Method and control device for controlling a drive motor of a muscle-powered vehicle

Abstract

A method for controlling a drive motor (4) of a human-powered vehicle (2) is described. The method comprises reading in (S1) state information of a crank (8) of the human-powered vehicle (2) and determining (S2) at least one of a crank angle acceleration and a crank angle jerk using a numerical method as a function of the read-in state information. The method further comprises controlling (S3) the drive motor (4) as a function of at least one of the determined crank angle acceleration and the determined crank angle jerk. A control device (6) for carrying out such a method is also described. A drive train having such a control device (6) and a human-powered vehicle (2) having such a drive train are also described.

Description

[0001] The present invention relates to a method for controlling a drive motor of a human-powered vehicle. The control of the drive motor is carried out as a function of at least one specific crank angle acceleration of a crank of the vehicle and a specific crank angle jerk of the crank. The present invention further relates to a control device configured to carry out such a method. Furthermore, the present invention relates to a drive train with such a control device and to a human-powered vehicle with such a drive train.

[0002] Some methods for controlling electric motors in muscle-powered vehicles, such as e-bikes, use information about the crank angle acceleration of a crank on the vehicle and, alternatively or additionally, information about the crank angle jerk. Both the crank angle acceleration and the crank angle jerk must be determined for this purpose.

[0003] From the prior art, it is known, for example from DE 10 2017 103 735 A1, that a crank angle acceleration is determined by differentiating a rotational speed of the crank.

[0004] DE 10 2009 000 919 A1 relates to a method for operating a motor-assisted bicycle in which the desired drive assistance is determined by a dynamic evaluation of the angular velocity and / or angular acceleration of the pedal crank, wherein a control and / or regulating device infers the torque requirement taking into account other operating parameters. DE 10 2010 048 592 A1 relates to a bottom bracket system for pedelecs with a mid-drive motor concept, which enables control of the motor power, particularly at high motor power and long gear ratios, through the measurement of torque and crank speed on both sides and the integration of a freewheel between the crankshaft and the drive sprocket.

[0005] The object of the invention is to provide an improved method for controlling a drive motor of a muscle-powered vehicle, starting from the prior art. This object is achieved by the subject matter with the features of the independent claims. Further developments are described in the dependent claims.

[0006] The present invention relates, in a first aspect, to a method for controlling a drive motor of a muscle-powered vehicle. The control can be either steering or, alternatively or additionally, regulating the drive motor. At least one control parameter can be determined for controlling the drive motor. The drive motor can be controlled by the determined control parameter. The drive motor can be an electric motor. The drive motor can be configured to provide motive power to assist the driver in propelling the vehicle. The drive motor can be used to relieve the driver during driving and, alternatively or additionally, to increase the vehicle's range. The drive motor can be configured to provide motive power to relieve the driver.For example, the vehicle can be propelled by either muscle power provided by the driver or by the drive motor. The drive motor's power can thus relieve the driver of some of the workload.

[0007] The vehicle can be a bicycle, an e-bike, a pedelec, or a cargo bike. The vehicle can have a crank with a crank axle. The crank can be rotatably mounted in the vehicle's frame. Crank arms can be rigidly connected to the crank axle, and a pedal can be rotatably mounted at one end of each crank arm. The crank axle can also be a pedal crank axle. The rider can apply muscle power as a driving force via the pedals, thus transferring it to the crank to propel the vehicle.

[0008] Some or all steps of the procedure can be performed by at least one component of the vehicle, for example, a vehicle control unit. The procedure can be a computer-implemented procedure.

[0009] The method involves reading state information from the crank of the human-powered vehicle. This reading can involve reading at least one piece of state information, for example, exactly one piece of state information or multiple pieces of state information. The state information can include measured values ​​for one or more states of the crank as a function of time. For example, the state information of the crank can include the time course of at least one state of the crank. The state information can be a measurable quantity, which can be measured, for example, with a sensor. For example, measured values ​​of the crank's state can be discrete in time, such as a time-based signal. The individual measured values ​​can be equidistant in time. The state of the crank can change while the driver of the vehicle is pedaling.Information about this changing state as a function of time can be read in when reading state information from the crank.

