Sensor device

Abstract

The present invention relates to a sensor device (100) comprising: a gyroscope unit (1) configured to acquire gyroscope measurement data (MD), an accelerometer unit (2) configured to acquire accelerometer measurement data (MB), and / or a magnetometer unit (3) configured to acquire magnetometer measurement data (MM), an evaluation unit (4) to which the gyroscope measurement data (MD) and the accelerometer measurement data (MB) and / or the magnetometer measurement data (MM) are provided, and which is configured to determine sensor output data (A) based on the gyroscope measurement data (MD) and the accelerometer measurement data (MB) and / or the magnetometer measurement data (MM) by means of a Kalman filter algorithm (4.1), and a monitoring unit (5) configured to monitor at least one auxiliary variable (R, GB) of the Kalman filter algorithm (4.1).1) to monitor and issue a warning message (W) if at least one monitored auxiliary variable (R, GB) exceeds a threshold value (SR, S-GB) assigned to the respective auxiliary variable.

Description

[0001] The present invention relates to a sensor device comprising: a gyroscope unit configured to acquire gyroscope measurement data, an accelerometer unit configured to acquire accelerometer measurement data, and / or a magnetometer unit configured to acquire magnetometer measurement data, and an evaluation unit to which the gyroscope measurement data as well as the accelerometer measurement data and / or the magnetometer measurement data are provided, and which is configured to determine sensor output data based on the gyroscope measurement data as well as the accelerometer measurement data and / or the magnetometer measurement data by means of a Kalman filter algorithm.

[0002] Such sensor devices with multiple inertial sensors are also referred to as inertial measurement units (IMUs) and are used, for example, for monitoring machines, with the sensor device also being used in particular as an inclination sensor to detect an inclination of a machine component.

[0003] A sensor device of the type mentioned above is known, for example, from US 2020 / 0141733 A1.

[0004] A Kalman filter algorithm suitable for use in a sensor device of the type mentioned above is described in detail in the technical report entitled “Indirect Kalman Filter for 3D Attitude Estimation” by the authors Nikolas Trawny and Stergios I. Roumeliotis.

[0005] From DE 198 18 860 A1 a method for detecting and localizing sensor errors in motor vehicles is known, wherein measurement data from a wheel speed sensor and a longitudinal acceleration sensor of the motor vehicle are evaluated using a Kalman filter algorithm.

[0006] The present invention is based on the objective of realizing a reliable sensor device.

[0007] This problem is solved according to the invention by a sensor device having the features of claim 1.

[0008] The sensor device according to the invention comprises a gyroscope unit configured to acquire gyroscope measurement data indicating one or more gyroscope rates or one or more rotational speeds. Gyroscope units, or simply gyroscopes, are also referred to as gyroscopic instruments, gyroscopic sensors, or simply gyroscopes. Preferably, the gyroscope unit is configured to acquire gyroscope measurement data for three mutually perpendicular axes, i.e., gyroscope measurement data indicating gyroscope rates with respect to three mutually perpendicular axes of rotation.

[0009] The sensor device according to the invention further comprises an acceleration sensor unit and / or a magnetometer unit, i.e. at least one of the two aforementioned units.

[0010] The accelerometer unit, if present, is configured to acquire accelerometer measurement data indicating one or more accelerations. Accelerometer units, or accelerometers for short, are also referred to as accelerometers, accelerometers, vibration sensors, or G-sensors. Preferably, the accelerometer unit is configured to acquire accelerometer measurement data for three mutually perpendicular axes, i.e., accelerometer measurement data indicating accelerations relative to three mutually perpendicular axes. By comparing the accelerometer measurement data with the acceleration due to gravity, the orientation of the accelerometer unit relative to the direction of the acceleration due to gravity can be determined in a known manner.

[0011] The magnetometer unit, if present, is configured to acquire magnetometer measurement data specifying one or more magnetic flux densities. Magnetometer units, or magnetometers for short, are also known as teslameters or gaussmeters. Preferably, the magnetometer unit is configured to acquire magnetometer measurement data for three mutually perpendicular axes, i.e., magnetometer measurement data specifying magnetic flux densities relative to three mutually perpendicular axes. By comparing the magnetometer measurement data with the Earth's magnetic field, the orientation of the magnetometer unit relative to the Earth's magnetic field can be determined in a known manner.

