A device for controlling training devices.
The training device uses sensors and an optimization algorithm to adjust support based on individual data, ensuring optimal effort and safety by minimizing stress.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- BITIFEYE DIGITAL TEST SOLUTIONS GMBH
- Filing Date
- 2022-02-11
- Publication Date
- 2026-05-19
AI Technical Summary
Training devices often fail to adjust to individual user capabilities, leading to insufficient effort or excessive stress, which can result in physical harm.
A training device equipped with a support unit, motion and body sensors, and a computing unit that uses an optimization algorithm to adjust the support based on individual physiological and mechanical data to maintain optimal effort levels.
The device effectively adjusts to individual responses to load changes, ensuring sufficient effort while minimizing excessive stress, thereby improving fitness safely.
Smart Images
Figure 0007862006000019 
Figure 0007862006000020 
Figure 0007862006000021
Abstract
Description
[Technical Field]
[0001] The present invention relates to a device for controlling a training device. [Background technology]
[0002] Numerous training devices exist that allow people to train and thereby improve their physical fitness. Electric bicycles are one example. Other examples include bicycle ergometers, abduction / adduction machines, and arm strengthening traction devices. During training sessions, it is crucial that training improves fitness while simultaneously avoiding excessive stress that could cause physical harm. It must be noted that the optimal stress level can vary significantly from person to person. It is important that training devices are used correctly or appropriately adjusted during training so that a person exerts sufficient effort without simultaneously experiencing excessive stress. Ideally, training devices should be designed to be adjustable for both people with weak hearts and those performing high-intensity exercise. An example of incorrect use of a training device would be setting the motor power of an electric bicycle too high. As a result, the person rides at a relatively high speed without exerting sufficient effort, simultaneously increasing the risk of accidents. [Overview of the project] [Problems that the invention aims to solve]
[0003] Therefore, an object of the present invention is to provide a device equipped with a training device that is controlled in such a way that a person training with the training device can make sufficient effort while simultaneously avoiding excessive stress on the person. [Means for solving the problem]
[0004] The apparatus according to the present invention for controlling a training device is - A training device configured to absorb the mechanical force applied by a person undergoing physical training, comprising a support unit configured to assist and / or make the training more difficult, and a motion measuring device configured to measure mechanical motion data BD(t) of the effort applied by the person during training, where t is time, the training device and - A body sensor configured to measure physiological data PD(t) of the human body, - A computing unit that stores a mathematical model in the form of mPD(t+T), and is configured to individually adjust mPD(t+T) and delay T for each person using an optimization algorithm so that mPD(t+T) approaches the measured physiological data PD(t+T), and to prepare a predicted mPD(t+T) of the physiological data PD(t+T) based on the model, - A control unit configured to take a predetermined reference variable for physiological data PD(t), take a predicted mPD(t+T) as a control variable, and control the support u(t) of the support unit as an manipulated variable. Includes.
[0005] formula:
number
[0006] Since the device uses the prediction mPD(t+T) where the time t+T is in the future by the delay time T as a control variable, the control unit can respond much more quickly to changes in training than it would if it used the physiological data PD(t) as a control variable. As a result, the control deviation of the control variable from the reference variable can be kept much lower than it would be if the physiological data PD(t) were used as a control variable. The control unit individually determines the coefficient a xi , the addend a 10 , the delay τ xi and the delay T so that the control deviation can be kept low for everyone. Each person responds to changes in the load acting on the person, for example generated by a training device, from the outside at different speeds. In the case of a person who has relatively little training, that person tends to respond slowly to changes, while in the case of a person who has relatively more training, that person responds relatively quickly to changes. The arithmetic unit is configured to adjust not only the coefficient a xi and the addend a 10 but also the delay τ xi and the delay T, so that the model can reflect the fact that each person responds to changes in the load at different speeds. As a result, the prediction has particularly high accuracy for each person, and thereby the control deviation is also particularly low. All that is still necessary is to identify the appropriate reference variable for each person's physiological data PD(t), while it can be assumed that the reference variable changes over time. For example, a sports doctor or a physiotherapist can be employed to set the reference variable. Since the control deviation is particularly low, it is possible here to control the training device so that, due to the person making sufficient effort, the person's physical strength improves and excessive stress on the person is avoided.