[0010] Furthermore, the method involves determining crank angle acceleration and crank angle jerk using a numerical procedure as a function of the input state information. The determination of the crank angle acceleration and, additionally, the crank angle jerk using the numerical procedure can be an approximate determination or an estimate. The numerical procedure can, for example, involve time differentiation of the crank's state information, which is input as a time signal. The numerical procedure can employ simple or multiple time differentiation, such as double or triple differentiation. For example, a simple, double, and alternatively or additionally a triple time derivative of the state information present as a time signal can be determined.

[0011] The method further features controlling the drive motor as a function of at least one of the specified crank angle acceleration and the specified crank angle jerk. For example, the specified crank angle acceleration and, alternatively or additionally, the specified crank angle jerk can be used as parameters for speed control and, alternatively or additionally, for speed regulation of the drive motor, for example, to improve the accuracy of the speed control and, alternatively or additionally, the speed regulation. Alternatively or additionally, the specified crank angle acceleration and, alternatively or additionally, the specified crank angle jerk can be used as parameters for torque control and, alternatively or additionally, for torque regulation of the drive motor, for example, to improve the accuracy of the torque control and, alternatively or additionally, the torque regulation.The control of the drive motor can be carried out depending on further parameters, which can be determined, for example, based on further measured variables.

[0012] This method allows for at least an approximate determination of crankshaft acceleration and crankshaft jerk. The determination is performed using a numerical method. Such a method can represent a good compromise between noise reduction and accuracy for determining crankshaft acceleration and, additionally, crankshaft jerk without the need for an additional sensor.

[0013] This eliminates the need for a dedicated sensor to determine crankshaft acceleration and crankshaft jerk. A method for controlling a drive motor with such a defined crankshaft acceleration and, alternatively or additionally, such a defined crankshaft jerk is therefore possible for many vehicles that do not have such a dedicated sensor. Furthermore, this method reduces the number of electrical interfaces in the vehicle, as a dedicated sensor is no longer required. This allows for simplified system integration of the method-executing control unit within the vehicle using additional sensors. Consequently, the number of potential sources of error for both the method and the control unit executing it can be reduced.

[0014] By determining the crank angle acceleration and crank angle jerk, the start and, alternatively or additionally, the end of a rider's pedaling can be detected. The drive motor can then be controlled based on the detected start and, alternatively or additionally, on the detected end of the pedaling.

[0015] Furthermore, the state variable of crank angle acceleration and, alternatively or additionally, of crank angle jerk can be determined using such a method. This allows this state variable to be determined for controlling the vehicle's drive motor, thus enabling more precise control.

[0016] According to a further embodiment, the method can be characterized in that the status information includes the crank angle. For example, a time-based signal with time-equidistant, time-discrete measured values ​​of the crank angle can be read in as status information. Furthermore, the crank angle can be detected using a crank angle sensor. For example, a time-dependent profile of the crank angle can be detected, for instance, in the form of a time-dependent signal. The vehicle can be equipped with the crank angle sensor. The crank angle sensor can be communicatively connected to the control unit. The detection of the crank angle can be performed prior to the reading of the status information. For example, the detection of the crank angle is performed for the purpose of reading the status information.In the step of reading status information, the crank angle detected by the crank angle sensor can be read in.

[0017] Determining the crank angle acceleration may involve calculating a second time derivative of the crank angle. Determining the crank angle jerk may involve calculating a third time derivative of the crank angle.

[0018] According to a further embodiment, the method can be characterized in that the status information includes a crank angular velocity of the crank. The crank angular velocity can be a crank cadence. For example, a time-based signal with time-equidistant, time-discrete measured values ​​of the crank angular velocity can be read in as status information. Furthermore, the crank angular velocity can be detected using a cadence sensor. For example, a time-dependent profile of the crank angular velocity can be detected, for instance, in the form of a time-dependent signal. The vehicle can be equipped with the cadence sensor. The cadence sensor can be communicatively connected to the control unit. The detection of the crank angular velocity can be performed prior to the reading of the status information.For example, the crankshaft angular velocity is measured for reading the status information. In the status information reading step, the crankshaft angular velocity measured by the cadence sensor can be read.