[0012] The sensor device according to the invention further comprises an evaluation unit to which the angular rate sensor measurement data, as well as the accelerometer measurement data and / or the magnetometer measurement data, are provided, and which is configured to determine sensor output data in a known manner based on the angular rate sensor measurement data, the accelerometer measurement data and / or the magnetometer measurement data using a Kalman filter algorithm. The evaluation unit is typically implemented by suitable programming of a microcontroller or another type of processor or arithmetic unit. With regard to a possible embodiment of the Kalman filter algorithm, reference is made to the contents of the aforementioned technical report, in particular to chapters 2 and 3 thereof.

[0013] The sensor device according to the invention further comprises a monitoring unit configured to continuously monitor at least one auxiliary variable of the Kalman filter algorithm during operation and to issue a warning message if at least one monitored auxiliary variable exceeds a threshold value assigned to the respective auxiliary variable. The monitoring unit is preferably implemented by suitable programming of the same microcontroller, processor, or arithmetic unit as the evaluation unit. However, it is also conceivable that the monitoring unit is implemented by suitable programming of a different microcontroller, processor, or arithmetic unit, to which the at least one auxiliary variable is provided by the evaluation unit during operation.

[0014] The auxiliary variable monitored by the monitoring unit is one from which information about an inaccuracy in the Kalman filter algorithm or the sensor output data determined using the Kalman filter algorithm can be derived directly or indirectly. Therefore, under normal operating conditions, at least one monitored auxiliary variable should remain below its assigned threshold. Exceeding the assigned threshold by at least one auxiliary variable is thus a strong indication of a malfunction or impairment of the sensor device, which is why the monitoring unit issues a warning message in this case. This warning message can be provided in any way, for example, by setting an error bit.

[0015] The monitoring unit according to the invention makes it possible to easily detect a potential malfunction or impairment of the sensor device by providing a warning message, and thus enables the realization of a reliable sensor device.

[0016] Preferably, the monitoring unit is configured to issue a warning message only if at least one monitored auxiliary variable exceeds the threshold value assigned to that auxiliary variable for a minimum duration assigned to that auxiliary variable. This prevents a warning message from being issued for measurement data that are only implausible for a short time, for example, due to temporary external disturbances, thereby reducing the number of false warnings.

[0017] In a preferred implementation, one of the auxiliary variables monitored by the monitoring unit is a so-called residual of the Kalman filter algorithm. Alternative terms for the residual of the Kalman filter algorithm include innovation, observation error, prediction error, or update. In the aforementioned technical report, for example, the residual of the Kalman filter algorithm is referred to as "residual r." The residual indicates a deviation between an erroneous measurement and a predicted state of the Kalman filter algorithm and is thus a kind of error measure that allows the Kalman filter algorithm to correct the difference between the model and the actual measurements. A very high residual is therefore a strong indicator that the sensor output data determined using the Kalman filter algorithm may be implausible.Therefore, by monitoring the residue, a particularly reliable sensor device can be implemented.

[0018] It is known that gyroscope units typically exhibit a systematic error called "gyroscope bias" or "time-varying offset," which can lead to a "deviation" or "drift" of the output measurement value. Therefore, the Kalman filter algorithm preferably includes a gyroscope data error model for estimating an error in the gyroscope data in order to correct the systematic error. Regarding a possible configuration of the gyroscope data error model, reference is made to the content of the aforementioned technical report, in particular to Chapter 2.1. Preferably, the monitoring unit is configured to monitor an auxiliary variable determined by the gyroscope data error model. Particularly preferably, the monitoring unit is configured to monitor an auxiliary variable that specifies the so-called gyroscope bias.A very high gyroscope bias is a strong indicator that the rotation rate sensor data, and therefore also the sensor output data determined using the Kalman filter algorithm, may be implausible. Monitoring the auxiliary variable determined by the rotation rate sensor data error model can therefore lead to the implementation of a particularly reliable sensor device.

[0019] An embodiment of the present invention is described below with reference to the attached Fig. 1 described, which shows a schematic diagram of a sensor device according to the invention.

[0020] Fig. Figure 1 shows a sensor device 100 according to the invention with a gyroscope unit 1, an accelerometer unit 2, a magnetometer unit 3, an evaluation unit 4 and a monitoring unit 5.

