Drive system for an electric bicycle, with calculation of a torque at the bottom bracket spindle for controlling the assistance power

EP4469341A1Inactive Publication Date: 2024-12-04BROSE ANTRIEBSTECHN GMBH & CO KGAA BERLIN
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
EP2023702073
Authority / Receiving Office
EP · EP
Patent Type
Applications
Current Assignee / Owner
Priority Date
2022-01-28
Filing Date
2023-01-24
Publication Date
2024-12-04
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

Existing drive systems for electric bicycles rely on expensive and space-consuming torque sensors to measure muscle power torque at the bottom bracket shaft, leading to higher costs and reduced driving comfort due to potential malfunctions and failures.

Method used

A drive system that uses an acceleration sensor and rotational angle sensor to calculate torque applied to the bottom bracket shaft, allowing for indirect measurement and control of support power via an electronic control unit, potentially using a mathematical model with sine or cosine transformations for accurate estimation.

Benefits of technology

This approach reduces costs and space requirements while enabling robust and fast control of support power, allowing the system to continue operating even if the torque sensor fails, maintaining driving comfort by switching to a secondary operating mode.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 1.1
    Figure 1.1
Patent Text Reader

Abstract

The invention relates in particular to a drive system for an electric bicycle (1), said drive system comprising: at least one acceleration sensor (2) for providing an acceleration signal (s2) that is representative of the acceleration of the electric bicycle (1); and at least one rotation angle sensor (3) for providing a rotational speed signal (s3) that is representative of a rotational speed of a bottom bracket spindle (T). An electronic control unit (SE) provided for controlling a drive motor (A) of the electric bicycle (1) is designed to use the acceleration signal (s2) and the rotational speed signal (s3) to calculate a torque (Mc) applied to the bottom bracket spindle (T) by muscle power and to control an assistance power provided by the drive motor (A) on the basis of the calculated torque (Mc).
Need to check novelty before this filing date? Find Prior Art

Description

[0001] Drive system for an electric bicycle with calculation of a torque on the bottom bracket shaft for controlling the assistance power

[0002] Description

[0003] The proposed solution relates in particular to a drive system for an electric bicycle and a method for controlling such a drive system.

[0004] Drive systems for electric bicycles (e.g., e-bikes or pedelecs) with at least one electric motor for the externally powered generation of additional drive torque and thus for providing assistance in addition to the drive power provided by human power are widely known. For example, DE 10 2018 001 795 A1 and DE 10 2019 201 812 B3 disclose drive systems with electric motors that enable continuous adjustment of a gear ratio between an input and an output. In each case, a planetary gear is part of the drive system, with changes in the gear ratio being controlled by a first electric motor.The support power for the downforce is provided via an additional electric motor and at the same time any actuating power occurring at the first electric motor is compensated in order to enable operation of the drive system without the supply of additional power from an energy storage device.

[0005] DE 10 2016 223 410 A1 discloses a drive system that enables continuous adjustment of a reduction gear and the provision of assistance with a single electric motor. In practice, it is common to make the assistance level, and thus its level, dependent on the rider's input. This rider's input is determined by measuring the torque currently applied to the bottom bracket shaft of the drive system by muscle power. The harder an electric bicycle rider pedals the pedals connected to the bottom bracket shaft, the higher the assistance level is selected to accelerate the electric bicycle with motor assistance. A torque sensor is typically provided on the bottom bracket shaft to measure the actual torque applied by muscle power.This makes it possible to directly measure the torque actually applied to the bottom bracket spindle, for example based on an inverse magnetoelastic effect or using at least one strain gauge.

[0006] However, directly measuring the actual torque applied is comparatively costly, at least in relation to other sensors and the overall cost of the drive system. A torque sensor also requires a comparatively large amount of space in the area of ​​the bottom bracket spindle. A possible malfunction or even failure of the torque sensor is also directly associated with significantly reduced riding comfort, as the assistance level can no longer be automatically adjusted in a meaningful way. Therefore, motor assistance is often suspended if a fault is detected in the torque sensor, meaning that no assistance level is available until the torque sensor is repaired.

[0007] Against this background, the proposed solution has the task of improving previously known drive systems and the control methods provided for them and at least reducing the disadvantages mentioned above.

[0008] This object is achieved both with a drive system of claim 1 and with a control method of claim 9.