[0019] Determining the crankshaft angular acceleration may involve calculating a simple time derivative of the crankshaft angular velocity. Determining the crankshaft angular jerk may involve calculating a double time derivative of the crankshaft angular velocity.

[0020] Thus, the process can be carried out by a control unit with a crank angle sensor and, alternatively or additionally, a cadence sensor for acquiring status information. A crank angle sensor and, alternatively or additionally, a cadence sensor for acquiring status information are already present in many human-powered vehicles, for example. Therefore, no further sensors or the information and measurements acquired by these sensors are necessary for carrying out the process. This allows for a particularly economical process, as such a process depends, for example, only on a crank angle sensor and, alternatively or additionally, a cadence sensor for acquiring status information.

[0021] Using such a method, the crank angle acceleration and, additionally, the crank angle jerk can be determined based on a measurement signal using numerical methods such as numerical differentiation. If the determination of the crank angle velocity and the crank angle jerk is performed as a function of the measured crank angle, an additional cadence sensor can be omitted. This can reduce system complexity. Alternatively, only a cadence sensor can be used, and a crank angle sensor can be omitted. This can also reduce system complexity.

[0022] If the crank angular velocity and, additionally, the crank angular jerk are determined as a function of the measured crank angular velocity, it is unnecessary to calculate an additional time derivative. This can reduce the computational effort of the procedure.

[0023] According to another embodiment, the acquisition of status information can be performed at a constant sampling rate. The acquisition of the crank angle and, alternatively or additionally, the acquisition of the crank angular velocity can be performed at a constant sampling rate. The acquisition can be performed cyclically and thus, for example, periodically over time.

[0024] By acquiring state information at a constant sampling rate, a temporal evolution of measured values ​​of the state information with time-equidistant measurements can be recorded. The numerical method can therefore be specifically designed for such a temporal evolution. For example, the method for determining the state can involve differentiating a temporal signal with time-equidistant measurements. This differentiation can be specifically configured for temporal evolutions with time-equidistant measurements. For example, the determination can involve only a single, or alternatively or additionally, multiple temporal differentiation of a temporal signal with time-equidistant measurements. For example, differentiation can be limited to a temporal signal with time-equidistant measurements.For example, it is therefore unnecessary to use a more complex, computationally intensive method for determining the crank angle acceleration and additionally the crank angle jerk, which is set up to determine and, for example, differentiate a temporal signal with measured values ​​that are not equidistant in time.

[0025] According to a further embodiment, the method can be characterized in that the numerical method is a differentiating low-pass filter. A differentiating low-pass filter can be used for the differentiating low-pass filtering. According to a systems-theoretical interpretation, determining the simple or multiple time derivatives of the state information using such a differentiating low-pass filter can be the numerical method for determining the crank angle acceleration and, additionally, the crank angle jerk as a function of the state information.

[0026] Such a method can be a more accurate numerical procedure for determining crank angle acceleration and crank angle jerk from the state information than using a difference quotient. A simple or multiple-order difference quotient of the state information as a function of time for determining the crank angle acceleration and, additionally, the crank angle jerk, can be less accurate than such a differentiating low-pass filter. Similarly, a subsequent low-pass filter of a value of the crank angle acceleration or crank angle jerk determined by a difference quotient from the time-dependent state information can be less accurate than determining the simple or multiple time derivatives of the state information using the differentiating low-pass filter.

[0027] Alternative methods involve first applying a low-pass filter and then calculating the difference quotient of simpler or higher order, depending on the result. Often, such application of the difference quotient, possibly with downstream or upstream low-pass filters, amplifies the measurement noise, and aliasing effects can occur as a result of the difference quotient. These aliasing effects cannot be compensated for by the downstream low-pass filter. The method presented here can therefore prevent such aliasing effects from occurring.

[0028] According to a further embodiment, the method can be characterized in that the differentiating low-pass filtering is linear and time-invariant. The differentiating low-pass filter can be a linear and time-invariant filter. The low-pass filter can exhibit a linear transfer characteristic and behave in a time-invariant manner, i.e., filter independently of time.

