Lower limb exoskeleton trajectory customization method based on generative model

Abstract

The invention discloses a lower limb exoskeleton trajectory customization method and device based on a generative model, equipment and a storage medium, and belongs to the field of assistance equipment, and the method comprises the steps: generating first step phase feedback data of a user in response to first mechanical feedback between the lower limb of the user and an exoskeleton in a first walking process of the user in a first walking state; determining an abnormal evaluation index value of the first walking behavior according to the first step phase feedback data; and adjusting the first feedback value according to the abnormal evaluation index value so as to reduce the abnormal evaluation index value. Based on the method provided by the embodiment of the invention, the individuation and real-time optimization of the exoskeleton equipment are realized through the real-time adjustment of the anomaly detection and the feedback value, the response capability and applicability of the power assisting equipment are improved, the defects in real-time optimization and individualized customization in the prior art are made up, and the user experience is improved. The problems that an existing power assisting device is single in power assisting mode and insufficient in universal power assisting effect are solved.

Description

Technical Field

[0001] The present application belongs to the field of power-assisting equipment, and specifically relates to a method, device, equipment and storage medium for customizing lower limb exoskeleton trajectories based on a generative model. Background Art

[0002] In most lower limb exoskeleton application scenarios, users wear robots to assist walking to enhance lower limb strength. However, traditional lower limb exoskeleton robots can only drive users to move according to preset trajectories, and cannot be optimized in real time to meet individual needs, which affects the assistance effect.

[0003] Currently, there are some methods on the market to optimize exoskeleton devices, for example, by improving the mechanical structure design to make it fit the human body better.

[0004] However, these methods are limited to improving the mechanical structure and cannot achieve personalized real-time optimization, which results in the device being unable to adjust according to the user's real-time needs during use. Summary of the invention

[0005] The present application aims to provide a method, device, equipment and storage medium for customizing the trajectory of a lower limb exoskeleton based on a generative model, which at least solves the problem that the power-assisting device has a single power-assisting mode and lacks the versatility of the power-assisting effect.

[0006] In a first aspect, an embodiment of the present application discloses a method for customizing a lower limb exoskeleton trajectory based on a generative model, comprising: In response to first mechanical feedback between the lower limbs of the user and the exoskeleton during a first walking process of the user in a first traveling state, first step phase feedback data of the user is generated; the first step phase feedback data is used to characterize the corresponding relationship between different lower limb postures of the user and the first feedback value of the first mechanical feedback during the first walking process; According to the first step phase feedback data, an abnormal evaluation index value for the first walking behavior is determined; the degree of difference between the abnormal evaluation index value and the step phase prediction feedback data for the first step phase feedback data is positively correlated; the step phase prediction feedback data is the target step phase feedback data of the target mechanical feedback between the lower limbs of the user and the exoskeleton in the target walking state of the user; According to the abnormal evaluation index value, the first feedback value is adjusted so that the abnormal evaluation index value is reduced.

[0007] In a second aspect, the present application also discloses a lower limb exoskeleton trajectory customization device based on a generative model, comprising: a data acquisition module, configured to generate first step phase feedback data of the user in response to first mechanical feedback between the lower limbs of the user and the exoskeleton during a first walking process of the user in a first walking state; the first step phase feedback data is used to characterize the corresponding relationship between different lower limb postures of the user and a first feedback value of the first mechanical feedback during the first walking process; An evaluation module is used to determine an abnormal evaluation index value for the first walking behavior according to the first step phase feedback data; the abnormal evaluation index value is positively correlated with the step phase prediction feedback data for the first step phase feedback data; the step phase prediction feedback data is the target step phase feedback data of the target mechanical feedback between the lower limbs of the user and the exoskeleton in the target walking state of the user; A feedback module is used to adjust the first feedback value according to the abnormal evaluation index value so that the abnormal evaluation index value decreases.

[0008] In a third aspect, an embodiment of the present application further discloses an electronic device, comprising a processor and a memory, wherein the memory stores programs or instructions that can be run on the processor, and when the program or instructions are executed by the processor, the steps of the method described in the first aspect are implemented.

[0009] In a fourth aspect, an embodiment of the present application further discloses a readable storage medium, on which a program or instruction is stored. When the program or instruction is executed by a processor, the steps of the method described in the first aspect are implemented.

[0010] In summary, in the embodiment of the present application, by obtaining the first mechanical feedback of the user in the first walking state, the first step phase feedback data of the user is generated, and the corresponding relationship between the user's lower limb posture and the mechanical feedback is recorded in real time, which provides basic data for subsequent trajectory optimization; then, according to the step phase feedback data, the abnormal evaluation index value of the first walking behavior is calculated, and by analyzing the difference between the step phase feedback data and the normal gait, the abnormality that may exist in the user during the power-assisting process is identified, and then the step phase prediction feedback data of the user in the target walking state is generated to provide a reference standard, and by comparing with the actual feedback data, the abnormal detection mechanism is further improved, and finally, according to the abnormal evaluation index value, the first feedback value is adjusted to reduce the abnormal evaluation index value, and the power-assisting trajectory is optimized in real time by dynamically adjusting the feedback value, ensuring that the exoskeleton device is adjusted according to the actual needs of the user, and improving the power-assisting effect. Therefore, based on the method of the embodiment of the present application, through the real-time adjustment of abnormal detection and feedback values, the personalization and real-time optimization of the exoskeleton device are realized, the responsiveness and applicability of the power-assisting device are improved, the deficiencies of the prior art in real-time optimization and personalized customization are made up, and the problems of the single power-assisting mode and insufficient general power-assisting effect existing in the prior power-assisting device are solved. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In the attached picture: Figure 1 It is a flowchart of the steps of a method for customizing lower limb exoskeleton trajectories based on a generative model provided in an embodiment of the present application; Figure 2 is a schematic diagram of the structure of a lower limb exoskeleton provided according to an embodiment of the present application; Figure 3 is a flowchart of another method for customizing lower limb exoskeleton trajectories based on a generative model provided in an embodiment of the present application; Figure 4 It is the conversion model data flow process under the embodiment of the present application; Figure 5 It is the adjustment process of the mechanical feedback under the embodiment of the present application; Figure 6 is a block diagram of a lower limb exoskeleton trajectory customization device based on a generative model provided in an embodiment of the present application; Figure 7 is a block diagram of an electronic device according to an embodiment of the present application; Figure 8 It is a block diagram of an electronic device of another embodiment provided by the embodiments of the present application. DETAILED DESCRIPTION