[0007] Support u(t) can be positive, thereby assisting training, and / or negative, thereby making training more difficult. The motor of an electric bicycle is an example of a support unit configured to assist training. In this case, the support may be, for example, the power applied by the motor. The brakes of a bicycle ergometer are an example of a support unit configured to make training more difficult. In this case, the support may be, for example, braking force. An example of a support unit configured to assist training and make training more difficult is the motor of an electric bicycle, which is configured to perform recovery, i.e., convert the person's pedaling force into an electric current. To keep the control deviation particularly low, it is preferable that the support unit be configured to control the support u(t) to a small increase. The increase may be, for example, up to 3%, specifically up to 1.5%, or up to 1%. 100% represents the maximum support u(t) when the support unit is configured to assist training. -100% corresponds to the maximum resistance to training when the support unit is configured to make training more difficult.
[0008] Exercise data BD(t) is characterized by the mechanical effort applied by a person during training to overcome a load. Exercise data BD(t) is zero when the person is at rest. Physiological data is characterized by how the systems and / or subsystems of the person's body function and includes variables that can be measured by sensors. The system or subsystem may be the cardiopulmonary system or a part thereof, or the musculoskeletal system or a part thereof. Physiological data PD(t) may be, for example, heart rate. There are several variables, such as knee adduction torque and / or knee abduction torque, which can cause problems with both exercise data BD(t) and physiological data PD(t).
[0009] It is preferable for j to be selected from the range of 2 to 5. We found that while j=2 requires only low computational power, it still achieves sufficient predictive accuracy, whereas j=5 achieves even higher predictive accuracy.
[0010] It is preferable to select k from the range of 1 to 4. We found that while k=1 requires only low computational power, it still achieves sufficient prediction accuracy, whereas k=4 achieves even higher prediction accuracy.
[0011] The training device preferably includes an altimeter configured to measure the altitude h(t) of the training device.
number
[0012] The training device preferably includes a temperature sensor for measuring the temperature Temp(t) around the training device, within the model,
number
[0013] The training device preferably includes an inclinometer configured to measure the inclination N(t) of the training device, within the model,
number
[0014] Delay τ xi is zero, and all delays τ xi It is preferable that i>1 be adjusted. The arithmetic unit is preferably configured to prepare a predicted mPD(t+T) for a time t+T that is at least T=5 seconds in the future.
[0015] The calculation unit, in order to take into account a person's basic physical fitness, uses the exercise data BD(t) and physiological data PD(t) confirmed in multiple training sessions, as well as optionally the altitude h(t), temperature Temp(t), and / or slope N(t) confirmed in multiple training sessions, to determine a coefficient a based on an optimization algorithm. xi , summand a 10 , delay τ xi It is preferable that the delay T be adjusted. Multiple training sessions could be, for example, all training sessions performed by a person. Alternatively, multiple training sessions could be a large number of training sessions performed very recently.
[0016] The computing unit uses the optimization algorithm 11 to calculate the coefficient a after the training session. xi , summand a 10 , delay τ xi It is preferable that the optimization algorithm 11 is configured to adjust the delay T, and a) coefficient a xi Each of the delays τ xi , summand a 10 And for delay T, a step of identifying multiple discrete values in each case, b)a xi a 10 , τ xi Steps a) and setting T to one of its values, c) calculating mPD(t+T) based on the model, d) calculating the modeling error between the measured physiological data PD(t+T) and mPD(t+T) for multiple t values, e) repeating steps b) to d) for all combinations of values, f) a xi a 10 , τ xi And for T, there is a step of selecting those values that yield the minimum modeling error. This is a way of focusing on calculations, while the value a xi a 10 , τ xi And T can nevertheless be determined with high accuracy, so the control deviation is particularly small. In step d), it is particularly preferable that the underestimation error is given more weight than the overestimation error.