[0029] This allows for effective suppression of measurement noise when determining crank angle acceleration and, additionally, crank angle jerk. The measurement noise can be captured during the acquisition of the state information. This enables high accuracy in determining the time derivative of the state information and thus in determining the crank angle acceleration and, additionally, the crank angle jerk. Simultaneously, the state information as a function of time, for example, the crank angle or the cadence, can be differentiated, and this differentiated signal can be smoothed, thereby filtering out measurement noise.

[0030] According to a further embodiment, the method can be characterized in that a filter for linear and time-invariant low-pass filtering is an FIR filter. The FIR filter, for example, has an impulse response of finite length.

[0031] A method that uses such an FIR filter as a differentiating low-pass filter can be guaranteed to be numerically stable when determining the crank angle acceleration and, additionally, the crank angle jerk. The implementation of the differentiating low-pass filter can be carried out as a convolution sum, for example, as a weighted sum of a finite number of measured values ​​of the state information. This avoids the need for recursion to determine the crank angle acceleration and the crank angle jerk, which could lead to unstable behavior of the method. The method can thus be executed computationally efficiently and stably.

[0032] Furthermore, the method can react quickly to changing values ​​of the state information. This can be relevant if, for example, the specific crank angle acceleration and, alternatively or additionally, the specific crank angle jerk are used for a safety function, thus requiring a short reaction time. The convolution sums can be efficiently calculated in a control unit executing the method, such as in an embedded system in a vehicle. For example, this can be done with a multiply-accumulate unit (MAC). This can reduce development effort, for example, if the method is implemented in such an embedded system using fixed-point arithmetic.

[0033] Unlike IIR filters, which have an infinite impulse response, this method allows for a guaranteed prediction of the transition time after which the change in crank angle acceleration, and alternatively or additionally the crank angle jerk, during actual pedaling by the rider, is reflected in the specific crank angle acceleration or crank angle jerk determined from the state information. Such a finite transition time can be determined. A method using an IIR filter, for example, can be numerically unstable and lead to overshoot when determining the crank angle acceleration or crank angle jerk. Therefore, a maximum settling time can be defined for the time until the specific crank angle acceleration, and alternatively or additionally the specific crank angle jerk, reaches its respective final value.This minimizes the development effort required for such a procedure as described here.

[0034] According to another embodiment, the method can be characterized in that the FIR filter is an algebraic numerical differentiator. An example and further explanation of an algebraic numerical differentiator, also referred to as a derivative estimator, can be found in the following documents: Algebraic Derivative Estimators in Theory and Application, Lothar Kiltz, 2017 Saarbrücken; Survey on algebraic numerical differentiation: historical developments, parameterization, examples, and applications, Othmane, Kiltz, Rudolph, 2022, International Journal of Systems Science. The algebraic numerical differentiator can be used for algebraic numerical differentiation. According to an approximation-theoretic interpretation of algebraic numerical differentiation, the desired simple or multiple time derivative of the state information can be locally approximated by a polynomial.The polynomial can be evaluated at the point in time when it locally approximates the desired time derivative of the state information. For a given time interval, the steps of local approximation by a polynomial and evaluation of the polynomial at a specific time can be repeated for different time points. This allows for numerical differentiation of the measured values ​​of the state information.

[0035] Determining the first, second, or third time derivative of the state information can be performed numerically, rather than analytically, for discrete values ​​of the state information as a function of time. Such an algebraic numerical differentiator can represent a good compromise for deriving the state information with good noise suppression and accuracy for determining the first, second, or third time derivative of the state information, and thus for determining the crank angle acceleration and, additionally, the crank angle jerk.

[0036] According to another embodiment, the method can be characterized in that the FIR filter is a Savitzky-Golay filter. According to another embodiment, the algebraic numerical differentiator can be a Savitzky-Golay filter. According to another embodiment, the linear and time-invariant differentiating low-pass filter can be a Savitzky-Golay filter. With such a Savitzky-Golay filter, the method can be a smoothing method based on a polynomial approximation according to the least-squares method.