[0012] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0013] The terms "first", "second", etc. in the specification and claims of the present application are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable under appropriate circumstances, so that the embodiments of the present application can be implemented in an order other than those illustrated or described here, and the objects distinguished by "first", "second", etc. are generally of one type, and the number of objects is not limited. For example, the first object can be one or more. In addition, "and / or" in the specification and claims represents at least one of the connected objects, and the character " / " generally indicates that the objects associated with each other are in an "or" relationship.

[0014] Figure 1 This embodiment provides a lower limb exoskeleton trajectory customization method based on a generative model.

[0015] The method may include the following steps: Step 101 , in response to a first mechanical feedback between a user's lower limbs and an exoskeleton during a first walking process in a first moving state, generates first step phase feedback data of the user.

[0016] The first step phase feedback data is used to characterize the corresponding relationship between different lower limb postures of the user and the first feedback value of the first mechanical feedback during the first walking process.

[0017] like Figure 2 As shown, in some embodiments of the present application, the lower limb exoskeleton provides mechanical feedback to the user's lower limbs through serial elastic actuators (SEA) to achieve driving assistance. The lower limb exoskeleton will respond to the first mechanical feedback between the user's lower limbs and the exoskeleton during the first walking process in the first moving state, and generate the user's first step phase feedback data. Specifically, the first step phase feedback data is used to characterize the correspondence between the user's different lower limb postures and the first feedback value of the first mechanical feedback during the first walking process. By recording the mechanical feedback data of the user at each gait stage when walking, the dynamic changes of the user's gait can be accurately reflected, providing reliable basic data for subsequent trajectory optimization.

[0018] For example, during the experiment, when a tester walks on flat ground, the lower limb exoskeleton device he wears will record his lower limb movement and corresponding mechanical feedback data in real time through the mechanical sensors built into SEA. These feedback data can include mechanical acquisition data such as interactive torque. Assuming that at a certain moment in the tester's walking, his right leg is in the support phase and his left leg is in the swing phase, the exoskeleton device collects the user's lower limb posture and force feedback data through sensors installed on each joint, and generates the corresponding first step phase feedback data. This data can reflect the mechanical characteristics of the right leg support force and the left leg swing trajectory, and provide data support for trajectory optimization and anomaly detection in subsequent steps.

[0019] Step 102: Determine an abnormal evaluation index value for the first walking behavior based on the first step feedback data.

[0020] Among them, the degree of difference between the abnormal evaluation index value and the step phase prediction feedback data of the first step phase feedback data is positively correlated; the step phase prediction feedback data is the target step phase feedback data of the target mechanical feedback between the user's lower limbs and the exoskeleton when the user is in the target walking state.

[0021] In some embodiments of the present application, an abnormal evaluation index value for the first walking behavior will be determined based on the first step phase feedback data. Specifically, the abnormal evaluation index value is used to characterize the degree of difference between the user's actual walking behavior and the expected walking behavior. The step phase prediction feedback data is the target step phase feedback data of the target mechanical feedback between the user's lower limbs and the exoskeleton under the target walking state. The abnormal evaluation index value is positively correlated with the degree of difference between the step phase prediction feedback data of the first step phase feedback data. This process can help identify abnormal situations that may exist during the user's walking process, thereby providing a basis for subsequent optimization.

[0022] For example, during the experiment, assuming that the tester is walking on flat ground, the first step phase feedback data of the gait phase records the mechanical feedback value between the lower limbs and the exoskeleton. By comparing these actual feedback data with the predicted step phase feedback data in the normal walking mode, the system can calculate the abnormal evaluation index value. Assuming that under normal circumstances, the mechanical feedback value of the right leg support phase is 50Nm, and the actual measured feedback value is 45Nm, the system calculates a positively correlated abnormal evaluation index value, through which it can be judged that there is a certain difference between the user's actual walking behavior and the expected walking behavior. This abnormal evaluation index value provides the necessary data basis for subsequent trajectory optimization and mechanical feedback adjustment. In this process, human-in-the-loop optimization (HIL optimization) is generally used for continuous calculation to finally obtain a stable and effective abnormal evaluation index value.

[0023] Step 103: adjust the first feedback value according to the abnormal evaluation index value so as to reduce the abnormal evaluation index value.

[0024] In some embodiments of the present application, the first feedback value will be adjusted according to the abnormal evaluation index value so that the abnormal evaluation index value is reduced. Specifically, when a high abnormal evaluation index value is detected, the system will adjust the first feedback value according to the index value, thereby optimizing the assistance trajectory of the exoskeleton and reducing the probability of abnormality. By dynamically adjusting the feedback value, it is possible to respond to the actual needs of users in real time and improve the responsiveness and applicability of the system.