[0009] A proposed drive system comprises at least one acceleration sensor for providing an acceleration signal representative of the acceleration of the electric bicycle and at least one rotation angle sensor for providing a speed signal representative of the rotational speed of the bottom bracket shaft. An electronic control unit provided for controlling at least one drive motor of the electric bicycle is configured to calculate a torque applied to the bottom bracket shaft by muscle power using the acceleration signal and the speed signal, and to control the assistance provided by the at least one drive motor based on this calculated torque.

[0010] The basic idea of ​​the proposed solution is to indirectly determine a torque at the bottom bracket shaft that is representative of the (current) drive power of an electric bicycle rider and to make it available for controlling the assistance power. Thus, the calculation based on the acceleration signal and the speed signal provides a torque signal for the calculated torque, which can be incorporated into the specification of the assistance power via at least one drive motor. The use of an indirect method for determining the muscle-powered torque applied to the bottom bracket shaft (often referred to as "pedaling torque") based on the acceleration signal and the speed signal that determines the angular position of the bottom bracket shaft can be implemented more cost-effectively than a direct measurement using a torque sensor.For example, a rotation angle sensor is many times cheaper than a torque sensor and also requires significantly less space. It has been shown that an indirect measurement method based on a (possibly continuous) measurement of the current acceleration of the electric bicycle and a (possibly continuous) measurement of the rotation angle of the rotating bottom bracket spindle can be used to comparatively accurately estimate the torque applied to the bottom bracket spindle by muscle power using a processor. In particular, this can be done so quickly and robustly that the assistance level can also be controlled using this method alone, if necessary. The calculated torque, or a signal derived from it, can, in principle, be used not only as an alternative to, but also in addition to a measurement signal from a torque sensor provided on the bottom bracket spindle for controlling the assistance level.

[0011] In one embodiment, a mathematical model is stored in the electronic control unit, which enables the torque at the bottom bracket spindle to be calculated from the acceleration signal and the speed signal. This mathematical model, which is implemented in software via an algorithm in the control logic of the electronic control unit, allows a torque signal for the calculated torque to be determined from the acceleration signal and the speed signal alone. As a first approximation, a Fourier transformation with a predefined window function can be used to calculate the torque within the mathematical model. For example, the use of a Hamming window, for example over a time range of 3 seconds, 5 seconds, or 7 seconds, has proven advantageous.With a view to faster calculation of a torque signal usable for controlling the at least one drive motor, the use of a sine transformation and / or a cosine transformation for the measured acceleration signal is considered advantageous in a further development. The mathematical model stored in the electronic control unit for calculating the torque then comprises a sine transformer and / or a cosine transformer, via which the acceleration signal is fed to a sine formation and / or a cosine transformation for calculating the torque (and thus a torque signal for the calculated torque). For an acceleration signal f over time t, for example, a sine formation and a cosine transformation are then specified as follows:

[0012] The two terms shown above carry the general spatial frequency 2TTft in the integral. To determine the torque on the bottom bracket shaft, the function Ys, c(f) is useful as an integral over one or half a revolution of the bottom bracket shaft, assuming that a rider of the electric bicycle pedals the pedals connected to the bottom bracket shaft with equal force using both legs (with Ys,c = Ys(f) - j ■ Y c(f)). Here, the sine or cosine transformation can only be performed for the relevant spatial frequency of the bottom bracket spindle's rotational speed. If the phase position is unknown, both the sine transformation and the cosine transformation are recommended. Thus, according to the proposed solution, the position and rotational speed of the bottom bracket spindle can be used to synchronize the transformation to the length of half a multiple of the bottom bracket spindle's rotation. If the phase position is known, for example, through indexing, the function Ys,c (f) can be divided into even and odd signals or signal values ​​in order to simplify it to a sine transformation or cosine transformation.In any case, by using a sine transformation and / or a cosine transformation of the acceleration signal, a processor-assisted, comparatively very fast estimation of the torque applied to the bottom bracket shaft by muscle power is possible with good accuracy, so that the assistance power of at least one drive motor of the electric bicycle can be effectively controlled.

[0013] As already explained above, the proposed solution includes, in particular, that in one embodiment, the assistance provided by the at least one drive motor is dependent exclusively on a torque signal calculated based on an acceleration signal and a speed signal. In this case, for example, the underlying indirect measurement method is somewhat less accurate than a direct measurement of the torque actually applied to the bottom bracket spindle by muscle power using at least one torque sensor, thereby achieving a cost and weight advantage.