[0037] According to a further embodiment, the method can be characterized in that, for controlling the drive motor, it is determined whether the vehicle's crank is moving, depending on the specified crank angle acceleration and, alternatively or additionally, the specified crank angle jerk. Moving the crank can mean rotating the crank in the vehicle's bearing. The temporal derivative of the state information, determined via the approximate derivation of the signal as a function of time—for example, the crank angle acceleration or the crank angle jerk—can thus be used to determine whether the driver is pedaling or not. For example, the specified crank angle acceleration and, alternatively or additionally, the specified crank angle jerk can be compared with a threshold value, and based on this comparison, it can be determined whether the crank is moving or not.When the rider pedals, the crank angle acceleration and crank angle jerk change over the course of a crank revolution. If the crank angle acceleration or crank angle jerk is above a certain first threshold, it can be determined that the rider has started pedaling. For example, the system detects that the crank is moving and the rider is pedaling. This allows the start of pedaling to be detected. If the crank angle acceleration or crank angle jerk is below a certain second threshold, it can be determined that the rider has stopped pedaling. For example, the system detects that the crank is no longer moving and the rider is no longer pedaling. This allows the cessation of pedaling to be detected. For example, the drive motor can provide assistance when it has been determined that the crank is moving.

[0038] According to a further embodiment, the method can be characterized in that the drive motor can be configured to provide motive power to assist the driver when propelling the vehicle. The control system can include switching on and, alternatively or additionally, switching off the assistance provided by the drive motor. For example, if it has been determined that the driver is pedaling, the assistance can be switched on. If it has been determined that the driver is not pedaling, the assistance can be switched off. This can be achieved by comparing the determined crank angle acceleration and, alternatively or additionally, the determined crank angle jerk with a first and a second threshold value. The first and second threshold values ​​can be the same or different.This allows the assistance provided by the drive motor to be switched on and off when propelling the vehicle. This can be done depending on a specific crank angle acceleration and, alternatively or additionally, a specific crank angle jerk, for example, solely depending on the specific crank angle acceleration and, alternatively or additionally, the specific crank angle jerk. Alternatively, the assistance can also be switched on and off depending on other state variables or measured variables of the vehicle. Such an additional measured variable or state variable could, for example, be the crank cadence.

[0039] A second aspect of the present invention relates to a control device. The control device can further comprise one or more sensors for acquiring status information. A sensor can, for example, be a crank angle sensor or a cadence sensor. The sensor can be arranged on the crank or on the frame.

[0040] The control unit can be configured to execute a method according to an embodiment of the first aspect of the present invention. Such a control unit with such a sensor can be designed to be particularly cost-effective and space-efficient. Additional sensors besides the one or more sensors for acquiring the status information may not be necessary to determine the crank angle acceleration and, alternatively or additionally, to determine the crank angle jerk.

[0041] The control unit can also be configured to control a vehicle's drive motor. For this purpose, the control unit can be configured to determine a control parameter as a function of a specific crank angle acceleration and, alternatively or additionally, of a specific crank angle jerk. The control unit can be communicatively connected to the drive motor. Furthermore, the control unit can be configured to send the specific control parameter for controlling the drive motor to the drive motor. Further features, embodiments, and advantages are described in the first aspect. Conversely, features, embodiments, and advantages of the second aspect also represent features, embodiments, and advantages of the first aspect.

[0042] A third aspect of the present invention relates to a drive train with a control device according to an embodiment of the second aspect of the present invention. Furthermore, the drive train comprises a drive motor and a crank. The drive train may be free of further sensors. Further features, embodiments, and advantages are described in the first and second aspects. Conversely, features, embodiments, and advantages of the third aspect also represent features, embodiments, and advantages of the first and second aspects.

[0043] A fourth aspect of the present invention relates to a muscle-powered vehicle with a drive train according to an embodiment of the third aspect of the present invention. The muscle-powered vehicle can be a bicycle, an e-bike, a pedelec, or a cargo bike. Further features, embodiments, and advantages of each aspect are described in the descriptions of the first through third aspects. Conversely, features, embodiments, and advantages of the fourth aspect also represent features, embodiments, and advantages of the first through third aspects.