[0025] For example, during the experiment, suppose the system detects that the abnormal evaluation index value of the user's right leg support phase during walking is 5, while under normal circumstances the value should be 0. Based on this abnormal evaluation index value, the system will adjust the first feedback value of the right leg support phase in real time to make it more in line with the actual needs of the user. For example, the support torque may be increased or the swing angle may be adjusted until the abnormal evaluation index value drops to a reasonable range close to 0. In this way, the system can effectively reduce abnormal situations during walking and ensure the power-assisting effect of the exoskeleton device and the walking stability of the user. This adaptive mechanism is based on the abnormal evaluation index value, which can guide the generation of personalized power assistance for users in the reinforcement learning algorithm (Proximal Policy Optimization, PPO), and will also dynamically adjust the impedance model of the exoskeleton to ensure the power-assisting effect while ensuring the comfort and safety of the wearer.

[0026] In summary, in the embodiment of the present application, by obtaining the first mechanical feedback of the user in the first walking state, the first step phase feedback data of the user is generated, and the corresponding relationship between the user's lower limb posture and the mechanical feedback is recorded in real time, which provides basic data for subsequent trajectory optimization; then, according to the step phase feedback data, the abnormal evaluation index value of the first walking behavior is calculated, and by analyzing the difference between the step phase feedback data and the normal gait, the abnormality that may exist in the user during the power-assisting process is identified, and then the step phase prediction feedback data of the user in the target walking state is generated to provide a reference standard, and by comparing with the actual feedback data, the abnormal detection mechanism is further improved, and finally, according to the abnormal evaluation index value, the first feedback value is adjusted to reduce the abnormal evaluation index value, and the power-assisting trajectory is optimized in real time by dynamically adjusting the feedback value, ensuring that the exoskeleton device is adjusted according to the actual needs of the user, and improving the power-assisting effect. Therefore, based on the method of the embodiment of the present application, through the real-time adjustment of abnormal detection and feedback values, the personalization and real-time optimization of the exoskeleton device are realized, the responsiveness and applicability of the power-assisting device are improved, the deficiencies of the prior art in real-time optimization and personalized customization are made up, and the problems of the single power-assisting mode and insufficient general power-assisting effect existing in the prior power-assisting device are solved.

[0027] Figure 3 This is another lower limb exoskeleton trajectory customization method based on a generative model provided in an embodiment of the present application.

[0028] The method may include the following steps: Step 201, in response to a first mechanical feedback between a user's lower limbs and an exoskeleton during a first walking process of the user in a first walking state, generates first step phase feedback data of the user and obtains a first walking trajectory of the user.

[0029] The first step phase feedback data is used to characterize the corresponding relationship between different lower limb postures of the user and the first feedback value of the first mechanical feedback during the first walking process.

[0030] In some embodiments of the present application, in response to the first mechanical feedback between the user's lower limbs and the exoskeleton during the user's first walking process in the first walking state, the user's first step phase feedback data is generated and the user's first walking trajectory is obtained. Specifically, the first step phase feedback data is used to characterize the correspondence between the user's different lower limb postures and the first feedback value of the first mechanical feedback during the first walking process. By recording the mechanical feedback data and the walking trajectory of the user, the necessary data basis can be provided for the optimization and adjustment of the exoskeleton trajectory.

[0031] For example, during the experiment, when a tester walks on flat ground wearing a lower limb exoskeleton device, the system will record the first step phase feedback data and trajectory of his lower limb movement in real time. Assuming that at a certain moment of walking, the tester's left leg is in the swing phase and the right leg is in the support phase, the exoskeleton device collects the corresponding mechanical feedback data through sensors installed at each joint and generates the first step phase feedback data. At the same time, the system also records the tester's trajectory during walking, which includes the correspondence between different postures of the left and right legs and the mechanical feedback. This information will be used for subsequent trajectory optimization and personalized power adjustment.

[0032] Optionally, in order to generate the user's first step phase feedback data, step 201 includes the following sub-steps: Sub-step 2011, in response to a first mechanical feedback between the lower limbs of the user and the exoskeleton during a first walking process of the user in the first walking state, obtain multiple gait postures of the user in the first walking state, and the mechanical collection value in each gait posture.

[0033] In some embodiments of the present application, in response to the first mechanical feedback between the user's lower limbs and the exoskeleton during the user's first walking process in the first walking state, multiple gait postures of the user in the first walking state and the mechanical acquisition values ​​at each gait posture are obtained. Specifically, during the user's walking process, the exoskeleton device monitors and records the user's multiple gait postures in real time through the body sensor, and records the corresponding mechanical feedback values ​​at each gait posture. These mechanical acquisition values ​​include mechanical parameters such as joint angles and torques. These data can reflect in detail the actual movement state of the user in each gait posture and the response of the exoskeleton.

[0034] For example, during the experiment, when a tester walks on flat ground wearing a lower limb exoskeleton device, the system will collect multiple gait posture data of the user while walking in real time through sensors such as the inertial measurement unit (IMU) and encoders on the device. These gait postures may include stages such as heel landing, sole landing, and toe off. In each gait posture, the system will also record the corresponding mechanical feedback value. For example, when the heel lands, the joint angle and torque value at that moment are recorded. By collecting these detailed mechanical feedback data, the system can generate the first step phase feedback data of the user in the first walking state, providing a basis for subsequent gait optimization and anomaly detection.

[0035] Sub-step 2012, compressing multiple one-to-one corresponding gait postures and mechanical acquisition values ​​into a code of a preset length to obtain the first step phase feedback data.

[0036] In some embodiments of the present application, multiple one-to-one corresponding gait postures and mechanical acquisition values ​​are compressed into a code of a preset length to obtain the first step phase feedback data. Specifically, the multiple gait postures and corresponding mechanical feedback values ​​recorded during the user's walking process are converted into a fixed-length code through a compression algorithm. This code not only retains the key features of the original data, but also effectively reduces data redundancy, which facilitates subsequent processing and analysis. The preset length code can be used for real-time analysis and processing to improve the response speed and efficiency of the system.

