58 results about "Stable gait" patented technology

The invention discloses a biped robot reinforcement learning control method which comprises the following steps: after a model prediction controller receives a walking instruction, outputting expected angle data of each joint of a robot to a reinforcement learning neural network, and meanwhile, returning joint angle data of an actual strategy of the reinforcement learning neural network to the neural network by a sensing system at a bottom layer; a joint angle taking time as a sequence and planned by model prediction is compared with a joint angle actually generated by a neural network, and model prediction control is fused into a reinforcement learning training process by setting a reward function for punishment, so that the reinforcement learning training efficiency is improved, and a stable gait is more quickly achieved. Excellent gaits planned by the MPC are transplanted into reinforcement learning control, and control robustness can be improved under the condition that the excellent gaits of the MPC are reserved; and meanwhile, the joint angle data which is obtained by taking time as a sequence and is obtained by taking MPC as a planner can accelerate the reinforcement learning training process, so that the training speed and the control effect of reinforcement learning are greatly improved.

The invention relates to the technical field of biomechanical analysis, in particular to a health data input method and system and a storage medium. The method comprises the following steps: collecting multi-modal data of a subject, and carrying out gait event recognition to obtain a gait event mark; bilateral gait synchronous alignment is carried out according to the gait event marks, bilateral gait cycle data are obtained, and the bilateral gait cycle data comprise plantar pressure data, ground force data, joint angle data and motion state data; plantar pressure power processing is carried out according to the plantar pressure data and the ground force data, and plantar power parameters are obtained; estimating body segment inertial parameters of the subject; joint torque energy processing is carried out according to the body segment inertial parameters and the joint angle data, and joint power parameters are obtained; and performing gait acceleration calculation according to the motion state data to obtain a gait stability index. The accuracy, comprehensiveness and convenience of gait analysis are remarkably improved through a biomechanical analysis technology.

The invention belongs to the technical field of smart old-age care product behavior debugging, and particularly discloses a smart old-age care product behavior dynamic adjustment method based on interactive data perception. According to the method, a sole mechanical fragment, a lower limb movement fragment and a heart rate fragment are extracted based on gait cycle alignment, gait stability, gait coordination and physiological conformity indexes are calculated, and coupling sensing of user movement behaviors and physiological states in the dynamic walking assisting process is achieved; the gait stability and the physiological conformity are fused to construct a gait physiological synchronism criterion, a hierarchical adaptive control strategy is implemented accordingly, and the limitation that a traditional supporting robot only depends on kinematics feedback control is broken through, so that on the premise that the user safety is guaranteed, response type physical and mental collaborative intelligent assistance is achieved, and the user experience is improved. And the risk prevention and control capability of rehabilitation training is effectively improved.

The invention discloses a Parkinson's disease patient rehabilitation training method based on video gait analysis, relates to the technical field of image analysis, solves the problem of dependence on subjective evaluation in the prior art, and realizes objective quantitative analysis of a rehabilitation effect. An athletic ability index Xpzs, a gait stability index Yb and a lower limb rehabilitation progress coefficient Ws of a patient are calculated and evaluated to ensure that an evaluation result is more reliable and accurate; meanwhile, an established real-time data feedback mechanism is combined with historical data, the problem that rehabilitation effect monitoring is insufficient is solved, and through regular tracking and effect evaluation, a training scheme and a rehabilitation target are dynamically adjusted, and it is ensured that the rehabilitation progress of a patient is tracked in a whole cycle; besides, the method is high in technical adaptability, personalized analysis can be provided according to illness states and rehabilitation progresses of different patients, the training intensity and the rehabilitation target can be dynamically adjusted, the flexibility and adaptability of a rehabilitation scheme are improved, the method is particularly suitable for patient groups with different illness states, and the rehabilitation effect is improved.

The invention discloses a wheel-foot composite robot and a motion control method thereof, a terminal and a storage medium, and the motion control method of the wheel-foot composite robot enables the wheel-foot composite robot to have remarkable robustness and dynamic stability in a complex environment by performing self-adaptive control on a four-foot form and a two-wheel form under a unified control framework. The method specifically comprises the following steps: identifying a current motion form, and performing real-time integral estimation on an error between an expected state and an actual state by using a robust adaptive model predictive control algorithm in a four-foot form to obtain an equivalent external disturbance vector for control solution, thereby realizing online disturbance compensation and adaptive torque distribution. The robot has stronger anti-interference capability on load change and ground disturbance in a four-foot form, and can keep stable gaits and accurate attitude control under a relatively high load condition.