[0014] Alternatively, the drive system can also additionally comprise at least one torque sensor for measuring the actual torque currently applied to the bottom bracket spindle by muscle power. In such a drive system, at least one torque sensor is therefore still provided for the direct, sensor-based measurement of the actual torque, but a processor-supported calculation of a torque is also carried out in accordance with the proposed solution. In a further development based on this, the electronic control unit can then be configured to use the measured actual torque (or a measurement signal for the actual torque) to control the assistance power and to use the calculated torque to verify the plausibility of the measurement of the actual torque by the at least one torque sensor. Here, too, the control of the assistance power is based on the calculated torque.However, the calculated torque is used (primarily or exclusively) to verify the plausibility of the actually measured torque when controlling the assistance power. If an (absolute) deviation above a tolerance value or an accumulation of such deviations between the actual, measured torque and the calculated torque above a threshold value indicates an error or even a failure of the torque sensor, the control electronics are configured to generate an error signal. The calculated torque thus allows for faster electronic detection of any malfunction of the torque sensor, especially before a complete failure of the torque sensor. In principle, the proposed solution also allows for mutual improvement in the accuracy of torque sensors with different operating principles. Such a system structure, based on so-called sensor fusion, can, for example,can be used to supplement a torque measuring system that only accesses one pedal to provide a two-sided effect.

[0015] Alternatively or additionally, the calculated torque can be used to provide a second operating mode of the drive system, in which the level of assistance provided is no longer dependent on the measurement signal from the torque sensor, but rather on the calculated torque. The drive system is thus able to access the calculated torque when needed, particularly in the event of a detected malfunction or failure of the torque sensor, in order to provide the electric bicycle rider with assistance and thus continue to maintain a certain level of riding comfort.In this case, the drive system can therefore be operated in at least one first operating mode and at least one second operating mode, wherein in the at least one first operating mode the assistance power depends on the actual torque measured with the at least one torque sensor (and thus on a measurement signal from the at least one torque sensor), and in the at least one second operating mode the assistance power depends on the calculated torque (and thus on a torque signal determined thereby). The control electronics thus enable switching between different operating modes and thus control scenarios as needed, depending on the basis on which the assistance power is to be controlled. Depending on the control scenario, a direct or indirect method for determining the torque at the bottom bracket spindle can be used for the assistance power.

[0016] In particular, the electronic control unit can be configured in this context to switch from the first operating mode to the second operating mode in the event of a malfunction (particularly electronically detected) or failure of the at least one torque sensor—either user-controlled, i.e., in response to a switching signal triggered by a user, or automatically. This provides a redundant system with a fallback level in which the control of the assistance power can depend on an indirectly determined torque on the bottom bracket shaft.The indirect method used here to determine the torque may then be less accurate than the also possible direct measurement of the actually applied torque, but it does at least allow a rider of the electric bicycle to continue using the at least one drive motor in the event of a malfunction or failure of the at least one torque sensor on the bottom bracket spindle. The usability of the calculated torque in a second operating mode can of course also be combined with the use of the calculated torque to check the plausibility of the measured torque, as explained above. The calculated torque can therefore primarily serve to check the plausibility of the measured torque in the first operating mode. Secondarily, the calculated torque is intended to control the assistance power in the second operating mode if the torque sensor malfunctions or fails.

[0017] In principle, the level of assistance can also depend on an adjustable assistance level, particularly one that can be adjusted depending on the riding situation. Depending on the selected assistance level, a specific (fixed) level of assistance is provided depending on the drive power applied via the bottom bracket spindle and, accordingly, the torque applied by muscle power. The goal here is to enable the user / rider of the electric bicycle to propel and accelerate the electric bicycle up to a maximum speed of, for example, 25 km / h with a comparatively moderate amount of muscle power applied.

[0018] The at least one drive motor can be controlled via the control unit to provide an assistance power that corresponds to the drive power currently applied to the bottom bracket spindle by muscle power multiplied by a factor specified by the control unit. For example, an assistance power of 0.3 times, 1.0 times, 1.5 times, or 2.5 times the currently applied drive power is provided, depending on the assistance level. The factor can be specified in steps. In particular, the control unit can vary the level of assistance to be provided by changing the size of this factor. The factor of the assistance power can thus be dynamically adapted to a riding situation of the electric bicycle, if necessary.