[0017] The calculation unit uses an algorithm to adjust for the current physical fitness level, taking into account the person's current physical fitness, by using the exercise data BD(t) and physiological data PD(t) observed during the training session, as well as optionally observed altitude h(t), temperature Temp(t), and / or slope N(t) during the training session, to calculate the coefficient a. xi and augment a 10It is preferable that it be configured to adjust the control deviation. The control deviation can be kept particularly low by taking into account the current physical strength.
[0018] The calculation unit uses an algorithm to adjust for current physical fitness to determine the difference Diff(t) = mPD(t) - PD(t) between the predicted physiological data mPD(t) and the measured physiological data PD(t), and if the difference Diff(t) exceeds the threshold Threshold1>0, the respective constants const 1xi By adding the coefficient a xi Modify and constant const 10 By adding the augment a 10 Correct the difference Diff(t) is Threshold M If the threshold is <0, each constant const Mxi By adding the coefficient a xi Modify and constant const M0 By adding the augment a 10 It is particularly preferable to configure it to modify the following. This is advantageous because it is less computationally intensive and is also suitable for execution during training sessions. It can also provide more thresholds. A suitable program code would look something like this, for example. if(Diff(t)>Threshold1) a x1 =a x1 + 1x1 a x2 =a x2 + 1x2 , ... elseif(Diff(t)>Threshold2) a x1 =a x1 + 2x1 , a x2 =a x2 + 2x2 , ... ... elseif(Diff(t)>Threshold M-1 ) a x1 =a x1 + (M-1)x1 , a x2 =a x2 + (M-1)x2 , ... elseif(Diff(t) <Threshold M ) a x1 =a x1 + Mx1 , a x2 =a x2 + Mx2 , ... ... elseif(Diff(t) <Threshold M+K ) a x1 =a x1 + (M+K)x1 , a x2 =a x2 + (M+K)x2 , ... End
[0019] Here, each has an if query, and coefficient a xi and augment a 10 All of the above has been corrected, and the following applies: Threshold1>Threshold2>...>Threshold M-1 >Threshold M+K >...>Threshold M+1 >Threshold M
[0020] The control unit is preferably a PID controller. The PID controller is particularly suitable for controlling the physiological data PD(t) because its integral term contributes to gradually reducing the control deviation, while its derivative term enables it to exceed the control deviation even before the control deviation actually occurs. The PID controller is [Number] is particularly preferably configured here to determine the assistance u(t) according to, where K P , K I and K D are control parameters, e(t) is the control deviation at time t, and the functions f1(e), f2(e) and f3(e) are selected such that the underestimation error is emphasized more strongly than the overestimation error. As a result, the deviation of the control variable from the reference variable is less likely than the deviation of the control variable having a value smaller than the reference variable. Thereby, it is possible to avoid excessive stress that may cause physical injury to a person. Particularly preferably, [Number] where f3(e) = 0 for e < 0 and f3(e) = e for e ≥ 0, while for f1(e) and f2(e), the polynomials may vary depending on the range of e.
[0021] The arithmetic unit is particularly preferably configured to adjust the control parameters K P , K I and K D individually for each person. In this way, it is possible to achieve a particularly low control deviation for each person.
