Force / torque sensor featuring redundant instrumentation and ready to detect faults

Abstract

Force / torque sensor comprehensive: a tool adapter plate (52) which is ready to be connected to a first object; a mounting adapter plate (54) which is ready to be connected to a second object; a number n of deformable supports (56a-e) which connect the tool adapter plate (52) to the mounting adapter plate (54), where n ≥ 4; an instrumentation comprising load measuring instruments (58l, 58s) attached to surfaces of four of the n deformable supports (56a-e); and a measuring circuit that is ready for operation, to separately determine the direction and magnitude of a force and torque between the first and second objects depending on electrical signals (g). 0-7) from the load measuring instruments (58l, 58s) of four different combinations of three of the four instrumented carriers (56a-d), to determine the specific force and torque outputs (f x , f y , f z , t x , t y , t z ) to compare the four provisions; and to report an error if at least one of the four force / torque outputs (f x , f y , f z , t x , t y , t z ) differs from the others by more than a predetermined tolerance.

Description

AREA OF INVENTION

[0001] The present invention relates in general to a force / torque sensor for robot applications and in particular to a force / torque sensor which has redundant instrumentation and is ready to detect and signal a fault depending on a comparison of force and torque values ​​derived from different groups of measuring instrument signals. BACKGROUND

[0002] German patent DE 10 2011 106 302 B3 discloses the determination of a measurement error of a multi-axis redundant sensor. For this purpose, a measurement matrix is ​​created, which depends, for example, on eight measuring points or sensor values. Subsequently, a pseudoinverse matrix of the measurement matrix is ​​generated. Based on the pseudoinverse matrix and the sensor values, a force and / or a moment is determined. To determine a measurement error, the forces and moments are transformed back into sensor values. If all measuring points are functioning correctly, the transformation yields the same values ​​as the measured values. If a discrepancy occurs between the sensor values ​​determined by the transformation and the original sensor values, a defective measuring point is present.

[0003] DE 10 2006 004 283 A1 discloses the calibration of force-torque sensors. Two strain gauges are arranged on each of four struts connecting an inner segment to an outer segment.

[0004] DE 102 17 019 C1 describes force-torque sensors. Strain gauges are arranged on connecting webs that extend from a central section and are connected to relief sections.

[0005] Robotics is a growing and increasingly important field in industry, medicine, science, and other applications. In many cases where a robot arm or a tool attached to it touches a workpiece, the applied force and / or torque must be monitored very precisely. Accordingly, a force / torque sensor is an important component of many robot systems.

[0006] A conventional type of force / torque sensor uses load-measuring instruments or strain gauges to measure the deformation of small supports connecting two mechanical parts—one connected to the robot arm and the other to a robot tool (or a mechanical coupling to the tool). For example, a central "middle part," known in the prior art as a tool adapter plate (TAP), is connected to a tool. Another body, arranged circularly around and spaced from the TAP, is known in the prior art as a mounting adapter plate (MAP) and is connected to a robot arm. The MAP and the TAP are connected by numerous relatively thin (and therefore mechanically deformable) supports arranged radially around the TAP—in some cases resembling the spokes of a wheel.A relative force or torque between objects that are appropriately attached to the TAP or the MAP attempts to move the MAP relative to the TAP, resulting in a slight deformation or bending of at least some of the supports.

[0007] Load measuring instruments are typically attached to all four surfaces of each support. These instruments translate tensile and compressive loads on the support surfaces, caused by mechanical deformation of the supports, into electrical signals. Once calibrated, signals from all four load measuring instruments are processed together to determine the magnitude and direction of any relative force and / or torque between the robot arm and the robot tool (and thus the force / torque applied by the tool to a workpiece).

[0008] Safety is paramount in any industrial environment where both humans and robots operate. A force / torque sensor on a robot is crucial for ensuring safety, as it allows the robot's control system to detect a collision or contact with an object (which could be a person) and interrupt its movement to prevent potential damage. Therefore, it is critical that the force / torque sensor is capable of accurately measuring and reporting applied loads at all times. If a force / torque sensor malfunctions—for example, due to a catastrophic failure of the load-measuring instrument (an open circuit), a partial load-measuring instrument failure (a resistance fault), detachment of the load-measuring strip, or sliding movement of metallic components, such as...If a deformable support, a wiring fault (a break in the line or a short circuit), or a fault in a measurement / processing circuit component occurs, the operational effect is the same as if the robot had no force / torque sensor at all; this is an unacceptable safety risk. Accordingly, a force / torque sensor must be capable of detecting and reporting any fault, even if the damage occurs when the system is not powered on, at least for some applications.