[0019] The proposed solution also includes an electric bicycle with a variant of a proposed drive system.

[0020] Furthermore, the proposed solution extends to a control method for a drive system of an electric bicycle. Here, a torque applied to the bottom bracket shaft by muscle power is calculated using an acceleration signal representative of the acceleration of the electric bicycle and a speed signal representative of the rotational speed of the bottom bracket shaft. The assist power of at least one drive motor of the drive system is controlled based on the calculated torque.

[0021] A variant of a proposed control method can thus be implemented in particular by a variant of a proposed drive system. The features and advantages explained above and below in connection with variants of a proposed drive system therefore also apply to variants of a proposed control method, and vice versa.

[0022] For example, in a proposed control process, the torque can be calculated using a mathematical model that incorporates the acceleration signal and the rotation signal as input variables. Using the stored mathematical model, a sine transformation and / or a cosine transformation of the acceleration signal can be performed to calculate the torque (and thus a torque signal for the calculated torque).

[0023] Analogous to a previously explained embodiment of a proposed drive system, in a variant of a proposed control method, the drive system can also additionally comprise at least one torque sensor for (directly) measuring the actual torque currently applied to the bottom bracket spindle by muscle power. The measured actual torque can then be used to control the assistance power, with this measurement of the actual torque being verified for plausibility by the at least one torque sensor with the calculated torque (continuously or at discrete times).

[0024] Alternatively or additionally, the drive system can be operated in at least one first operating mode and in at least one second operating mode, wherein in the at least one first operating mode the assistance power depends on the actual torque measured by the at least one torque sensor, and in the at least one second operating mode the assistance power depends on the calculated torque. Thus, switching between different control scenarios as needed is also possible here, as already explained above in connection with an embodiment variant of a proposed drive system. In particular, switching between the different operating modes is possible in the event of a malfunction or failure of the at least one torque sensor.

[0025] The proposed solution further comprises a computer program product having instructions which, when executed by at least one processor of an electronic control unit for a drive system of an electric bicycle, cause the at least one processor to execute an embodiment variant of a proposed method.

[0026] The attached figures illustrate possible embodiments of the proposed solution.

[0027] Here we show:

[0028] Figure 1 shows a schematic diagram of an electric bicycle in which the proposed solution is used, illustrating a measurement data acquisition for the development of a mathematical model subsequently implemented in the control electronics of a drive system of the electric bicycle, with the aid of which a torque applied to a bottom bracket shaft of the electric bicycle by muscle power is estimated with the aid of a processor, based exclusively on an acceleration signal and a speed signal;

[0029] Figure 2 shows a diagram showing the speed of the electric bicycle during a test ride.

[0030] Figure 1 shows a driving speed plotted against time;

[0031] Figure 3A is a diagram showing acceleration values ​​measured with an inertial measuring unit of the electric bicycle of Figure 1 over time for the test ride;

[0032] Figure 3B is a diagram showing calculated moving averages for the

[0033] Measurement data of Figure 3A are plotted against time; Figure 4 is a three-dimensional representation of the measurement data for the

[0034] Acceleration in one direction of travel after a Fourier transformation, wherein in the diagram of Figure 4 the amplitudes are plotted against frequency and time for a calculated Fourier transform;

[0035] Figure 5: Measurement data recorded for a trip of the electric bicycle

[0036] acceleration signal;

[0037] Figure 6 is a diagram in which the measured data in Figure 5 are calculated

[0038] Torque values ​​are plotted over time;

[0039] Figure ? a diagram comparing calculated torque values ​​with measured torque values ​​for riding an electric bicycle;

[0040] Figure 8 shows partial parts of an embodiment variant of a proposed drive system with an electronic control unit that implements a mathematical model on the software side.

[0041] Figure 1 shows a variant of a proposed electric bicycle 1 with a front wheel 10 and a rear wheel 11, wherein the rear wheel 11 can be driven with electric motor assistance via at least one drive motor A of a drive system. An assistance power can be transmitted to the rear wheel 11 via a power transmission link, for example a chain 12, via the drive motor A. The assistance power of the drive motor A is provided in addition to a drive power that a rider of the electric bicycle 1 applies by muscle power via pedals of the electric bicycle 1 connected to a bottom bracket shaft T. The assistance power provided by the drive motor A and its level are controlled via an electronic control unit SE of the drive system.The assistance provided should depend in particular on the torque applied by the rider of the electric bicycle 1 to the bottom bracket shaft T.