[0037] For example, during the experiment, the system collected the user's gait posture and mechanical feedback value data when walking on flat ground. These data include joint angles and torque values ​​in multiple gait stages. The system converts these data into fixed-length codes through a compression algorithm, such as codes with a length of 128. In this process, the system extracts key features in the gait posture and mechanical feedback values, such as maximum torque value, average joint angle, etc. In this way, the generated first step phase feedback data encoding can not only effectively characterize the user's walking state, but also simplify the data processing process, providing data support for subsequent trajectory optimization and power adjustment.

[0038] Optionally, in order to obtain the first travel trajectory of the user, step 201 includes the following sub-steps: Sub-step 2013, in response to the first mechanical feedback between the lower limbs of the user and the exoskeleton during the first walking process of the user in the first moving state, obtain multiple gait postures of the user in the first moving state and the time corresponding to each gait posture.

[0039] In some embodiments of the present application, in response to the first mechanical feedback between the lower limbs of the user and the exoskeleton during the first walking process of the user in the first walking state, multiple gait postures of the user in the first walking state and the time corresponding to each gait posture are obtained. Specifically, the user's walking process is monitored in real time by sensors, and the gait posture and the corresponding time points are recorded. The gait posture includes the postures of the user at different stages of walking, such as heel landing, sole landing, toe off, etc. These postures occur at different time points during the walking process. By recording these data, the user's travel trajectory can be obtained.

[0040] For example, during the experiment, a tester walked while wearing a lower limb exoskeleton device. The system collects multiple gait posture data and corresponding moments of the tester while walking in real time through sensors such as the inertial measurement unit (IMU) and encoders installed on the device. For example, at a certain moment in the gait cycle, the tester's right heel touches the ground, and the system records this moment and the corresponding posture data; at another moment, the left toe leaves the ground, and the system also records this moment and the corresponding posture data. In this way, the system can record in detail multiple gait postures and their corresponding moments during the user's walking process, providing data support for generating the first travel trajectory.

[0041] In sub-step 2014, a time sequence generated according to a plurality of one-to-one corresponding gait postures and moments is determined as a first moving trajectory.

[0042] In some embodiments of the present application, a time series sequence generated according to a plurality of one-to-one corresponding gait postures and moments is determined as a first moving trajectory. Specifically, by collecting a plurality of gait postures of the user during walking and their corresponding moment data, these data can be arranged in chronological order to form a complete time series sequence. This time series sequence can accurately reflect the motion trajectory of the user during walking and provide basic data for trajectory optimization of the exoskeleton system.

[0043] For example, during the experiment, the system recorded multiple gait posture data and corresponding moments when the user was walking on flat ground, such as the moment when the right leg heel touched the ground was 0.5 seconds, the moment when the left leg sole touched the ground was 1.2 seconds, and the moment when the right leg toe left the ground was 1.8 seconds. By arranging these gait posture data and their moments in chronological order, the system generated the user's first walking trajectory. This trajectory can reflect the user's dynamic movement during the entire walking process, providing data support for subsequent trajectory optimization and power adjustment.

[0044] Step 202: input the first traveling trajectory into a conversion model to obtain a predicted trajectory of a second traveling trajectory of the user in a second traveling state.

[0045] In some embodiments of the present application, the first moving trajectory of the user in the first moving state is input into the conversion model to obtain a predicted trajectory of the second moving trajectory of the user in the second moving state. Specifically, the conversion model generates a predicted moving trajectory of the user in the second moving state by analyzing the data of the first moving trajectory and combining the mechanical feedback information of the user in different moving states. This process can generalize the movement pattern of the user in one moving state to other moving states, thereby providing support for the multi-task adaptability of the exoskeleton.

[0046] For example, Figure 4 As shown in the figure, the table q represents the trajectory value of the trajectory, and t represents time. During the experiment, a tester's trajectory data is recorded and input into the conversion model in the first state of walking on flat ground. Assume that at this time, the conversion model predicts the tester's trajectory in the second state of going up and down stairs based on these data. By analyzing the motion pattern of walking on flat ground, the model can generate the tester's gait and mechanical feedback information when going up and down stairs, thereby realizing the prediction and adaptation of the trajectory under different task states. This predicted trajectory provides a data basis for subsequent personalized assistance adjustments. W wm is the trajectory encoding in the first traveling state (walking on flat ground), and is mapped to the trajectory encoding under the second walking state (stair climbing) task, and Fr is a mapping formed by the transformation model. The above process can be simply expressed as: ,This task switching combined with personalized control enables the exoskeleton not only to handle complex ,task scenarios, but also to make adaptive adjustments based on the body parameters and gait habits of different ,users.

[0047] Step 203: Determine an abnormal evaluation index value for the first walking behavior based on the first step phase feedback data.

[0048] Among them, the degree of difference between the abnormal evaluation index value and the step phase prediction feedback data of the first step phase feedback data is positively correlated; the step phase prediction feedback data is the target step phase feedback data of the target mechanical feedback between the user's lower limbs and the exoskeleton when the user is in the target walking state.

[0049] The method shown in this step has been explained in step 102 and will not be repeated here.

[0050] Optionally, step 203 includes the following sub-steps: Sub-step 2031, inputting the first phase feedback data into the prediction model to obtain the first phase prediction feedback data for the first phase feedback data.

[0051] In some embodiments of the present application, the first step phase feedback data is input into the prediction model to obtain the first step phase prediction feedback data for the first step phase feedback data. Specifically, the first step phase feedback data includes the mechanical feedback values ​​of the user at each gait stage during walking. The prediction model generates corresponding prediction feedback data by analyzing and calculating these data. The prediction feedback data reflects the mechanical feedback characteristics that the user should have in a normal walking state, and is used to evaluate the deviation of the actual feedback data.