[0022] The arithmetic unit is preferably configured to execute a calibration method in which the step response of the physiological data PD(t) is generated by a rapid change in the manipulated variable, and the arithmetic unit determines the control parameters K P , K I and K DIt is preferably configured to determine. The arithmetic unit can be configured to record physiological data PD(t) in order to continuously generate a step response. The arithmetic unit is configured to switch the support unit from a certain first support u1 to a certain second support u2 at time T0 in order to cause a rapid change in the variable operated thereby. For example, u1 can be 80% to 100%, and u2 can be 0% to 20%. It is possible to show here information indicating that a person should train at a frequency as constant as possible, such as the number of times of stepping on a pedal. The arithmetic unit is configured to wait for a sufficient time for the physiological data PD(t) to stabilize at approximately the value of PD1 before the switch and at approximately the value of PD2 after the switch both during the first support u1 and during the second support u2. The arithmetic unit can be configured to wait for at least 2 minutes both before and after the switch. It can further be assumed that the arithmetic unit is configured to generate a second step response. For this purpose, after the movement data BD(t) or the physiological data PD(t) stabilizes following a rapid change in the operated variable, the arithmetic unit can be configured to switch the support from u2 to u1 and wait again until the physiological data PD(t) stabilizes.
[0023] The arithmetic unit is configured to identify at least one rapid change in the operated variable and the resulting step response of the physiological data PD(t) after a training session, and the arithmetic unit determines control parameters K P , K I and K D from at least one such step response. It is preferably configured to determine. The arithmetic unit uses a calibration method to coarsely adjust the control parameters K P , K I and K D , and it can be assumed that it is configured to use at least one such step response identified outside the calibration method after a training session to finely adjust the control parameters K P , K I and K D .
[0024] Preferably, the motion data BD(t) includes power, specifically in the case of a bicycle, specifically in the case of an electric bicycle or bicycle ergometer, pedaling force, running force, rowing force, speed, torque, rotational speed, angular velocity and / or knee abduction torque.
[0025] The support unit preferably includes an electric motor, a transmission, and / or brakes.
[0026] Physiological data PD(t) preferably includes heart rate, heart rate variability, electrocardiogram, blood oxygen saturation, blood pressure, nerve activity, specifically electroencephalogram, knee abduction torque, adduction, specifically knee adduction and / or knee flexion and extension.
[0027] The present invention will be described in more detail below with reference to the attached schematic diagram. [Brief explanation of the drawing]
[0028] [Figure 1] A schematic diagram of the apparatus according to the present invention is shown. [Figure 2] Details of the overview diagram according to the present invention are shown. [Figure 3] The graphs of f1(e) and f2(e) are shown. [Figure 4] The graph of f3(e) is shown. [Figure 5] This graph shows the step response of the physiological data PD(t) caused by abrupt changes in the manipulated variable. [Figure 6] This shows graphs of various measurement variables recorded during the training session. [Modes for carrying out the invention]
[0029] Figures 1 and 2 show the device 1 for controlling the training device 2. - A training device 2 configured to absorb mechanical force 9 applied by a person 8 undergoing physical training, and including a support unit 6 configured to assist and / or make training more difficult, and including a motion measuring device 5 configured to measure mechanical motion data BD(t) of the effort applied by the person during training, where t is time, the training device 2 and - A body sensor 7 configured to measure the physiological data PD(t) of a person 8, - Format
number
number
[0030] Training device 2 may include an altimeter configured to measure the altitude h(t) of training device 2, within the model,
number
[0031] The control unit can be, for example, a PID controller. The PID controller can be configured to determine the assistance u(t) according to, for example, [Number] where K P , K I and K D are control parameters, e(t) is the control deviation at time t, and the functions f1(e), f2(e) and f3(e) are selected such that the underestimation error is weighted more strongly than the overestimation error. Here, [Number] where f3(e) = 0 for e < 0 and f3(e) = e for e ≥ 0, while for f1(e) and f2(e), the polynomials can vary depending on the range of e. Figure 3 shows an exemplary graph of f1(e) = f2(e), and Figure 4 shows an exemplary graph of f3(e). As can be seen from Figure 3, the functions f1(e) and f2(e) can have a bisector and are only above the bisector within each range of 0 < e < E1 or 0 < e < E2. Specifically, when the physiological data is the heart rate, for example, f1(e) = f2(e) = e for e > 12 or e < 0, and f1(e) = f2(e) = 2 * e - 0.082 * e2 This can be applied. As can be seen from Figure 4, for example, f3(e) can be defined by f3(e)=e for e>0 and f3(e)=0 for e≦0.