The invention discloses a chain type hybrid deep reinforcement learning quadruped robot control system. The system comprises a sensing module, a data processing module, a chain type hybrid reinforcement learning model module, a feedback module and the like. The invention further discloses a control method of the chain type hybrid deep reinforcement learning quadruped robot control system, and the method comprises the steps: S1, collecting bionic motion data, robot state data and environment data, and training based on an attention mechanism space-time diagram neural network and antagonistic imitation learning to obtain a reference gait model containing motion prior; and S2, constructing a three-level parallel simulation environment, setting a progressive course target, training the progressive course target by adopting an improved PPO algorithm, and optimizing the gait stability, the movement efficiency and the environmental adaptability. By means of high-precision data acquisition with the sampling frequency larger than or equal to 240 Hz and antagonistic imitation learning, motion priori knowledge is introduced, the training period can be shortened by 50% or above, abnormal gaits are avoided, and the training process is controllable.

The invention discloses a humanoid robot motion control method based on sequence feature processing, and the method comprises the steps: firstly, collecting motion data of a robot in different task scenes, and carrying out the preprocessing, so as to improve the data quality; then, a dynamic updating mechanism is adopted to extract local motion features, and the capturing capacity of short-time motion changes is enhanced; the motion trend of the robot is analyzed through an autoregression method, and the continuity and stability of a control instruction are improved. Based on the processing result, a self-adaptive control strategy is generated, so that the robot can optimize motion parameters in real time according to environment changes. The method shows high gait stability, energy consumption efficiency and motion adaptability in a simulation environment and an actual robot platform, and the motion control level of the humanoid robot in a complex task scene can be effectively improved. The method is suitable for various humanoid robot motion control tasks, and is especially suitable for an intelligent robot system with high stability and real-time performance requirements.

The invention relates to the technical field of data processing, in particular to a gait correction suggestion generation method and system. Comprising the steps that a three-dimensional coordinate data set of a patient is obtained through wearable equipment, a relevance feature vector between a gravity center transfer track and a foot landing point is calculated, and whether the track needs to be smoothed or not is judged according to the feature vector. And calculating a matching degree score of the gravity center transfer trajectory and the foot landing point in the dynamic scene through the smoothed trajectory data set, and generating a gait stability evaluation result according to the matching degree score. Based on the evaluation result, a personalized gravity center optimization path is generated, an optimization path guidance data set is transmitted to the equipment, the gait of the patient is adjusted, and an adjusted three-dimensional coordinate data set is generated. And finally, calculating a fall risk assessment index and generating an assessment report, and providing a targeted gait optimization suggestion. According to the method, the gait stability can be effectively improved, the falling risk is reduced, and a basis is provided for personalized gait correction.

The invention provides a gait balance training method and system based on virtual reality, and is applied to the technical field of data processing. The method comprises the following steps: processing knee joint sequence image information of a target user to generate knee joint motion range interval information; processing the physiological parameter information of the target user in a preset time period, and generating target diagnosis interval data information of the target user; processing the gait training purpose information of the target user and the target diagnosis interval data information of the target user to generate a gait training adjustment factor; processing the knee joint activity range interval information and the clinical attribute information of the target user based on a target gait training generation model, and generating gait posture stability information and gait stability time information; and processing the gait posture stability information, the gait stability time information and the gait training adjustment factor to generate target gait training information.

The invention discloses a quadruped robot supporting leg counter-acting force planning control method, which belongs to the technical field of robot control, and comprises the following steps: firstly, acquiring current gait information of a quadruped robot, and identifying a current supporting leg; secondly, constructing a ground support polygon model; then setting stable support constraint conditions based on a ground support polygon model, and calculating a current expected state of the mass center of the robot; constructing a virtual model of the quadruped robot, and calculating the reactive force of the current supporting legs of the quadruped robot; and finally, the output torque of each motion joint of the quadruped robot is calculated to generate a robot planning control instruction. A stable supporting constraint condition is formed by recognizing supporting legs of the quadruped robot and establishing a ground supporting polygon model, and the current reactive force of the supporting legs is calculated based on the stable supporting constraint condition, so that a robot planning control instruction is generated, planning control over the reactive force is achieved, and the gait stability of the robot is improved.