[0052] For example, during the experiment, the system inputs the first step feedback data of a tester when walking on flat ground into the prediction model. The feedback data records the mechanical feedback values ​​of the user at each gait stage (such as heel strike, sole strike, toe off, etc.). The prediction model generates a set of prediction feedback data by analyzing these data. It is assumed that in the heel strike stage, the predicted mechanical feedback value is 48Nm, and the actual recorded feedback value is 50Nm. In this way, the system can obtain the user's gait prediction feedback data in a normal walking state, providing a reference for subsequent abnormality detection and feedback adjustment.

[0053] Sub-step 2032, determining the second-order moment of the first-phase feedback data and the first-phase predicted feedback data as an abnormal evaluation index value.

[0054] In some embodiments of the present application, the second-order moment of the first-phase feedback data and the first-phase predicted feedback data is determined as the abnormal evaluation index value. Specifically, the second-order moment is a term in statistics that is used to measure the dispersion or variation of a data set. The specific formula can be expressed as: Among them, s represents the abnormal evaluation index value, x t Indicates the first step phase feedback data, Represents the first step phase prediction feedback data. Here, the user's actual gait feedback data is compared with the gait prediction feedback data generated by the prediction model, and the second-order moment between the two is calculated to determine the abnormal evaluation index value. This index value reflects the degree of deviation between the user's actual gait and the predicted gait, providing a basis for subsequent feedback adjustments.

[0055] For example, during the experiment, the system recorded the first step phase feedback data when the user was walking on flat ground, and generated the first step phase prediction feedback data through the prediction model. Assuming that at a certain moment, the first step phase feedback data value is (50Nm, 20Nm), and the prediction feedback data value is (48Nm, 19Nm). By calculating the second-order moment between the two, the system obtains the abnormal evaluation index value of ||(2,1)||²=5. This value is used to reflect the deviation between the actual gait and the predicted gait. The system can make subsequent feedback adjustments based on this index value to ensure the optimization effect of the exoskeleton device and the walking stability of the user.

[0056] Step 204: adjust the first feedback value according to the abnormal evaluation index value so as to reduce the abnormal evaluation index value.

[0057] The method shown in this step has been explained in step 103 and will not be repeated here.

[0058] Optionally, step 204 includes the following sub-steps: Sub-step 2041, bringing the abnormal evaluation index value into a preset weight function to obtain the abnormal weight function value of the user.

[0059] In some embodiments of the present application, the abnormal evaluation index value is brought into a preset weight function to obtain the user's abnormal weight function value. Specifically, by analyzing the abnormal evaluation index value and inputting it into the preset weight function, the weight function calculates the corresponding abnormal weight function value according to the size of the abnormal evaluation index value. This process converts the degree of abnormality into a weight value that is convenient for subsequent processing through numerical calculation, providing a basis for the system to further optimize the feedback of the exoskeleton.

[0060] For example, during the experiment, the system detected that the abnormal evaluation index value of the user during walking was 5. By inputting this value into the preset weight function, the weight function calculates an abnormal weight function value based on the abnormal evaluation index value. Assume that the calculated abnormal weight function value is 0.8. This abnormal weight function value reflects the relative degree of the abnormal evaluation index value and is used to guide the system to adjust the feedback value in subsequent steps to ensure that the exoskeleton device can better adapt to the actual needs of the user.

[0061] Sub-step 2042: bringing the abnormal weight function value into a preset impedance model to obtain an updated value for the first feedback value, and updating the first feedback value by the updated value.

[0062] In some embodiments of the present application, the abnormal weight function value is brought into a preset impedance model to obtain an updated value for the first feedback value, and the first feedback value is updated by the updated value. Specifically, by inputting the calculated abnormal weight function value into the preset impedance model, the model adjusts the current impedance parameter according to the weight function value to generate an updated value for the first feedback value. Then, the system applies this updated value to the actual mechanical feedback control to optimize the power-assisting effect of the exoskeleton and reduce abnormalities during walking.

[0063] For example, during the experiment, the system detected that the user's abnormal evaluation index value was 5, and calculated the abnormal weight function value to be 0.8 through the preset weight function. This weight function value is input into the preset impedance model, and the model adjusts the impedance parameters according to the weight value to generate an updated value for the current feedback value. Assuming that the initial first feedback value is 50Nm, the updated value becomes 55Nm after adjustment. The system applies the updated feedback value to the user's walking process, and adjusts the exoskeleton's power assistance strength in real time, thereby reducing the occurrence of abnormalities and improving the power assistance effect and the user's walking stability.

[0064] Step 205: Generate a second feedback value of a second mechanical feedback between the user's lower limbs and the exoskeleton in a second traveling state according to the obtained predicted trajectory, and adjust the second feedback value according to the abnormal evaluation index value.

[0065] In some embodiments of the present application, a second feedback value of the second mechanical feedback between the user's lower limbs and the exoskeleton in the second traveling state is generated based on the obtained predicted trajectory. Specifically, the system obtains the predicted trajectory of the user in the second traveling state through the conversion model, and then generates the mechanical feedback data of the user in the second traveling state based on the predicted trajectory. Then, according to the previously calculated abnormal evaluation index value, the second feedback value is adjusted in real time to reduce the possibility of abnormality and optimize the power assistance effect.

[0066] For example, during the experiment, when a tester is walking up and down stairs, the system will use the predicted trajectory to generate the user's second feedback value. If during the experiment, the system detects that the second feedback value deviates from the expected trajectory and the abnormal evaluation index value is high, the system will adjust the second feedback value in real time, such as adjusting the support torque or posture angle, to ensure that the abnormal evaluation index value is reduced. In this way, the system can effectively reduce the user's movement abnormalities when going up and down stairs, and improve the exoskeleton assistance effect and walking stability.