[0044] According to one embodiment, the muscle-powered vehicle can be characterized in that the control unit can be integrated into a housing with the drive motor. The housing can be mechanically connected to a frame of the vehicle, for example, to a downtube of a two-wheeled vehicle. The sensor for acquiring status information can be located outside the housing. Alternatively, the sensor for acquiring status information can be integrated into the housing. For example, at least one of the crank angle sensor and one of the cadence sensor can be integrated into the housing. This provides protection for the control unit and, for example, the sensors from physical impacts. Fig. Figure 1 schematically shows the steps of a procedure for controlling a drive motor of a muscle-powered vehicle. Fig. Figure 2 schematically shows a vehicle with a drive train and a control unit for executing steps of the schematically described above. Fig. 1. Procedure shown.

[0045] Fig. Figure 1 schematically shows the steps of a procedure for controlling a drive motor 4 of a muscle-powered vehicle 2. Fig. Figure 2 schematically shows the muscle-powered vehicle 2. In the embodiment shown, the muscle-powered vehicle 2 is a pedelec with a drive motor 4 designed as an electric motor. In an alternative embodiment, the vehicle 2 is a cargo bike.

[0046] Vehicle 2 has a drivetrain comprising a control unit 6, a drive motor 4, and a crank 8. Vehicle 2 also has pedals (not shown). The pedals are connected to the crank 8 via crank arms. A driver of vehicle 2 can apply muscle power to the crank 8 via the pedals, and this muscle power can be used to propel vehicle 2. The drive motor 4 is configured to provide power to propel the vehicle and thus assist the driver in propelling vehicle 2. The control unit 6 is communicatively connected to the drive motor 4. The control unit 6 and the drive motor 4 are integrated into a housing 12. The housing 12 is mechanically connected to a frame of vehicle 2.

[0047] Furthermore, vehicle 2 has a crank angle sensor 10. The crank angle sensor 10 is configured to detect S0.1 of the crank angle of the crank 8. The crank angle is recorded as a measured value, representing the state of the crank 8. Measurement values ​​of the crank angle are recorded as a function of time. This detection S0.1 is performed at a constant sampling rate, i.e., a constant sampling rate over time for a constant, cyclical detection S0.1 of the crank angle of the crank 8. During the detection step S0.1, time-equidistant measurements of the crank angle are recorded.

[0048] Furthermore, vehicle 2 has a cadence sensor 11. The cadence sensor 11 is configured to detect S0.2 a crank angular velocity of the crank 8. The crank angular velocity is referred to as cadence. The crank angular velocity is recorded as a measured value, representing the state of the crank 8. Measured values ​​of the crank angular velocity are acquired as a function of time. This acquisition S0.2 is performed at a constant sampling rate, i.e., a constant sampling rate over time for a constant cyclical acquisition S0.2 of the crank angular velocity of the crank 8. In each acquisition step S0.2, time-equidistant measured values ​​of the crank angular velocity are recorded.

[0049] In the embodiment shown, the vehicle 2 has both the crank angle sensor 10 and the cadence sensor 11. In an alternative embodiment, the vehicle 2 has either the crank angle sensor 10 or the cadence sensor 11.

[0050] The procedure involves reading in state information S1 of the crank 8 of the muscle-powered vehicle 2. The crank angle previously recorded in the acquisition step S0.1 is read in. The crank angular velocity previously recorded in the acquisition step S0.2 is also read in. Thus, measured and acquired time-dependent values ​​of the crank angle and crank angular velocity are read in. The state information is read in as a discrete function of time.

[0051] Furthermore, the method includes determining S2 of at least one crank angle acceleration and one crank angle jerk of crank 8. The determination of S2 is performed numerically as a function of the input state information. The determination of S2 involves temporal differentiation.

[0052] According to the illustrated embodiment, both the measured crank angle and the measured crank angular velocity are read in as state information. To determine S2 of the crank angular acceleration, the method employs a two-stage time differentiation of the read-in crank angle. To determine S2 of the crank angular acceleration, the method employs a single-stage time differentiation of the read-in crank angular velocity. In this embodiment, the crank angular acceleration is thus determined based on two different measured quantities: the crank angle and the crank angular velocity. This allows for a more precise determination of the crank angular acceleration by comparing the individual results.