The invention discloses an orthopedic postoperative rehabilitation training method based on biomechanical feedback. The method comprises the following steps: synchronously collecting joint kinematics, myoelectricity and plantar pressure multi-source data of a patient; performing time sequence alignment and noise suppression on the data; constructing an individualized biomechanical function evaluation model, and generating a dynamic score covering the joint motion range, the muscle symmetry, the gait stability and the load tolerance; the training intensity, the auxiliary support and the rest time are adjusted in a self-adaptive mode based on the deviation between the score and the stage target; voice with clinical semantics and visual feedback are generated; and the evaluation and regulation strategy is continuously optimized through the closed-loop learning module. According to the technical scheme, precision, individualization and intelligentization of rehabilitation training are achieved, the recovery efficiency is effectively improved, and the risk of secondary injury is reduced.

The invention relates to the technical field of rehabilitation training, in particular to a Parkinson patient gait rehabilitation training device which comprises a mounting frame and a treadmill, a first linear sliding table is fixedly mounted at the top of the inner wall of the mounting frame, a mounting plate is slidably connected to the bottom of the first linear sliding table, and a sliding plate is slidably connected to the bottom of the mounting plate; the bottom of the sliding plate is fixedly connected with a vest type restraint strap through a furling assembly, the lower end of the vest type restraint strap is fixedly connected with a waist restraint strap, and a first pressure sensor is fixedly mounted at the bottom of the mounting plate. The pressure change of the sliding plate is detected in real time through the first pressure sensor to evaluate the gait stability, when it is detected that the frozen gait inclines forwards, running of the treadmill is stopped immediately, a patient is prevented from falling down and being injured, the first linear sliding table and the first retractor work cooperatively, the inclined patient is straightened to return to the normal posture, and the working efficiency is improved. Training time is prolonged, uncomfortable and nervous moods are relieved, and adverse effects on the body caused by long-time inclination are avoided.

This invention discloses a safe navigation and localization method for bipedal robots based on discrete control barrier functions and a hierarchical architecture, belonging to the field of robot control technology. The method employs the Relative Dynamics Test (RRT) algorithm to generate a global path and combines it with a Linear Inverted Pendulum Model (LIPM) to generate a stable gait. A discrete control barrier function (D-CBF) is used to ensure obstacle avoidance in dynamic environments, and the safe control input is solved through an optimization problem. Sensor fusion technology is used for real-time localization, and moving average filtering is applied to eliminate oscillation errors caused by gait. This invention, by combining discrete control barrier functions (D-CBF) with hybrid zero dynamics (HZD), achieves high-precision navigation and safe obstacle avoidance for bipedal robots in complex environments, effectively solving the problems of insufficient stability and lack of safety in traditional methods.

The application discloses a gait and balance training robot system and method; the system comprises a motion generating unit, the motion generating unit is used for generating gait trajectory motion, and the leg is driven to move in a gait; the motion generating unit comprises a parallel mechanism, a rocker mechanism and a connecting plate; the connecting plate and the connecting rod of the rocker mechanism are connected with the parallel mechanism, and are used for linking the parallel mechanism and the rocker mechanism; the rocker mechanism is used for generating gait trajectory motion; and the parallel mechanism is used for ensuring that the gait trajectory motion generated by the rocker mechanism is stable. The gait trajectory motion is generated by the rocker mechanism, the gait motion of the leg is simulated, and the leg is driven to move in a gait. The rocker mechanism has the advantages of simple structure and stable operation; and the parallel mechanism is linked with the rocker mechanism, so that the stable gait trajectory motion is further ensured, thereby achieving a better rehabilitation training effect.

The application discloses an electroencephalogram guided hip-knee-ankle orthopedic rehabilitation exoskeleton adaptive adjustment method, and relates to the technical field of intelligent rehabilitation engineering and biomedical engineering. On the basis of identifying the continuous electroencephalogram fluctuation characteristics in the intention transition period, the support response of the lower limb three joints is combined to construct a traceable transition identification zone. By extracting the driving dislocation point to set the joint response delay, the problem of early driving before the support is completed is avoided. The forward inclination trend is predicted in combination with the torso posture angle and the foot bottom pressure change, and a reverse support hierarchical matrix is dynamically arranged among the three joints to ensure the timing matching and compensation of the support forces among the joints when the abnormal driving trend occurs. Through the rhythmized inhibition and micro-assistance alternating mechanism, the joint output forms a flexible buffer gait, so that the gait stability, the neural participation degree and the safety of the patient in the rehabilitation training process are significantly improved.