[0067] Optionally, in some embodiments of the present application, an update value of the first feedback value and / or the second feedback value may be obtained by the following impedance model: ; Among them, C d represents the preset first expected impedance model parameter, K d represents the preset second expected impedance model parameter, q represents the joint angular torque, q d represents the desired joint angular torque, τ e represents the human-computer interaction torque, w(s) represents the abnormal weight function value, represents the first-order derivative of the joint angular torque q with respect to time, Denotes the expected joint angular torque q d The first derivative with respect to time.

[0068] In the embodiment of the present application, the feedback value is dynamically adjusted by the above formula to optimize the assistance effect of the exoskeleton in real time. d and K d As well as the abnormal weight function value w(s), ensure the rationality of the power assistance intensity and feedback value. By controlling the joint angle and its change rate, it can better adapt to the actual needs of users and improve the responsiveness and applicability of the system.

[0069] like Figure 5 As shown, in a specific embodiment of the present application, the specific impact on mechanical feedback is achieved through steps R1 to R4: Step R1: The proprioceptive data of the first phase of gait and torque are compressed into a compact representation. The purpose of this step is to reduce the dimensionality of the data to make it easier to process and analyze.

[0070] Step R2: The compressed data is input into the prediction model for encoding (Encoder) and decoding (Decoder), and the representation of the embedded space (Latent space) is generated through reparameterization. Then, the representation of the embedded space is used by the decoder to generate the predicted value of the original data.

[0071] Step R3: The generated prediction data is used to evaluate the abnormal situation. The abnormal evaluation index value is determined by human-in-the-loop optimization (HIL optimization), and a time series graph is generated to show the change of the abnormal situation.

[0072] Step R4: Based on the abnormal evaluation index value, generate a weight function w(s) and apply it to the impedance model (Variable impedance) to adjust the system response.

[0073] In summary, in the embodiment of the present application, by obtaining the first mechanical feedback of the user in the first walking state, the first step phase feedback data of the user is generated, and the corresponding relationship between the user's lower limb posture and the mechanical feedback is recorded in real time, which provides basic data for subsequent trajectory optimization; then, according to the step phase feedback data, the abnormal evaluation index value of the first walking behavior is calculated, and by analyzing the difference between the step phase feedback data and the normal gait, the abnormality that may exist in the user during the power-assisting process is identified, and then the step phase prediction feedback data of the user in the target walking state is generated to provide a reference standard, and by comparing with the actual feedback data, the abnormal detection mechanism is further improved, and finally, according to the abnormal evaluation index value, the first feedback value is adjusted to reduce the abnormal evaluation index value, and the power-assisting trajectory is optimized in real time by dynamically adjusting the feedback value, ensuring that the exoskeleton device is adjusted according to the actual needs of the user, and improving the power-assisting effect. Therefore, based on the method of the embodiment of the present application, through the real-time adjustment of abnormal detection and feedback values, the personalization and real-time optimization of the exoskeleton device are realized, the responsiveness and applicability of the power-assisting device are improved, the deficiencies of the prior art in real-time optimization and personalized customization are made up, and the problems of the single power-assisting mode and insufficient general power-assisting effect existing in the prior power-assisting device are solved.

[0074] refer to Figure 6 , which shows a lower limb exoskeleton trajectory customization device 30 based on a generative model provided in an embodiment of the present application, comprising: The data acquisition module 301 is used to generate the user's first step phase feedback data in response to the first mechanical feedback between the user's lower limbs and the exoskeleton during the first walking process of the user in the first walking state; the first step phase feedback data is used to characterize the corresponding relationship between different lower limb postures of the user and the first feedback value of the first mechanical feedback during the first walking process; The evaluation module 302 is used to determine an abnormal evaluation index value for the first walking behavior according to the first step phase feedback data; the abnormal evaluation index value is positively correlated with the step phase prediction feedback data for the first step phase feedback data; the step phase prediction feedback data is the target step phase feedback data of the target mechanical feedback between the user's lower limbs and the exoskeleton in the target walking state; The feedback module 303 is used to adjust the first feedback value according to the abnormal evaluation index value so as to reduce the abnormal evaluation index value.

[0075] Optionally, the data acquisition module 301 includes the following submodules: a collection submodule, configured to obtain a plurality of gait postures of the user in the first walking state and a mechanical collection value in each gait posture in response to a first mechanical feedback between the lower limbs of the user and the exoskeleton during a first walking process of the user in the first walking state; The compression submodule is used to compress multiple one-to-one corresponding gait postures and mechanical acquisition values ​​into a code of a preset length to obtain the first step feedback data.

[0076] Optionally, the evaluation module 302 includes the following submodules: A first prediction submodule is used to input the first step phase feedback data into the prediction model to obtain the first step phase prediction feedback data for the first step phase feedback data; The evaluation submodule is used to determine the second-order moment of the first-step phase feedback data and the first-step phase prediction feedback data as an abnormal evaluation index value.

[0077] Optionally, the feedback module 303 includes: The weight submodule is used to bring the abnormal evaluation index value into the preset weight function to obtain the abnormal weight function value of the user; The regulating submodule is used to bring the abnormal weight function value into a preset impedance model to obtain an updated value for the first feedback value, and update the first feedback value through the updated value.

[0078] Optionally, the device 30 further includes: A trajectory module, configured to obtain a first moving trajectory of the user in response to a first mechanical feedback between a lower limb of the user and the exoskeleton during a first walking process of the user in a first moving state; a prediction module, configured to input the first travel trajectory into a conversion model to obtain a predicted trajectory of a second travel trajectory of the user in a second travel state; The adjustment module is used to generate a second feedback value of a second mechanical feedback between the user's lower limbs and the exoskeleton in a second traveling state according to the obtained predicted trajectory, and adjust the second feedback value according to the abnormal evaluation index value.