[0053] To determine S2 of the crank angle jerk, the method employs a three-stage time differentiation of the input crank angle. To determine S2 of the crank angle jerk, the method employs a two-stage time differentiation of the input crank angular velocity. In this embodiment, the crank angle jerk is thus determined based on two different measured variables: the crank angle and the crank angular velocity. This allows for a more precise determination of the crank angle jerk by comparing the individual results.

[0054] In the illustrated embodiment, both the crank angle acceleration and the crank angle jerk are determined. In an alternative embodiment, either the crank angle acceleration or the crank angle jerk is determined.

[0055] In an alternative embodiment, either the crank angle or the crank angular velocity is read as state information. In such an embodiment, one of the crank angle sensor 10 and the cadence sensor 11 are omitted. In such an embodiment, either the crank angle is acquired (S0.1) or the crank angular velocity is acquired (S0.2). Determining the crank angular acceleration or crank angular jerk (S2) then involves only one differentiation step. Depending on which state information is read and which quantity, crank angular acceleration or crank angular jerk, is determined, either a single, double, or triple differentiation step is performed. Further differentiation steps are omitted in such an embodiment.The crank angle acceleration and, alternatively or additionally, the crank angle jerk can each be determined using exactly one simple, two, or three differentiation steps. This makes such a system less complex, and the computational effort required to determine S2 of the crank angle acceleration or crank angle jerk is reduced. Furthermore, one of the sensors can be omitted. In this alternative embodiment, either both the crank angle acceleration and the crank angle jerk are determined, or either the crank angle acceleration or the crank angle jerk is determined.

[0056] The numerical method employed is a differentiating low-pass filter for the approximate determination of at least one of the crank angle acceleration and the crank angle jerk. The determination of S2 of at least one of the crank angle acceleration and the crank angle jerk is performed using a differentiating low-pass filter. The differentiating low-pass filter is a linear and time-invariant filter. Therefore, the differentiating low-pass filter is linear and time-invariant.

[0057] The differentiating low-pass filter for linear and time-invariant low-pass filtering is an FIR filter. This FIR filter is implemented as a convolution sum and, in the illustrated embodiment, can be represented as a weighted sum of a finite number of state information values ​​as a function of time. This ensures that such an FIR filter is numerically stable. The FIR filter is an algebraic numerical differentiator. In the illustrated embodiment, the FIR filter is a Savitzky-Golay filter.

[0058] After at least one of the crank angle acceleration and the crank angle jerk has been approximately determined, the control S3 of the drive motor 4 of the vehicle 2 is carried out. This control S3 is performed as a function of at least one of the determined crank angle acceleration and the determined crank angle jerk. In the embodiment shown, the control S3 is performed as a function of both the determined crank angle acceleration and the determined crank angle jerk. This ensures that the control S3 of the drive motor 4 is carried out exactly. In an alternative embodiment, the control S3 of the drive motor 4 is performed either as a function of the determined crank angle acceleration or as a function of the determined crank angle jerk. This eliminates the need to determine S2 of either of the two quantities.

[0059] To control S3 of the drive motor 4, a determination S3.0 is performed, depending on at least one of the determined crank angle acceleration and the determined crank angle jerk, to ascertain whether the crank 8 of the vehicle 2 is moving. In this case, moving the crank 8 means turning or rotating the crank 8. Determining S3.0 establishes whether the rider starts or stops pedaling. From this, it is determined whether the crank 8 is moving and the rider is pedaling. For example, if it is determined that the crank 8 is not moving, the control S3 of the drive motor differs from if it is determined that the crank 8 is moving. The drive motor 4 is driven with a target torque that is zero if it is determined that the crank 8 is not moving. If it is determined that the crank 8 is moving, the drive motor 4 is driven with a target torque that is not zero.

[0060] Control S3 involves engaging support S3.1 from the drive motor 4. When it is determined that the crank 8 is being moved, support S3.1 is engaged. Drive power is then provided by the drive motor 4 to propel the vehicle 2. The rider is relieved of effort while propelling the vehicle 2, specifically when the rider is pedaling.

[0061] Control S3 includes a shutdown function S3.2 of the drive motor's assistance. If it is determined that the crank 8 is not moving, the assistance is shut down S3.2. Then, no drive power is provided by the drive motor 4 to propel the vehicle 2. The rider is not relieved of any effort while propelling the vehicle 2, for example, because the rider is not pedaling.