[0042] With the aim of indirectly determining the torque applied to the bottom bracket spindle T by muscle power, i.e., without measuring the actual torque using at least one torque sensor provided on the bottom bracket spindle T, the electric bicycle 1 of Figure 1 is equipped with an inertial measuring unit IMU, via which an acceleration of the electric bicycle 1 can be measured. The inertial measuring unit IMU is capable of measuring not only the acceleration of the electric bicycle 1 in a direction of travel f, but also accelerations in the transverse direction q and in the vertical direction s. An example acceleration measurement with measurement data provided by the international measuring unit IMU M u is illustrated in a diagram D1A of Figure 1.

[0043] In addition, the drive system of the electric bicycle 1 of Figure 1 has a rotation angle sensor, via which an angular position and thus a rotational speed of the bottom bracket shaft T can be measured. Measurement data m? provided by the rotation angle sensor during a ride with the electric bicycle 1 are illustrated by way of example using a diagram D1 B of Figure 1.

[0044] In order to develop a mathematical model which can be used effectively, i.e. sufficiently quickly and accurately, to infer a torque applied to the bottom bracket shaft T by muscle power from only an acceleration signal representative of the acceleration of the electric bicycle 1 and a speed signal representative of the speed of the bottom bracket spindle T, a test ride is carried out with the electric bicycle 1. This test ride extends, for example, over approximately 400 m. The test route intended for this purpose is initially flat and then changes to a 4 percent incline. The incline is followed by a flat ride of 100 m, at the end of which there is a gentle decline of approximately 50 m, during which the rider of the electric bicycle 1 does not pedal on the bottom bracket spindle T.For this test track, a curve of measured data m for the travel speed recorded results for a measured travel speed v of the electric bicycle 1. v as shown in diagram D2 of Figure 2.

[0045] For the test ride with electric bicycle 1, the inertial measurement unit (IMU), used here as an acceleration sensor with a sampling rate of 100 Hz, recorded measured values ​​for the acceleration of electric bicycle 1 in the direction of travel f, the transverse direction q, and the vertical direction s. Measurement data rrif for the accelerations in the direction of travel f, measurement data m q for the accelerations in the transverse direction q and measured data m sFor the accelerations in the vertical direction s, 3 are plotted against time for the test run in diagram D3A of Figure 3A. Filtered signals are determined from these raw signals for the different accelerations. Diagram D3B of Figure 3B shows the moving averages rnwA.f, rriMA.q, and rriMA.s determined for the measured data m. f , m q and m s Figure 3A plotted over time (10 filter points).

[0046] A Fourier transformation with a Hamming window and a window width of 5 seconds produces the graph in Figure 4 from the (filtered) acceleration signal for the direction of travel f. Here, the amplitude of the function obtained from the Fourier transformation is plotted against frequency and time. As expected, a signal with twice the frequency of the bottom bracket spindle speed, in this case from approximately 2.5 Hz to 3 Hz, is shown in signal range B, since the rider of the electric bicycle 1 pedaled at a cadence of 80 revolutions per minute during the test ride. It can be seen that in the function obtained by the Fourier transformation, or rather a signal determined with it, the increase in speed during acceleration and the decrease in amplitude towards the end of the section and the transition to the flat in the last part of the test section lead to significant changes and are therefore detectable.On this basis, the acceleration signal can be used to indirectly determine the torque applied to the bottom bracket spindle T by muscle power, at least provided the rotational speed of the bottom bracket spindle T is also known. The Fourier transformation performed thus results in an analyzable signal, from which, however, only measured vibrations need to be subtracted during actual operation of the electric bicycle 1.

[0047] The Fourier transformation chosen here as a first approximation

[0048] However, a Hamming window of more than 5 seconds may be too slow to determine a torque signal from this, as it can be used to control at least one drive motor A (particularly with regard to ride comfort for the rider of the electric bicycle 1). Against this background, one embodiment proposes replacing the Fourier transformation for a real signal f of the acceleration sensor (e.g., the acceleration sensor in the form of the inertial measurement unit IMU) with a sine and cosine transformation of the following form:

[0049] Both terms of the transformations shown above carry the general spatial frequency 2rrft in the integral. However, to calculate the torque, the function Ys,c(f) must be set as an integral over one or half a revolution of the bottom bracket spindle T, provided that the rider of the electric bicycle 1 pedals with equal force with both legs. This means that the sine or cosine transformation is only carried out for the relevant spatial frequency of the rotational speed of the bottom bracket spindle T. If the phase position is unknown, the sine and cosine transformation can be carried out. If the phase position and thus any imbalance when pedaling on the pedals connected to the bottom bracket spindle T is known, e.g. by an index, the function can be divided into even or odd, so that it is then simplified to a sine or cosine transformation.