[0079] Optionally, the trajectory module includes the following submodules: a trajectory submodule, configured to obtain a plurality of gait postures of the user in the first walking state and a time corresponding to each gait posture in response to a first mechanical feedback between a lower limb of the user and the exoskeleton during a first walking process of the user in the first walking state; The sequence submodule is used to determine a time sequence generated according to a plurality of one-to-one corresponding gait postures and moments as a first moving trajectory.

[0080] In summary, in the embodiment of the present application, by obtaining the first mechanical feedback of the user in the first walking state, the first step phase feedback data of the user is generated, and the corresponding relationship between the user's lower limb posture and the mechanical feedback is recorded in real time, which provides basic data for subsequent trajectory optimization; then, according to the step phase feedback data, the abnormal evaluation index value of the first walking behavior is calculated, and by analyzing the difference between the step phase feedback data and the normal gait, the abnormality that may exist in the user during the power-assisting process is identified, and then the step phase prediction feedback data of the user in the target walking state is generated to provide a reference standard, and by comparing with the actual feedback data, the abnormal detection mechanism is further improved, and finally, according to the abnormal evaluation index value, the first feedback value is adjusted to reduce the abnormal evaluation index value, and the power-assisting trajectory is optimized in real time by dynamically adjusting the feedback value, ensuring that the exoskeleton device is adjusted according to the actual needs of the user, and improving the power-assisting effect. Therefore, based on the method of the embodiment of the present application, through the real-time adjustment of abnormal detection and feedback values, the personalization and real-time optimization of the exoskeleton device are realized, the responsiveness and applicability of the power-assisting device are improved, the deficiencies of the prior art in real-time optimization and personalized customization are made up, and the problems of the single power-assisting mode and insufficient general power-assisting effect existing in the prior power-assisting device are solved.

[0081] Reference Figure 7 , the electronic device 500 may include one or more of the following components: a processing component 502 , a memory 504 , a power component 506 , a multimedia component 508 , an audio component 510 , an input / output (I / O) interface 512 , a sensor component 514 , and a communication component 516 .

[0082] The processing component 502 generally controls the overall operation of the electronic device 500, such as operations associated with display, phone calls, data communications, camera operations, and recording operations. The processing component 502 may include one or more processors 520 to execute instructions to complete all or part of the steps of the above-mentioned method. In addition, the processing component 502 may include one or more modules to facilitate the interaction between the processing component 502 and other components. For example, the processing component 502 may include a multimedia module to facilitate the interaction between the multimedia component 508 and the processing component 502.

[0083] The memory 504 is used to store various types of data to support the operation of the electronic device 500. Examples of such data include instructions for any application or method operating on the electronic device 500, contact data, phone book data, messages, pictures, multimedia, etc. The memory 504 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0084] The power supply component 506 provides power to the various components of the electronic device 500. The power supply component 506 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to the electronic device 500.

[0085] The multimedia component 508 includes an interface that provides an output interface between the electronic device 500 and the user. In some embodiments, the interface may include a liquid crystal display (LCD) and a touch panel (TP). If the interface includes a touch panel, the interface may be implemented as a touch screen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touch, slide, and gestures on the touch panel. The touch sensor may not only sense the boundaries of the touch or slide action, but also detect the duration and pressure associated with the touch or slide operation. In some embodiments, the multimedia component 508 includes a front camera and / or a rear camera. When the electronic device 500 is in an operating mode, such as a shooting mode or a multimedia mode, the front camera and / or the rear camera may receive external multimedia data. Each front camera and rear camera may be a fixed optical lens system or have a focal length and optical zoom capability.

[0086] The audio component 510 is used to output and / or input audio signals. For example, the audio component 510 includes a microphone (MIC), and when the electronic device 500 is in an operation mode, such as a call mode, a recording mode, and a voice recognition mode, the microphone is used to receive an external audio signal. The received audio signal can be further stored in the memory 504 or sent via the communication component 516. In some embodiments, the audio component 510 also includes a speaker for outputting audio signals.

[0087] The input / output I / O interface 512 provides an interface between the processing component 502 and the peripheral interface module, which may be a keyboard, a click wheel, buttons, etc. These buttons may include but are not limited to: a home button, a volume button, a start button, and a lock button.

[0088] The sensor assembly 514 includes one or more sensors for providing various aspects of status assessment for the electronic device 500. For example, the sensor assembly 514 can detect the open / closed state of the electronic device 500, the relative positioning of components, such as the display and keypad of the electronic device 500, and the sensor assembly 514 can also detect the position change of the electronic device 500 or a component of the electronic device 500, the presence or absence of user contact with the electronic device 500, the orientation or acceleration / deceleration of the electronic device 500, and the temperature change of the electronic device 500. The sensor assembly 514 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact. The sensor assembly 514 may also include an optical sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor assembly 514 may also include an accelerometer, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

[0089] The communication component 516 is used to facilitate wired or wireless communication between the electronic device 500 and other devices. The electronic device 500 can access a wireless network based on a communication standard, such as WiFi, a carrier network (such as 2G, 3G, 4G or 5G), or a combination thereof. In an exemplary embodiment, the communication component 516 receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 516 also includes a near field communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology and other technologies.

[0090] In an exemplary embodiment, the electronic device 500 may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to implement the methods provided in the embodiments of the present application.

[0091] In an exemplary embodiment, a non-transitory computer-readable storage medium including instructions is also provided, such as a memory 504 including instructions, and the instructions can be executed by a processor 520 of an electronic device 500 to perform the above method. For example, the non-transitory storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, etc.

[0092] Figure 8FIG. 6 is a block diagram of an electronic device 600 according to another embodiment of the present invention. For example, the electronic device 600 may be provided as a server.