[0032] The calculation unit 3 controls the control parameter K individually for each person 8. P , K I and K D It can be predicted that it will be configured to adjust the control parameter K. For this purpose, the calculation unit 3 can be configured to perform a calibration method in which the step response of physiological data PD(t) occurs at time T0 due to a rapid change in the manipulated variable, and the calculation unit 3 adjusts the control parameter K from the step response. P , K I and K DIt is configured to determine the following. An exemplary step response is illustrated in Figure 5. The calculation unit 3 can be configured to record physiological data PD(t) to generate step responses sequentially. The calculation unit 3 can be configured to switch the support unit 6 from a constant first support u1 to a constant second support u2 in order to cause abrupt changes in the variable being manipulated. For example, u1 may be 80% to 100% and u2 may be 0% to 20%. Information can be presented here indicating that a person should train at as constant a frequency as possible, for example, the number of times they press the pedals. The calculation unit 3 can be configured to wait for a sufficient amount of time during both the first support u1 and the second support u2 for the physiological data PD(t) to stabilize at approximately the value of PD1 before the switch and at approximately the value of PD2 after the switch. The calculation unit 3 can be configured to wait for at least 2 minutes both before and after the switch. To determine the control parameters from the step response, the calculation unit 3 can be configured to apply an inflection tangent 13 to the step response. Before applying the inflection tangent 13, PD(t) can be adjusted by a function, such as a polynomial, and the inflection tangent 13 can be applied to the adjusted function. At least the squared error method can be used to adjust the function. The intersection of the inflection tangent 13 and PD(t)=PD1 is at a delay time T starting at T0. u Determined, the intersection of the inflection tangent 13 and PD(t)=PD2 is T u The starting time T at the end G This determines the control parameter, for example, K. P =1.2 * T G / (K S * T U ), K I =0.6 * T G / (K S * (T U ) 2 ) and K D =0.6 * T G / K SThis can be determined by, where K S This is the amplification constant, and can be calculated as the ratio of the change in the control parameter to the change in the support parameter.
[0033] The calculation unit is configured to identify a step response from which at least one abrupt change in the manipulated variable and physiological data PD(t) or motor data BD(t) after the training session are obtained, and the calculation unit identifies a control parameter K from that at least one step response. P , K I and K D It can be assumed that the calculation unit is configured to determine the control parameter K. P , K I and K D A calibration method is used to roughly adjust the control parameter K P , K I and K D To fine-tune it, it could also be configured to use at least one step response identified outside of the calibration method after the training session.
[0034] The computation unit may further be configured to generate a second step response. For this purpose, the computation unit may be configured to switch support from u2 to u1 after the physiological data PD(t) stabilizes following a rapid change in the manipulated variable, and then wait again until the motor data BD(t) or physiological data PD(t) stabilizes. Control parameter K P , K I and K D This may change as the support u(t) increases or decreases.
