT hip The gain is set as the sum of the proportional term (Kp3·ΔH) of the height deviation ΔH and the differential term of the height deviation with respect to time (Kd·(dΔH / dt)). Kp3 ranges from 50 to 80 N·m / m, falling between Kp1 and Kp2, reflecting the support force required during the transition phase. The differential gain Kd ranges from 10 to 20 N·m / (m / s). Introducing the differential term dampens the exoskeleton's response, enabling it to predict and suppress rapid changes in center of gravity height, thus providing smoother cushioning and more stable support upon foot strike. Knee joint torque adjustment ΔT knee Set to the hip joint torque adjustment ΔT hip The fact that they are equal in strength and in the same direction indicates that during this critical transition phase, the hip and knee joints need to provide synergistic assistance of equal strength to ensure the overall stiffness and stability of the limb at the moment of ground contact, effectively absorb the impact, and prepare for the subsequent support phase.

[0049] In some embodiments described above, this application proposes calculating the joint torque adjustments required to maintain user center of gravity stability based on terrain pre-sensing, and generating real-time control command sequences for the exoskeleton's hip and knee joint drive motors. However, in practical applications, effectively converting these calculated joint torque adjustments into specific output torque values ​​for the drive motors at different gait phases to ensure the smoothness, adaptability, and effectiveness of the assistive force, while avoiding unnecessary impact or energy waste, is a problem requiring meticulous handling. Simply outputting the adjustment amounts directly may result in poor assistive effects or a decreased user experience.

[0050] Reference Figure 7 As shown, this application further proposes a method for generating real-time control command sequences for the drive motors of the hip and knee joints of the exoskeleton based on the torque adjustment of each joint. This method first divides a gait cycle into N control beats on an average time basis. Here, N is a preset integer, such as 100 or 200, and its specific value depends on the sampling frequency of the control system and the required control precision. By discretizing the continuous gait cycle into a series of equal-length time intervals, fine-grained control of the exoskeleton's auxiliary torque output can be achieved, thereby better adapting to the dynamic changes in human gait.

[0051] Subsequently, the torque adjustment values ​​for each joint are multiplied by an envelope curve that varies over time to obtain the instantaneous torque value corresponding to each control beat. This envelope curve is pre-designed or generated in real time based on the biomechanical characteristics of human gait and the needs of exoskeleton assistance. Its function is to dynamically weight the calculated torque adjustment values, making the output of the auxiliary torque more consistent with the actual needs of the gait phase. Specifically, in the early and middle stages of the swing phase, the legs usually require a large auxiliary force to complete the swing, so the envelope curve is set to 100% of the torque adjustment value in this stage to ensure sufficient assistance. In the late stage of the swing phase, as the legs are about to touch the ground, the auxiliary force should gradually decrease to avoid impact upon landing, so the envelope curve linearly decreases to 0%. Entering the early stage of the support phase, the legs need to establish stable support as soon as they touch the ground, and the auxiliary force should gradually increase, but considering the participation of the body's own muscles, the envelope curve linearly rises to 80% of the torque adjustment value. In the late stage of the support phase, as the legs prepare to leave the ground, the auxiliary force should gradually decrease again to smoothly transition to the swing phase, so the envelope curve linearly decreases to 0%. This piecewise linear envelope curve design allows the output of the auxiliary torque to smoothly follow changes in gait phase, avoiding sudden changes in torque.

[0052] Finally, the instantaneous torque value of each control beat is added to the basic assist torque value of the corresponding joint to obtain the final output torque value of the control beat, and arranged in chronological order to form a control command sequence. The basic assist torque value is a preset assist torque provided by the exoskeleton to maintain the user's normal gait on flat ground or in the absence of special terrain conditions, and is usually obtained through analysis or experimental calibration of a large amount of normal gait data. Superimposing this value with the instantaneous torque value adjusted by the envelope curve aims to ensure that the exoskeleton maintains stable basic assist function while providing terrain-adaptive adjustments. These final output torque values ​​are arranged in chronological order according to their corresponding control beats to form a complete command sequence, which is sent to the drive motor to achieve real-time assist control.

[0053] Through the above technical solution, this application can effectively convert the calculated joint torque adjustment amount into the specific output torque value of the drive motor under different gait phases, thereby ensuring the stability, adaptability, and effectiveness of the auxiliary force. By finely dividing a gait cycle into multiple control beats and introducing a time-varying envelope curve to weight the joint torque adjustment amount, this application can dynamically adjust the output intensity of the auxiliary torque according to different gait phases. Specifically, in the early and middle stages of the swing phase, which require greater assistance, the exoskeleton provides the full torque adjustment amount; in the late swing phase and late support phase, which require a smooth transition, the auxiliary torque decreases linearly to avoid unnecessary impact; and in the early support phase, which requires establishing stable support, the auxiliary torque increases moderately, balancing the auxiliary effect with the user's own muscle participation. In addition, by superimposing the instantaneous torque value adjusted by the envelope curve with the basic assist torque value, the exoskeleton can maintain a stable basic assistance while providing terrain-adaptive adjustments. This makes the generated control command sequence smoother and more natural, significantly improving the gait adaptability of the exoskeleton and the user's wearing comfort, and effectively avoiding instability or discomfort that may be caused by sudden torque changes.