[0021] The gyroscope unit 1 is set up to acquire gyroscope measurement data MD, which specifies gyroscope rates relative to three mutually perpendicular axes X, Y, Z.

[0022] The accelerometer unit 2 is set up to acquire accelerometer measurement data MB, which specifies accelerations relative to the three mutually perpendicular axes X, Y, Z.

[0023] The magnetometer unit 3 is set up to record magnetometer measurement data MM, which specify magnetic flux densities with respect to the three mutually perpendicular axes X, Y, Z.

[0024] The evaluation unit 4 and the monitoring unit 5 are both implemented by suitable programming of a microcontroller 6.

[0025] The evaluation unit 4 is provided with the gyroscope measurement data MD, the accelerometer measurement data MB and / or the magnetometer measurement data MM.

[0026] The evaluation unit 4 is designed to determine sensor output data A based on the gyroscope measurement data MD, the accelerometer measurement data MB and / or the magnetometer measurement data MM using a Kalman filter algorithm 4.1, wherein the Kalman filter algorithm 4.1 includes a gyroscope data error model 4.1.1 for estimating an error of the gyroscope data MD.

[0027] The monitoring unit 5 is designed to monitor a residual R determined as an auxiliary variable by the Kalman filter algorithm 4.1 and to compare it with a threshold value SR assigned to the residual R.

[0028] The monitoring unit 5 is designed to issue a warning message W if the residual R exceeds the threshold value SR assigned to the residual R for a minimum duration DR assigned to the residual R.

[0029] The monitoring unit 5 is further configured to monitor a gyroscope bias GB determined as an auxiliary variable by the gyroscope rate sensor data error model 4.1.1 of the Kalman filter algorithm 4.1 and to compare it with a threshold value S-GB assigned to the gyroscope bias GB.

[0030] The monitoring unit 5 is designed to issue the warning message W when the gyroscope bias GB exceeds the threshold value S-GB assigned to the gyroscope bias GB for a minimum duration D-GB assigned to the gyroscope bias GB. Reference symbol list 100 sensor device 1 gyroscope unit 2 Accelerometer unit 3 Magnetometer unit 4 evaluation units 4.1 Kalman Filter Algorithm 4.1.1 Gyroscope Data Error Model 5 monitoring unit 6 microcontrollers A sensor output data D-GB Minimum duration DR minimum duration GB Auxiliary variable Gyroscope bias MB accelerometer measurement data MD gyratory rate sensor measurement data MM Magnetometer measurement data R Auxiliary quantity Residual S-GB threshold SR threshold W Warning message X, Y, Z axes

Claims

[1] Sensor device (100) comprising: a gyroscope unit (1) configured to acquire gyroscope measurement data (MD), an acceleration sensor unit (2) which is designed to Acquiring accelerometer measurement data (MB), and / or a magnetometer unit (3) configured to acquire magnetometer measurement data (MM), an evaluation unit (4) to which the gyroscope measurement data (MD) as well as the accelerometer measurement data (MB) and / or the magnetometer measurement data (MM) are provided, and which is configured to determine sensor output data (A) based on the gyroscope measurement data (MD) as well as the accelerometer measurement data (MB) and / or the magnetometer measurement data (MM) using a Kalman filter algorithm (4.1), and a monitoring unit (5) which is configured to monitor at least one auxiliary variable (R, GB) of the Kalman filter algorithm (4.1) and to issue a warning message (W) when at least one monitored auxiliary variable (R, GB) exceeds a threshold value (SR, S-GB) associated with the respective auxiliary variable. [2] Sensor device (100) according to claim 1, wherein the monitoring unit (5) is configured to issue the warning message (W) only if at least one monitored auxiliary variable (R, GB) exceeds the threshold value (SR, S-GB) assigned to the respective auxiliary variable (R, GB) for a minimum duration (DR, D-GB) assigned to the respective auxiliary variable (R, GB). [3] Sensor device (100) according to one of the preceding claims, wherein a monitored auxiliary variable (R) is a residue of the Kalman filter algorithm (4.1). [4] Sensor device (100) according to one of the preceding claims, wherein the Kalman filter algorithm (4.1) comprises a rate-of-rotation sensor data error model (4.1.1) for estimating a rate-of-rotation sensor data error (MD), and wherein a monitored auxiliary variable (GB) is determined by the rate-of-rotation sensor data error model (4.1.1).

Images