[0035] The calculation unit 3, in order to consider the basic physical fitness of person 8, uses the exercise data BD(t) and physiological data PD(t) confirmed in multiple training sessions, as well as the altitude h(t), temperature Temp(t), and / or slope N(t) confirmed in multiple training sessions, which are optionally confirmed in multiple training sessions, to determine the coefficient a based on the optimization algorithm 11 (see Figure 2). xi , summand a 10 , delay τ xi It can also be configured to adjust the delay T. For this purpose, the computing unit 3 uses the optimization algorithm 11 to adjust the coefficient a after the training session. xi , summand a 10 , delay τ xi The optimization algorithm 11 can be configured to adjust the delay T, and the optimization algorithm 11 is a) coefficient a xi Each of the augends a 10 , delay τ xi For each of the and delay T, a step of identifying multiple discrete values in each case, b) a xi a 10 , τ xi Steps a) and setting T to one of the values, c) calculating mPD(t+T) based on the model, d) calculating the modeling error between the measured physiological data PD(t+T) and mPD(t+T) for multiple t values, e) repeating steps b) to d) for all combinations of values, f) a xi a 10 , τ xi The process involves selecting values for T that result in the lowest modeling error. In step d), underestimation error may be given more weight than overestimation error.11
[0036] As can be seen in Figure 2, the calculation unit 3 uses an algorithm 12 to adjust the coefficient a, taking into account the current physical fitness of person 8, using the exercise data BD(t) and physiological data PD(t) confirmed during the training session, as well as optionally confirmed altitude h(t), temperature Temp(t), and / or slope N(t) confirmed during the training session. xi and augment a 10 It can be configured to adjust the following. For this purpose, the computing unit, for example, uses algorithm 12 for adjusting the current physical strength to determine the difference Diff(t) = mPD(t) - PD(t) between the predicted mPD(t) of the physiological data and the measured physiological data PD(t), and if the difference Diff(t) exceeds the threshold Threshold1>0, the respective constant const 1xi By adding the coefficient a xi Modify and constant const 10 By adding the augment a 10 Correct the difference Diff(t) is Threshold M If the threshold is <0, each constant const Mxi By adding the coefficient a xi Modify and constant const M0 By adding the augment a 10 It can be configured to modify it.
[0037] Coefficient a confirmed by optimization algorithm 11 xi And delay τ xi and the coefficient a confirmed by algorithm 12 for adjusting T and the current form xi The augmented number a, confirmed by the optimization algorithm 11 and the algorithm for checking the current physical strength. 10 This is used in step 10 to prepare the predicted mPD(t+T). The predicted mPD(t+T) is a control variable within control unit 4, and the manipulated variable is the support u(t).
[0038] The exercise data BD(t) may be, for example, power, specifically pedal force, running force, rowing force, speed, torque, rotational speed, angular velocity, and / or knee abduction torque in the case of a bicycle, specifically an electric bicycle or bicycle ergometer. If the training device 2 is a bicycle or bicycle ergometer, the power 9 applied by person 8 during training and absorbed by the training device 2 is pedal force. The training device 2 may also be, for example, a rowing ergometer or a rowboat, and the exercise data may be rowing force. The training device may also be an abduction / adduction machine, and the exercise data may be knee abduction torque.
[0039] The support unit 6 may include, for example, an electric motor, a transmission, and / or a brake. The support u(t) provided by the support unit 6 may be positive, thereby assisting the training, and / or negative, thereby making the training more difficult. An electric motor is an example of a support unit 6 configured to assist training. In this case, the support u(t) may be, for example, power provided by the electric motor. If the motion data BD(t) is power, the control unit 4 will, P M (t)=u(t) * K * The power P of the electric motor according to BD(t) MAlternatively, it can be assumed that the unit is configured to determine the maximum motor assistance. Factor K indicates the maximum motor assistance possible. K can be, for example, 1 to 5, specifically 3. For example, the brakes on a bicycle ergometer are an example of an assistance unit configured to make training more difficult. In this case, the assistance could be, for example, braking force. An example of an assistance unit configured to support and make training more difficult is an electric motor configured to perform recovery, i.e., to convert the person's pedaling force into an electric current. Assistance unit 6 can be configured to control the assistance u(t) to a small increase. For example, an increase of up to 3%, specifically up to 1.5%, or up to 1%, can be assumed. Here, 100% corresponds to the maximum assistance u(t) when the assistance unit is configured to support training. When the assistance unit is configured to make training more difficult, -100% corresponds to the maximum resistance to training.
[0040] Physiological data PD(t) may include heart rate, heart rate variability, electrocardiogram, blood oxygen saturation, blood pressure, neural activity, specifically electroencephalogram, adduction, specifically knee adduction and / or knee flexion and extension. Adduction and / or knee flexion and extension can be determined, for example, using multiple inertial measurement units attached to the person 8, configured to determine acceleration values and / or rotational data.