According to the outdoor walking auxiliary type portable foot drop correction electrical stimulation device provided by the invention, through innovative design, the correction effect and the user experience are remarkably improved; the device adopts a multi-mode sensor array and an original gait phase prediction algorithm, real-time gait recognition with response delay lower than 50 ms is achieved, the walking rhythm of a patient is accurately matched, and it is ensured that electrical stimulation and a motion period are synchronous; based on a 3D stimulation mode library of terrain recognition and a self-adaptive intensity adjustment algorithm, stimulation parameters can be dynamically adjusted according to different terrains, and the gait stability in a complex environment is effectively improved; the bionic electrode layout and the magnetic type quick wearing system are combined with an invisible design, so that the wearing comfort and convenience are remarkably improved; clinical tests show that the device enables the step speed of a patient to be increased, the number of tumble times to be reduced, the comfort level of a user to be high, technical spanning from clinic to daily life is achieved, and an efficient, safe and easy-to-popularize rehabilitation solution is provided for the foot drop patient.

The invention relates to the technical field of gait stability evaluation, in particular to a gait stability detection method based on multi-cell-body set membership estimation. The method comprises the steps that based on gait data of normal walking of a subject, a stability area template based on multi-cell bodies is constructed in an off-line mode; acquiring real-time gait data of a subject, deducing a prediction state set at the current moment by adopting a set member filtering algorithm and utilizing a posterior state set, carrying out measurement constraint on the prediction state set based on actual measurement data, and carrying out online recursion on a state feasible set; and carrying out a set inclusion test on the state feasible set and the stability region template, judging whether the current gait is stable or not, calculating a risk score according to the distance between the state feasible set and the stability region template, and generating real-time early warning. Compared with the prior art, the method can provide more accurate fall risk assessment, and significantly improves the safety and life quality of old people and special groups.

The present application relates to a kind of soft landing reward function-based quadruped robot low-noise gait control method and system, the method includes: constructing quadruped robot motion control environment, and defining foot contact determination standard;Design soft landing reward function based on vertical contact force, for the negative punishment of large contact force at the moment of landing, guide quadruped robot to learn low contact force landing mode;Collect the real-time contact force of quadruped robot foot and historical contact state, with soft landing reward function as core optimization target, using reinforcement learning algorithm, in combination with gait stability constraint, iterative training is carried out to strategy network, and low-noise gait strategy network is obtained, for optimizing output gait control signal, to control the gait action of quadruped robot accordingly.Compared with prior art, the present application can consider the noise control and gait stability of robot by accurately determining the moment of landing, designing contact force negative reward and combining reinforcement learning to actively optimize gait strategy.

The invention discloses a leg structure with bevel gear drive and a leg-foot type robot. The leg structure comprises a hip joint motor, a thigh joint motor and a shank joint motor. The hip joint motor is connected with a thigh joint motor, a thigh transmission shaft is installed on the thigh joint motor and meshed with a second thigh gear on the thigh rotating shaft through a first thigh gear, and a shank transmission shaft is installed on the shank joint motor and meshed with a second thigh gear on the thigh rotating shaft through a second thigh gear. The shank transmission shaft is meshed with a second shank gear on the shank rotating shaft through a first shank gear, and the first thigh gear, the second thigh gear, the first shank gear and the second shank gear are all bevel gears. The bevel gear transmission system adopted in the invention effectively amplifies the torque output of the motor, so that the robot can still maintain a stable gait when facing an abrupt slope or under the condition of large load, the motor load is effectively reduced, the service life of the motor is prolonged, and the energy consumption is reduced.

The application relates to the technical field of robot control, and discloses a biped robot walking control method based on model predictive control and a landing point optimization. The method aims to solve the problems of insufficient stability and poor adaptability of the prior art when coping with external disturbances and uneven terrains. The core of the application is to perform optimization control on the supporting leg and the swing leg respectively: model predictive control (MPC) is adopted to realize real-time optimization of the three-dimensional ground reaction force of the supporting leg, so that the effective support of the supporting leg on the torso is ensured; meanwhile, a gait stability feedback controller is designed, real-time state feedback is provided through a state observer, the swing leg trajectory is optimized, the landing point position of the robot is dynamically corrected, and the anti-interference ability of the system is enhanced.