[0093] Reference Figure 8 , the electronic device 600 includes a processing component 622, which further includes one or more processors, and a memory resource represented by a memory 632, for storing instructions that can be executed by the processing component 622, such as an application. The application stored in the memory 632 may include one or more modules, each corresponding to a set of instructions. In addition, the processing component 622 is configured to execute instructions to perform the method provided in the embodiment of the present application.

[0094] The electronic device 600 may also include a power supply component 626 configured to perform power management of the electronic device 600, a wired or wireless network interface 650 configured to connect the electronic device 600 to a network, and an input / output (I / O) interface 658. The electronic device 600 may operate based on an operating system stored in the memory 632, such as Windows Server™, Mac OS X™, Unix™, Linux™, FreeBSD™ or the like.

[0095] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the application disclosed herein. The present application is intended to cover any modification, use or adaptation of the present application, which follows the general principles of the present application and includes common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present application are indicated by the following claims.

[0096] It should be understood that the present application is not limited to the precise structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A method for customizing lower limb exoskeleton trajectories based on a generative model, characterized in that: include: In response to first mechanical feedback between the lower limbs of the user and the exoskeleton during a first walking process of the user in a first traveling state, first step phase feedback data of the user is generated; the first step phase feedback data is used to characterize the corresponding relationship between different lower limb postures of the user and the first feedback value of the first mechanical feedback during the first walking process; According to the first step phase feedback data, an abnormal evaluation index value for the first walking behavior is determined; the degree of difference between the abnormal evaluation index value and the step phase prediction feedback data for the first step phase feedback data is positively correlated; the step phase prediction feedback data is the target step phase feedback data of the target mechanical feedback between the lower limbs of the user and the exoskeleton in the target walking state of the user; According to the abnormal evaluation index value, the first feedback value is adjusted so that the abnormal evaluation index value is reduced.

2. The method according to claim 1, characterized in that The step of generating the first step phase feedback data of the user in response to the first mechanical feedback between the lower limbs of the user and the exoskeleton during the first walking process of the user in the first walking state comprises: In response to first mechanical feedback between the lower limbs of the user and the exoskeleton during a first walking process of the user in a first walking state, acquiring a plurality of gait postures of the user in the first walking state and a mechanical acquisition value in each of the gait postures; A plurality of one-to-one corresponding gait postures and mechanical acquisition values ​​are compressed into a code of a preset length to obtain the first step phase feedback data.

3. The method according to claim 1, characterized in that Determining the abnormal evaluation index value of the first walking behavior according to the first phase feedback data includes: Inputting the first-phase feedback data into a prediction model to obtain first-phase prediction feedback data for the first-phase feedback data; The second-order moment of the first-phase feedback data and the first-phase predicted feedback data is determined as the abnormality evaluation index value.

4. The method according to claim 1, characterized in that The adjusting the first feedback value according to the abnormal evaluation index value so as to reduce the abnormal evaluation index value includes: Substituting the abnormal evaluation index value into a preset weight function to obtain the abnormal weight function value of the user; The abnormal weight function value is brought into a preset impedance model to obtain an updated value for the first feedback value, and the first feedback value is updated by the updated value.

5. The method according to claim 4, characterized in that The impedance model is: ； Among them, C d represents the preset first expected impedance model parameter, K d represents the preset second expected impedance model parameter, q represents the joint angular torque, q d represents the desired joint angular torque, τ e represents the human-computer interaction torque, w(s) represents the abnormal weight function value, represents the first-order derivative of the joint angular torque q with respect to time, Denotes the expected joint angular torque q d The first derivative with respect to time.

6. The method according to claim 1, characterized in that The method further comprises: In response to a first mechanical feedback between a lower limb of the user and the exoskeleton during a first walking process of the user in a first walking state, acquiring a first moving trajectory of the user; inputting the first traveling trajectory into a conversion model to obtain a predicted trajectory of a second traveling trajectory of the user in a second traveling state; A second feedback value of a second mechanical feedback between the lower limbs of the user and the exoskeleton in the second traveling state is generated according to the obtained predicted trajectory, and the second feedback value is adjusted according to the abnormal evaluation index value.

7. The method according to claim 1, characterized in that The step of acquiring a first moving trajectory of the user in response to a first mechanical feedback between the lower limbs of the user and the exoskeleton during a first walking process of the user in a first moving state comprises: In response to first mechanical feedback between a lower limb of the user and an exoskeleton during a first walking process of the user in a first walking state, acquiring a plurality of gait postures of the user in the first walking state and a time corresponding to each of the gait postures; A time sequence generated according to a plurality of one-to-one corresponding gait postures and moments is determined as the first moving trajectory.

8. A lower limb exoskeleton trajectory customization device based on a generative model, characterized in that: include: a data acquisition module, configured to generate first step phase feedback data of the user in response to first mechanical feedback between the lower limbs of the user and the exoskeleton during a first walking process of the user in a first walking state; the first step phase feedback data is used to characterize the corresponding relationship between different lower limb postures of the user and a first feedback value of the first mechanical feedback during the first walking process; An evaluation module is used to determine an abnormal evaluation index value for the first walking behavior according to the first step phase feedback data; the abnormal evaluation index value is positively correlated with the step phase prediction feedback data for the first step phase feedback data; the step phase prediction feedback data is the target step phase feedback data of the target mechanical feedback between the lower limbs of the user and the exoskeleton in the target walking state of the user; A feedback module is used to adjust the first feedback value according to the abnormal evaluation index value so that the abnormal evaluation index value decreases.

9. An electronic device, characterized in that: include: A processor, a memory for storing instructions executable by the processor; The processor is configured to execute the instructions to implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: When the instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device is enabled to perform the method as claimed in any one of claims 1 to 7.

Images