[0041] The physiological data PD(t), exercise data BD(t), and support u(t) of a training session performed on an electric bicycle as training device 2 are graphically plotted in Figure 6. Physiological data PD(t) is heart rate in beats per minute (bpm). Heart rate can be measured, for example, by a body sensor 7 attached to a chest strap. Exercise data BD(t) is pedal force in watts. Pedal force can be determined, for example, by measuring torque and angular velocity. In particular, to obtain high-quality torque, the torque in Figure 6 was measured using an Innotorq torque sensor, as described in International Publication No. 2015 / 028345A1. Angular velocity was measured by measuring the rotation of the pole ring using a magnetic sensor. Support unit 6 in Figure 6 is the electric motor of an electric bicycle, whose support is controlled from 0% to 100%. If the motor is capable of recovery, the support can be controlled from -100% to 100%. The dashed line in the upper graph of Figure 6 represents the reference variable. It can be seen that the reference variable may change over time. It can also be seen that the measured heart rate is always a good approximation of the reference variable. [Explanation of symbols]
[0042] 1 device 2 Training devices 3 Processing Units 4. Control Unit 5 Exercise measuring device 6 Support Units 7 Body Sensors 8 people 9 Power 10. Preparation for predicting mPD(t+T) 11 Optimization Algorithms 12. Algorithm for adjusting current physical strength 13. Tangent line at the inflection point BD(t) exercise data PD(t) Physiological Data mPD(t+T) Prediction of physiological data u support t time T uDelay time T v Set time The time of rapid change in T0 support
Claims
1. A device for controlling the training device (2), - A training device (2) configured to absorb mechanical force (9) applied by a person (8) undergoing physical training, comprising a support unit (6) configured to assist and / or make the training more difficult, and a motion measuring device (5) configured to measure mechanical motion data BD(t) of the mechanical force (9) applied by the person during the training, where t is time, the training device (2) and - A body sensor (7) configured to measure the physiological data PD(t) of the person (8), - A computing unit (3) that stores a mathematical model in the form of mPD(t+T), and is configured to individually adjust mPD(t+T) and delay T for each person using an optimization algorithm (11) such that mPD(t+T) approaches the measured physiological data PD(t+T), and to prepare a predicted mPD(t+T) of the physiological data PD(t+T) based on the mathematical model, - A control unit (4) is configured to provide a predetermined reference variable for the physiological data PD(t), take the predicted mPD(t+T) as a control variable, and control the support u(t) of the support unit (6) as an operated variable. A device equipped with the following features. 【Request Item 2】 【Number 1】 The calculation unit (3) then uses the optimization algorithm (11) to individually calculate coefficient a for each person in such a way that mPD(t+T) approaches the measured physiological data PD(t+T). ｘｉ , summand a １０ and delay τ ｘｉ The apparatus according to claim 1, configured to adjust at least partially.
3. The training device (2) includes an altimeter configured to measure the altitude h(t) of the training device (2), and within the mathematical model, [Math 2] The apparatus according to claim 2.
4. The training device (2) includes a temperature sensor configured to measure the temperature Temp(t) around the training device (2), and within the mathematical model, [Math 3] The apparatus according to claim 3.
5. The training device (2) includes an inclinometer configured to measure the inclination N(t) of the training device (2), and within the mathematical model, [Math 4] The apparatus according to claim 4.
6. The apparatus according to any one of claims 1 to 5, wherein the calculation unit (3) is configured to prepare the predicted mPD(t+T) for a time T that is at least T=5 seconds in the future.
7. The calculation unit (3) takes into account the basic physical fitness of the person (8) and, after the training session, uses the mechanical motion data BD(t) and physiological data PD(t) confirmed in multiple training sessions, as well as the altitude h(t), temperature Temp(t), and / or slope N(t) confirmed in multiple training sessions, to determine the coefficient a based on the optimization algorithm (11). ｘｉ , the summand a １０ , the aforementioned delay τ ｘｉ The apparatus according to claim 5, further configured to adjust the delay T.