[0050] Under this assumption, a system model of the electric bicycle 1 with its drive system can be developed, for example, using the SIMULINK® software. This model can be used to verify a mathematical model that can be used to estimate the torque at the bottom bracket spindle T from the acceleration signal and a speed signal representative of the rotational speed of the bottom bracket spindle T. Such a system model can simulate the relevant motion differential equations.

[0051] Based on the system model, the usefulness of a developed mathematical model MM can be verified, which is then implemented in the electronic control unit SE to control the drive motor A. The mathematical model MM, which can be represented, for example, in SIMULINK® by several continuous and discrete switching blocks, has as input variables an acceleration signal representative of the acceleration of the electric bicycle 1 and a speed signal representative of the rotational speed of the bottom bracket spindle T. Part of the mathematical model MM is, in particular, a sine-cosine transformer, via which the sine and / or cosine transformation of the acceleration signal explained above is performed.As an output variable, the mathematical model MM, which is implemented in the software via a corresponding algorithm in the electronic control unit SE, outputs a calculated torque and thus a typically continuous torque signal based on calculated torque values ​​that can be used to control the drive motor A. This then results, for example, in a curve of measured data if corresponding to diagram D7 in Figure 5 for measured accelerations in the direction of travel f during a ride of the electric bicycle 1. c , which is illustrated in diagram D8 of Figure 6.

[0052] If the curve of the calculated torque M c with a curve of an actual torque M applied to the bottom bracket shaft T, measured using a torque sensor maccording to diagram D9 of Figure 7, it can be seen that the calculated and thus estimated values ​​for the torque M c with the measured values ​​for the actual torque M m qualitatively and partly also quantitatively agree well. The calculated torque M c can thus easily be used to specify the level of assistance provided by the drive motor A. A possible reduction in the accuracy of the estimated torque value is offset by lower costs and lower weight compared to the use of a torque sensor on the bottom bracket shaft T.

[0053] Thus, for example, according to Figure 8, an electronic control unit SE can be provided which implements the mathematical model MM and which uses an acceleration signal S2 provided by an acceleration sensor 2 of the electric bicycle 1 and a speed signal S3 for the speed of the bottom bracket shaft T provided by a rotation angle sensor 3 of the electric bicycle 1 in order to estimate a torque at the bottom bracket shaft T. Based on this estimated / calculated torque, the electronic control unit SE controls the at least one drive motor A using a control signal s A , through which the support service is specified.

[0054] Alternatively, the drive system may additionally comprise a torque sensor 4, via which the torque M actually applied to the bottom bracket shaft T by muscle power is measured. mcan be measured directly. From this torque sensor 4 of the electric bicycle 1, the electronic control unit SE receives torque signals S4, which the electronic control unit SE uses to generate the control signal s A The torque M calculated using the mathematical model MM with the acceleration signal S2 and the speed signal S3 c The electronic control unit SE uses (a) for a plausibility check of the measured values ​​supplied by the torque sensor 4 and / or (b) as a fallback level for sole control of the drive motor 4 in the event of a malfunction or failure of the torque sensor 4. List of reference symbols

[0055] 1 electric bike

[0056] 10 front wheel

[0057] 11 Rear wheel

[0058] 12 Chain (transmission link)

[0059] 2 acceleration sensor

[0060] 3 Angle sensor

[0061] 4 Torque sensor

[0062] A drive motor

[0063] B Signal area f Direction of travel

[0064] IMU Inertial Measurement Unit

[0065] M c Calculated torque m f , miMu, m q , measurement data m s , m v , my MM Mathematical Model

[0066] Mm Measured actual torque mMA.S, miVIA.q, mMA,f Moving averages q Transverse direction