The utility model relates to a kind of negative pressure adsorption type four-foot wall-climbing robots, including torso main body, four leg components, adsorption component and drive assembly.Torso main body includes left, right, front torso plate and rear torso rod, both sides are equipped with connecting plate;Leg component is two-degree-of-freedom bionic structure, respectively articulated on connecting plate, it includes thigh and shank skeleton, and by joint rudder independent drive.Adhesion component includes four suction discs and four independent corresponding micro diaphragm pumps arranged in lower end of shank, form independent vacuum unit of sub-foot;Lower end of shank is connected with suction disc universal self-adapting by ball shaft seat, ball bearing and installation column with air passage.The utility model mechanical structure simplification and light weight, greatly improve the fault tolerance of system, security and the adaptive ability of rough, inclined special-shaped wall surface, can be with stable gait on a variety of material wall surface steady crawl.

The invention provides a multi-gait controller generation method, a controller and a quadruped robot, and relates to the technical field related to robot control, in the multi-gait controller generation method, a two-stage training framework is designed based on deep reinforcement learning, the controller is generated through the two-stage training framework, and the quadruped robot is controlled through the controller. The control mode can be divided into automatic posture control and manual posture control, so that when kinematics and dynamics models of the quadruped robot are not needed, the quadruped robot can be controlled to stably move in various gaits only by using an IMU and an encoder of a joint motor as body sensors, and in addition, the stability of the quadruped robot is improved. It is guaranteed that the quadruped robot can pass under the complex terrain and stably move in various gaits.

The application discloses a wheel-foot composite robot, a motion control method thereof, a terminal and a storage medium, and the motion control method of the wheel-foot composite robot is characterized in that the four-foot mode and the two-wheel mode are respectively controlled adaptively under a unified control framework, so that the wheel-foot composite robot has remarkable robustness and dynamic stability in a complex environment. Specifically, the current motion mode is identified, the error between the expected state and the actual state is estimated by using a robust adaptive model predictive control algorithm in the four-foot mode, an equivalent external disturbance vector is obtained and is used for control solution, so as to realize online disturbance compensation and adaptive torque distribution, so that the robot has stronger anti-interference ability to load change and ground disturbance in the four-foot mode, and can maintain stable gait and accurate attitude control under the condition of a higher load.

The invention relates to a gait analysis and fall risk early warning method based on chronic diseases of old people, and belongs to the technical field of health monitoring and medical assistance. The method comprises the steps of collecting gait data of old people in real time through a wearable device, extracting stride and stride frequency features, combining chronic disease information, constructing a personalized fall risk prediction model by using a machine learning algorithm, adjusting model parameters through an optimization algorithm based on gait features and individual physiological parameters, and generating a personalized early warning scheme. The gait stability and the risk level are evaluated according to the monitoring module, and when a high risk is detected, early warning is given out through vibration or sound. The system adopts portable equipment, gets rid of laboratory limitation and is suitable for daily health monitoring, the system optimizes the model through continuous data updating, the prediction precision is improved, effective early warning of gait analysis and the falling risk is achieved, and the system is suitable for popularization and application. The falling risk of old people is reduced.

The invention discloses a humanoid robot foot bearing structure which comprises a front foot bottom plate, a rear foot bottom plate, an imitated foot arch piece and an imitated foot sole fascia, the front foot bottom plate, the rear foot bottom plate and the imitated foot arch piece form a triangular arch support capable of being longitudinally deformed, the imitated foot sole fascia elastically deforms along with the imitated foot arch piece, a flexible rotating mechanism is arranged in the front arch section of the imitated foot arch piece, and the front arch section of the imitated foot arch piece is connected with the imitated foot sole fascia. Therefore, the foot arch imitating piece can transversely rotate, human foot tarsal transverse joint movement is simulated, and transverse disturbance is relieved. The device has the advantages that the front arch section and the rear arch section can rotate inwards and outwards around the rigid connecting shaft by utilizing the flexible rotating mechanism, so that the adaptability to complex terrains and the gait stability are improved; the triangular arch structure composed of the foot arch imitating pieces can be dynamically opened, closed and deformed, the impact force generated when the robot walks is dispersed, the vertical rigidity is reduced, and the buffering time is prolonged. The flexible foot sole imitating fascia simulates a human foot sole fascia mechanism, energy is stored when the foot arch imitating piece is compressed, restoring force is generated when the foot is lifted, and the gait efficiency and foot naturalness of the robot are improved.