8. The arithmetic unit (3) uses the optimization algorithm (11) to adjust the coefficient a after the training session ｘｉ , the augend a １０ , the delay τ ｘｉ and the delay T, and the optimization algorithm (11) is configured to a) The coefficient a ｘｉ Each of the above, the augment a １０ , the aforementioned delay τ ｘｉ For each of the above and the delay T, a step of identifying a plurality of discrete values in each case, b) a ｘｉ a １０ , τ ｘｉ and the step of setting T to one of the discrete values, c) A step of calculating mPD(t+T) based on the mathematical model, d) A step of calculating the modeling error between the measured physiological data PD(t+T) and mPD(t+T) for multiple t values. e) Repeat steps b) to d) for all combinations of the discrete values. f) a ｘｉ a １０ , τ ｘｉ And the step of selecting the discrete value that yields the minimum modeling error for T. The apparatus according to claim 7, including the following:
9. The apparatus according to claim 8, wherein in step d), the underestimation error is given greater weight than the overestimation error.
10. The calculation unit (3) uses an algorithm to adjust the coefficient a during the training session, taking into account the current physical fitness (12) of the person (8), using the mechanical motion data BD(t) and physiological data PD(t) confirmed during the training session, as well as the altitude h(t), temperature Temp(t), and / or incline N(t) confirmed during the training session. ｘｉ and the augment a １０ The apparatus according to claim 5, configured to adjust the
11. The calculation unit uses the algorithm for adjusting the current physical strength (12) to determine the difference Diff(t) = mPD(t) - PD(t) between the predicted mPD(t) of the physiological data and the measured physiological data PD(t), and when the difference Diff(t) is set to the threshold Threhold １ > If it is greater than 0, each constant const １ｘｉ By adding the coefficient a ｘｉ Modify and constant const １０ By adding the adenant a １０ Correct the difference Diff(t) is Threhold Ｍ If the threshold is <0, each constant const Ｍｘｉ By adding the coefficient a ｘｉ Modify and constant const Ｍ０ By adding the adenant a １０ The apparatus according to claim 10, configured to modify [something].
12. The apparatus according to any one of claims 1 to 11, wherein the control unit (4) is a PID controller.
13. The aforementioned PID controller is [Math 5] The system is configured to determine the support u(t) according to, where K Ｐ _K Ｉ and K Ｄ is the control parameter, e(t) is the control error at time t, and the function f １ (e), f ２ (e) and f ３ (e) The apparatus according to claim 12, wherein underestimation errors are selected to be given greater weight than overestimation errors.
14. The calculation unit (3) is configured to perform a calibration method in which the step response of the physiological data PD(t) or the mechanical motion data BD(t) is caused by a sudden change in the manipulated variable, and the calculation unit (3) calculates the control parameter K from the step response. Ｐ _K Ｉ and K Ｄ The apparatus according to claim 13, configured to determine the following.
15. The calculation unit (3) is configured to identify the step response from which at least one abrupt change in the manipulated variable and the physiological data PD(t) or mechanical motion data BD(t) after the training session are obtained, and the calculation unit identifies the control parameter K from at least one of the step responses. Ｐ _K Ｉ and K Ｄ The apparatus according to claim 14, configured to determine the following.
16. The apparatus according to any one of claims 1 to 15, wherein the mechanical motion data BD(t) is pedaling force in the case of a bicycle, or running force, rowing force, speed, torque, rotational speed, angular velocity and / or knee abduction torque in the case of a bicycle ergometer.
17. The support unit (6) includes an electric motor, a transmission and / or a brake, as described in any one of claims 1 to 16.
18. The apparatus according to any one of claims 1 to 17, wherein the physiological data PD(t) includes heart rate, heart rate variability, electrocardiogram, blood oxygen saturation, blood pressure, nerve activity, adduction and / or knee flexion and extension.