[0067] S2 acceleration signal

[0068] S3 speed signal

[0069] S4 torque signal

[0070] SÄ control signal

[0071] SE Electronic Control Unit

[0072] T bottom bracket spindle

[0073] V Vertical direction

Claims

Claims 1. Drive system for an electric bicycle (1), with - a bottom bracket shaft (T) for applying muscle power to propel the electric bicycle (1), - at least one drive motor (A) to provide externally powered support power in addition to the drive power and - an electronic control unit (SE) for controlling the support power provided via the at least one drive motor (A), characterized in that - the drive system includes at least one acceleration sensor (2) to provide an acceleration signal (S2) representative of the acceleration of the electric bicycle (1) and at least one rotation angle sensor (3) to provide a rotation speed signal (S3) representative of a rotation speed of the bottom bracket shaft (T) and - the electronic control unit (SE) is set up to apply a muscle-powered torque (M) to the bottom bracket axle (T) using the acceleration signal (S2) and the speed signal (S3). c ) to calculate and based on the calculated torque (M c ) to control the support service.

2. Drive system according to claim 1, characterized in that a mathematical model (MM) is stored in the electronic control unit (SE) which calculates the torque (M) from the acceleration signal (s2) and the speed signal (s3). c ) is made possible.

3. Drive system according to claim 2, characterized in that the stored mathematical model (MM) comprises a sine transformer and / or a cosine transformer, via which the acceleration signal (S2) is used to calculate the torque (M). c ) is subjected to a sine transformation and / or a cosine transformation.

4. Drive system according to one of claims 1 to 3, characterized in that the drive system additionally includes at least one torque sensor (4) for measuring the actual torque (M) currently applied to the bottom bracket shaft (T) by muscle power. m ) includes.

5. Drive system according to claim 4, characterized in that the electronic control unit (SE) is configured to control the support power using the measured actual torque (M). m ) to use and the calculated torque (M c ) to verify the plausibility of the measurement of the actual torque (M m ) by using at least one torque sensor (4).

6. Drive system according to claim 4 or 5, characterized in that the drive system can be operated in at least one first operating mode and at least one second operating mode, wherein in the at least one first operating mode the support power is determined by the actual torque (M) measured with the at least torque sensor (4). m ) is dependent and in which at least one second operating mode the support power depends on the calculated torque (M c is dependent on.

7. Drive system according to claim 6, characterized in that the electronic control unit (SE) is configured to switch from the first operating mode to the second operating mode in the event of a malfunction or failure of the at least one torque sensor (4).

8. Electric bicycle with a drive system (A) according to one of the preceding claims.

9. Method for controlling a drive system of an electric bicycle (1), in which a drive power for the propulsion of the electric bicycle (1) can be applied via a bottom bracket shaft (T) by muscle power and an additional support power can be applied via at least one drive motor (A) operated in a driving mode by external power, characterized in that a torque (M) applied to the bottom bracket shaft (T) by muscle power c ) using an acceleration signal (s2) representative of the acceleration of the electric bicycle (1) and a rotational speed signal (s3) representative of a rotational speed of the bottom bracket axle (T) and the support power is calculated based on the calculated torque (M c ) is controlled.

10. Method according to claim 9, characterized in that the calculation of the torque (M c) using a mathematical model (MM) which incorporates the acceleration signal (s2) and the rotational speed signal (s3) as input variables.

11. Method according to claim 10, characterized in that the stored mathematical model (MM) is used to calculate the torque (M). c ) a sine transformation and / or a cosine transformation of the acceleration signal (s2) is performed.

12. Method according to one of claims 9 to 11, characterized in that the drive system additionally includes at least one torque sensor (4) for measuring an actual torque (M) currently applied to the bottom bracket shaft (T) by muscle force. m ) includes.

13. Method according to claim 12, characterized in that the measured actual torque (M) is used to control the support power. m ) is used, whereby the measurement of the actual torque (M) m) by the at least one torque sensor (4) with the calculated torque (M c ) is made plausible.

14. Method according to claim 12 or 13, characterized in that the drive system can be operated in at least one first operating mode and at least one second operating mode, wherein in the at least one first operating mode the support power is determined by the actual torque (M) measured with the at least torque sensor (4). m ) is dependent and in which at least one second operating mode the support power depends on the calculated torque (M c is dependent on.

15. Method according to claim 14, characterized in that in the event of a malfunction or failure of the at least one torque sensor (4), the system switches from the first operating mode to the second operating mode.

16. Computer program product comprising instructions which, when executed by at least one processor of an electronic control unit (SE) for a drive system of an electric bicycle (1), cause the at least one processor to execute a method according to any one of claims 9 to 15.