[0062] To execute the steps of the method, the control unit 6, the crank angle sensor 10, and the cadence sensor 11 are configured. In an alternative embodiment, to execute the steps of the method, the control unit 6 is configured together with a sensor that includes the crank angle sensor 10 and the cadence sensor 11.

[0063] The control unit 6 with the crank angle sensor 10 and the cadence sensor 11 is thus designed to save space and installation area by eliminating the need for additional sensors. Nevertheless, the method determines at least one of the crank angle acceleration and the crank angle jerk for controlling S3 of the drive motor 4. Furthermore, such a control unit 6 with these sensors 10, 11 is easier to maintain and less expensive to purchase and operate compared to a vehicle 2 with additional sensors for directly acquiring further parameters for controlling S3 of the drive motor 4. Moreover, the crank angle velocity and the crank angle jerk can be determined as state variables of the vehicle 2 using this method. Controlling S3 of the drive motor 4 as a function of at least one of these two parameters is more precise compared to conventional methods for controlling drive motors.At the same time, determining S2 of at least one of the two quantities using the numerical method is computationally efficient.

[0064] Reference sign 2 vehicles 4 Drive motor 6 Control unit 8 crank 10 Crank angle sensor 11 Cadence sensor 12 cases S0.1 Capturing a crank angle S0.2 Detecting a crank angle velocity S1 Reading crank status information S2 Determining at least one of a crank angle acceleration and a crank angle jerk S3 Controlling the drive motor S3.0 Determine whether the vehicle's crank is being moved S3.1 Activating support from the drive motor S3.2 Switching off support from the drive motor

Claims

[1] Method for controlling a drive motor (4) of a muscle-powered vehicle (2) comprising the steps: reading (S1) state information of a crank (8) of the muscle-powered vehicle (2); determining (S2) a crank angle acceleration and a crank angle jerk using a numerical method as a function of the read state information; and controlling (S3) the drive motor (4) as a function of at least one of the determined crank angle acceleration and the determined crank angle jerk. [2] Method according to claim 1, characterized by , that the status information includes a crank angle of the crank (8) and that a detection (S0.1) of the crank angle is carried out with a crank angle sensor (10). [3] Method according to any one of the preceding claims, characterized by, that the status information has a crank angular velocity of the crank (8) and that a detection (S0.2) of the crank angular velocity is carried out with a cadence sensor (11). [4] Method according to one of claims 2 or 3, characterized by , that the acquisition (S0.1; S0.2) of state information is performed at a constant sampling rate. [5] Method according to any one of the preceding claims, characterized by that the numerical method is a differentiating low-pass filter. [6] Method according to claim 5, characterized by that the differentiating low-pass filtering is linear and time-invariant. [7] Method according to claim 6, characterized by , that a filter for linear and time-invariant low-pass filtering is an FIR filter. [8] Method according to claim 7, characterized by that the FIR filter is an algebraic numerical differentiator. [9] Method according to one of claims 7 or 8, characterized by that the FIR filter is a Savitzky-Golay filter. [10] Method according to any one of the preceding claims, characterized by , that for controlling (S3) the drive motor (4) a determination (S3.0) is carried out depending on at least one of the determined crank angle acceleration and the determined crank angle jerk, as to whether the crank (8) of the vehicle (2) is moved. [11] Method according to any one of the preceding claims, characterized by , that the drive motor (4) is configured to provide driving power to assist a driver of the vehicle (2) in propelling the vehicle (2) and that the control (S3) includes switching on (S3.1) and switching off (S3.2) the assistance provided by the drive motor (4). [12] Control device (6) which is configured to perform a method of the preceding claims. [13] Drive train with a control device (6) according to claim 12 as well as with a drive motor (4) and with a crank (8). [14] Muscle-powered vehicle (2) with a drive train according to claim 13. [15] Muscle-powered vehicle (2) according to claim 14, characterized by , that the control device (6) is integrated in a housing (12) with the drive motor (4) and the housing (12) is mechanically connected to a frame of the vehicle (2).

Images