[0054] Furthermore, based on the above embodiments, this application introduces a real-time feedback and adaptive adjustment mechanism in addition to generating a real-time control command sequence, specifically including: At the start of the next gait cycle, current commands are sent to the drive motors of the hip and knee joints according to the final output torque value of the first control beat in the control command sequence. This step aims to ensure that the exoskeleton can immediately provide assistance based on the pre-planned torque value of the first control beat at the start of a new gait cycle. By sending current commands to the drive motors, the output torque of the motors can be directly controlled, thereby achieving precise application of assistive force. This lays the foundation for assistive control throughout the gait cycle, ensuring the timeliness and continuity of assistance.

[0055] At the end of each control beat, the actual angle of the current joint is read. If the deviation between the actual angle and the desired angle exceeds a preset threshold of 5°, the final output torque value of the control beat is increased by 20% and used as the command value for the next control beat. Otherwise, the output continues according to the next control beat value in the original sequence. This step introduces a real-time feedback mechanism. Angle sensors installed at the hip and knee joints can accurately acquire the actual angle of the joints. The desired angle is usually preset by a gait planning algorithm based on terrain and the user's gait phase, representing the ideal joint movement trajectory. Comparing the actual angle with the desired angle quantifies the difference between the exoskeleton's assistive effect and the user's actual movement. The preset threshold of 5° defines an acceptable deviation range; when the deviation exceeds this range, it indicates that adjustments are needed to correct the exoskeleton's assistive behavior. When the deviation between the actual joint angle and the desired angle exceeds the preset threshold, the system determines that the current assistive force may be insufficient or inappropriate, requiring intervention. By increasing the output torque value of the next control beat by 20%, the system can respond quickly and provide greater assistive force to help the user pull the joint movement back to the desired trajectory. If the deviation is within an acceptable range, it indicates that the current assistance effect is good and no additional adjustments are needed. Continue to execute according to the preset control command sequence to maintain the stability and predictability of the system. This feedback-based adjustment strategy allows the exoskeleton to dynamically adapt to the user's actual movement state.

[0056] The above process is repeated until all control beats in the control command sequence have been executed. This step ensures that the aforementioned real-time feedback and adaptive adjustment mechanism can be maintained throughout the entire gait cycle. By performing angle readings, deviation judgments, and torque adjustments at the end of each control beat, the exoskeleton can continuously monitor the user's movement and make fine adjustments as needed, thereby providing continuous and stable assistance. This repetitive execution mechanism ensures that the exoskeleton maintains a high degree of synchronization and adaptability with the user's movement throughout the entire gait cycle.

[0057] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A terrain-pre-sensing-based exoskeleton gait adaptive adjustment method, characterized in that, Includes the following steps: The system acquires real-time data streams from inertial measurement units and distance sensors mounted on the exoskeleton's feet and lower legs, including acceleration, angular velocity, and the relative distance between the feet and the ground. Multi-source sensor data in the real-time data stream are fused to generate a fused sequence of foot posture angle changes, the rate of change of distance between the foot and the ground, and the hip and knee joint angles. Based on the foot posture angle change sequence and distance change rate, the tilt angle of the ground in front and its changing trend are calculated, including: during the foot swing phase, recording the foot pitch angle at multiple consecutive sampling moments, fitting a curve of pitch angle changing with time, using the first derivative of the curve at the last sampling moment before foot contact with the ground as the tangent of the tilt angle, and obtaining the tilt angle through arctangent calculation; comparing the distance change rate with a preset foot swing speed threshold, when the distance change rate jumps from negative to positive and the absolute value exceeds the threshold, it is determined that the ground in front changes from downhill to uphill, and vice versa, thus determining the changing trend. The system analyzes the slope and predicts its changes over several gait cycles based on the trend. This includes: combining the tilt angle calculated in the current gait cycle with the historical tilt angle values ​​from the last three gait cycles to form a sequence; calculating the sign of the first difference of the sequence; if the sign is positive twice consecutively, the slope is predicted to increase by a preset step length; if the sign is negative twice consecutively, the slope is predicted to decrease by a preset step length; if the sign changes alternately, the slope is predicted to remain unchanged. The system also determines the user's current gait phase based on the hip and knee angles. Based on the tilt angle, predicted slope changes, and current gait phase, a gait planning algorithm calculates the torque adjustments required to maintain the user's center of gravity stability for each joint. This includes: establishing a center of gravity reference coordinate system with the user's hip joint as the origin; calculating the target height of the center of gravity in the direction perpendicular to the slope based on the tilt angle; calculating the target forward velocity of the center of gravity along the slope based on the predicted slope changes; subtracting the measured center of gravity height and forward velocity from the target height and forward velocity in the current gait phase to obtain height deviation and velocity deviation, respectively; multiplying the height deviation by a preset height scaling factor to obtain a height correction torque; multiplying the velocity deviation by a preset velocity scaling factor to obtain a velocity correction torque; and using the sum of the height correction torque and the velocity correction torque as the torque adjustment for the hip joint. Based on the torque adjustment of each joint, a real-time control command sequence for the drive motors of the hip and knee joints of the exoskeleton is generated, including the output torque values ​​of each joint at different times in the next gait cycle. Execute a real-time control command sequence to drive the corresponding joint motor to output auxiliary force.