This application discloses a gait analysis system and method. The gait analysis system includes multiple inertial measurement modules, multiple pressure measurement modules, a depth camera, an optical motion capture module, a data filtering module, and a main control module. Through the cooperation of multiple modules, various types of data reflecting gait characteristics are collected. The data can compensate for each other and collaboratively reflect the user's gait characteristics. Finally, the main control module obtains the user's gait analysis results, realizing a comprehensive analysis of gait characteristics such as stride length, stride width, stride frequency, gait rhythm, gait periodicity, gait stability, and gait coordination. This gait analysis system can eliminate the problem of inaccurate gait analysis caused by the limited information collected by a single sensor, reduce the error interference of external environmental influences on gait analysis, improve detection accuracy and reliability, and achieve accurate and comprehensive analysis of the user's gait.

The invention discloses a biped semi-passive halving upper body walking device. The biped semi-passive halving upper body walking device comprises a hip part, two leg parts, ankle joints and foot parts, the hip part comprises a hip part torsion motor and a hip part balance plate, is connected with the hip part rotating piece through a vertical rotating shaft in a rotating fit manner, and is connected with a first back-and-forth swinging motor of the thigh part through a hip joint driving mechanism of a hip joint driving module and a leg opening motor; a second front-back swinging motor arranged on the thigh part and a driving motor arranged on the shank part are respectively connected with a knee joint shaft and an ankle joint shaft through belt transmission; the foot part is of a ship-shaped foot structure and is connected with the ankle joint shaft through a sole connecting piece, and damping springs are evenly distributed around the foot part. The walking stability can be ensured in the walking process, the foot flexibility is kept through the semi-passive ankle joint mechanism, multiple motors are used for driving, and the structure is compact and simple. By controlling a plurality of motors, human gait-simulated stable walking can be well realized.

This invention discloses a humanoid robot motion adaptive learning system, relating to the field of robot control technology. It includes the following steps: during the humanoid robot's movement, collecting foot pressure distribution data, ground friction coefficient data, ground texture images, and body posture data; performing feature fusion processing on the foot pressure distribution data, ground friction coefficient data, and body posture data to obtain a comprehensive feature vector of ground conditions; constructing a condition database and determining the current condition based on the ground condition feature vector; generating gait parameter correction instructions based on the current condition, and controlling the humanoid robot to correct its gait according to the gait parameter correction instructions. This invention can accurately locate the current condition, dynamically adjust the correction coefficient by combining the current control parameter deviation value and texture features, providing precise adaptation and local fine-tuning for gait parameters, and achieving stable gait control of the humanoid robot in scenarios with sudden changes in ground conditions and local differences within the same condition.

The application discloses a kind of based on real-time evaluation's old person exercise capacity recession intervention management method and system, it is related to old person exercise capacity management field, including: establishing exercise capacity baseline database;Through smart bracelet and plantar pressure shoe pad real-time collection data under the condition of exercise;Get four core exercise capacity indexes of gait stability, balance ability, exercise intensity and fatigue degree;Build exercise capacity recession real-time evaluation model, generate intervention threshold, determine current exercise capacity grade and recession risk;Execute individualized exercise intervention, intensity adjustment, rest prompt and fall risk early warning;Trigger baseline review, re-evaluate the exercise capacity state of old people, generate weekly health report.The application has the advantages that: through multi-source data fusion and four-stage progressive evaluation, realize the real-time, individualized monitoring and dynamic classification intervention of old person exercise capacity, effectively improve the accuracy of recession risk early warning and the adaptability of intervention measures.

The invention discloses a design method and a control strategy for an energy recovery system of an exoskeleton robot, and the method specifically comprises the following steps: S1, analyzing the energy transfer process of a motor in the exoskeleton robot, and obtaining the type of energy to be recovered; s2, calculating the magnitude of positive and negative work applied by the motor under the symmetrical stable gait according to the power change of the motor; and S3, the energy recovery structure is analyzed, and an energy recovery system is constructed.