2. The exoskeleton gait adaptive adjustment method based on terrain pre-sensing according to claim 1, characterized in that: The process of fusing multi-source sensor data in a real-time data stream includes: After Kalman filtering is applied to the acceleration and angular velocity to eliminate the gravitational acceleration component and zero-bias drift, the pitch angle change sequence of the foot in three-dimensional space is obtained by quaternion attitude calculation as the foot attitude angle change sequence. The rate of change of distance between the foot and the ground is obtained by performing a first-order difference operation on the relative distance; The angular velocities collected by the inertial measurement units installed at the hip and knee joints are integrated, and kinematic calculations are performed in conjunction with the geometric constraints of the exoskeleton mechanical structure to obtain the hip and knee joint angles.

3. The exoskeleton gait adaptive adjustment method based on terrain pre-sensing according to claim 1, characterized in that: The preset step size is dynamically determined based on the maximum value of the distance change rate within the current gait cycle: the maximum value is multiplied by a fixed time coefficient to obtain the step size, and the step size is limited to between 0.5 degrees and 3 degrees. When the predicted slope will increase, the predicted slope value in the next gait cycle is equal to the current tilt angle plus the step length. When the predicted slope will decrease, the predicted slope value in the next gait cycle is equal to the current tilt angle minus the step length.

4. The exoskeleton gait adaptive adjustment method based on terrain pre-sensing according to claim 1, characterized in that: The steps for determining the user's current gait phase based on hip and knee angles include: When the hip angle begins to decrease from its maximum flexion position and the knee angle increases from its extended position to more than 30 degrees, it is determined that the current phase is the early swing phase. When the hip joint angle decreases to below zero degrees and the knee joint angle increases to its peak and then begins to decrease, it is determined that the current phase is in the middle of the swing phase. When the hip joint angle continues to extend backward and the knee joint angle decreases to near zero degrees, it is determined that the current phase is the late swing phase. When the hip joint angle changes from extension to flexion and the knee joint angle remains less than 10 degrees, it is determined that the current position is the support phase.

5. The exoskeleton gait adaptive adjustment method based on terrain pre-sensing according to claim 1, characterized in that: If the gait phase is in the support phase, then the hip joint torque adjustment is set to: ΔT hip =Kp1·ΔH+Kv1·ΔV; Where ΔH is the height deviation, ΔV is the speed deviation, Kp1 is the height proportional gain with a value range of 80–120 N·m / m, Kv1 is the speed proportional gain with a value range of 40–60 N·m / (m / s); and the knee joint torque adjustment ΔT knee For ΔT hip Half of them and in the same direction; If the gait phase is in the swing phase, then the hip joint torque adjustment is set to: ΔT hip =Kp2·ΔH; Where Kp2 is less than Kp1 and its value ranges from 20 to 40 N·m / m; the knee joint torque adjustment ΔT knee Set to zero; If the gait phase is in the late swing phase or the early stance phase transition stage, then the hip joint torque adjustment is set as follows: ΔT hip =Kp3·ΔH+Kd·(dΔH / dt); Where Kp3 ranges from 50 to 80 N·m / m, Kd is the differential gain and ranges from 10 to 20 N·m / (m / s); the knee joint torque adjustment ΔT knee Set with ΔT hip They are equal and have the same direction.

6. The exoskeleton gait adaptive adjustment method based on terrain pre-sensing according to claim 1, characterized in that: Based on the torque adjustment of each joint, a real-time control command sequence for the drive motors of the hip and knee joints of the exoskeleton is generated, including: Divide a gait cycle into N control beats on an average time basis; Multiply the torque adjustment of each joint by an envelope curve that varies with time to obtain the instantaneous torque value corresponding to each control beat. The envelope curve takes the value of 100% of the torque adjustment in the early and middle stages of the swing phase, linearly decreases to 0% in the later stage of the swing phase, linearly increases to 80% of the torque adjustment in the early stage of the support phase, and linearly decreases to 0% in the later stage of the support phase. The instantaneous torque value of each control cycle is added to the basic assist torque value of the corresponding joint to obtain the final output torque value of the control cycle, and then arranged in chronological order to form a control command sequence.

7. The exoskeleton gait adaptive adjustment method based on terrain pre-sensing according to claim 6, characterized in that: Execute a real-time control command sequence to drive the corresponding joint motors to output auxiliary force, including: At the start of the next gait cycle, current commands are sent to the drive motors of the hip and knee joints according to the final output torque value of the first control beat in the control command sequence. At the end of each control cycle, the actual angle of the current joint is read. If the deviation between the actual angle and the expected angle exceeds the preset threshold of 5°, the final output torque value of the control cycle is increased by 20% and used as the instruction value for the next control cycle. Otherwise, the output continues according to the next control cycle value in the original sequence. Repeat the above process until all control cycles in the control command sequence have been executed.

Images