[0009] The background section of this document is provided to place embodiments of the invention in a technological and operational context, thereby assisting those skilled in the art in understanding their scope or benefits. Unless explicitly described as such, no information shall be considered prior art merely by virtue of its inclusion in the background section. SUMMARY

[0010] The following is a simple summary of the disclosure to provide the person skilled in the art with a basic understanding. This summary is not a comprehensive overview of the disclosure and is not intended to identify key or critical elements of embodiments according to the invention or to define the scope of the invention. The sole purpose of this summary is to present some concepts disclosed herein in a simplified form as an introduction to the more detailed description that will be presented later.

[0011] According to one or more embodiments described and claimed herein, a force / torque sensor comprises a number n of deformable supports connecting the TAP to the MAP, where n ≥ 4. At least four of the n deformable supports are equipped with load measuring instruments mounted on the surfaces of the supports, such that each support outputs two measuring instrument signals. The eight measuring instrument signals are grouped into four groups of six measuring instrument signals, such that each group contains the measuring instrument signals from three of the four equipped supports. Each group of six measuring instrument signals is multiplied by a calibration matrix, resulting in a set of six force and torque values. The four sets of force and torque values ​​are then compared.If one group does not match the other three groups when taking into account a predetermined tolerance, a sensor error is reported.

[0012] One embodiment relates to a force / torque sensor. The sensor comprises a tool adapter plate (TAP) ready to be connected to a first object and a mounting adapter plate (MAP) ready to be connected to a second object. The sensor also comprises a number n of deformable supports connecting the TAP to the MAP, where n ≥ 4, and instrumentation comprising load-measuring instruments attached to the surfaces of four of the n deformable supports. The sensor further comprises a measuring circuit ready to separately determine the direction and magnitude of a force and torque between the first and second objects based on electrical signals from the load-measuring instruments of four different combinations of three of the four configured supports.to provide; to compare the resolved force and torque outputs of the four solutions; and to report an error if at least one of the four force / torque outputs differs from the others by more than a predetermined tolerance.

[0013] The measuring circuit is particularly ready to output the respective determined force and torque values ​​when, in the comparison step, three of the four determinations agree within a predetermined tolerance and one determination does not agree within the tolerance with the other three solutions.

[0014] Another embodiment relates to a method for detecting a fault in a calibrated force / torque sensor with a number n of deformable carriers, where n ≥ 4, and wherein four of the n deformable carriers are configured to each generate two instrument signals. Eight instrument signals are iteratively sampled from the four configured carriers. The eight instrument signals are grouped into four separate groups of six signals each, with each group containing instrument signals from three of the four configured carriers. Each group of six sampled instrument signals is multiplied by a separate calibration matrix assigned to the respective group to calculate force and torque values, resulting in four groups of force and torque values. Either one group of force and torque values ​​or the mean of the four groups of force and torque values ​​is output.The four groups of force and torque values ​​are compared. If one of the four groups of force and torque values ​​differs from the other three groups by more than a predetermined tolerance, a sensor error is reported.

[0015] If exactly three of the four groups of force and torque values ​​match within the predetermined tolerance, the group of force and torque values ​​that does not match the other three groups will be output (preferably as correct force and torque values). BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The present invention will now be described in detail with reference to the accompanying figures, in which embodiments of the invention are illustrated. However, this invention should not be considered limited by the embodiments shown here. Instead, these embodiments are provided to ensure that this disclosure is complete and that the scope of the invention is fully described to those skilled in the art. The same reference numerals denote identical elements throughout. Fig. Figure 1 is a top view of a force / torque sensor with three supports according to the state of the art. Fig. Figure 2 is an enlarged representation of a sensor carrier. Fig. 1 according to the state of the art. Fig. Figure 3 is a cross-section and a functional schematic representation of a circuit of a half-bridge circuit topology of load measuring instruments on a carrier according to the state of the art. Fig. Figure 4 is a top view of a force / torque sensor with five supports. Fig. Figure 5 is a perspective view of the force / torque sensor with five supports. Fig. 4. Fig. Figure 6 is an enlarged view of two supports of the force-torque sensor of the Fig. 4, showing different load measuring instrument lengths. Fig. Figure 7 presents four matrix equations that map measuring instrument outputs to resulting forces and torques. Fig. Figure 8 is a flowchart which describes a procedure for detecting a fault in the force / torque sensor of the Fig. 4 represents. DETAILED DESCRIPTION

[0017] For the sake of simplicity and clarity, the present invention is described essentially with reference to an exemplary embodiment thereof. Numerous specific details are set forth in the following description to provide a complete understanding of the present invention. However, it is obvious to those skilled in the art that the present invention can also be carried out without these specific details. Well-known methods and structures are not described in detail in this description in order to avoid unnecessarily obscuring the present invention.

[0018] Fig. Figure 1 shows a top view of a conventional force / torque sensor 10, which has three deformable supports 16a, 16b, 16c connecting a TAP 12 to a MAP 14. Four load measuring instruments 18 are attached to each support 16 - one on each surface (the lower surface of the measuring instrument is in Fig. 1 not visible).

[0019] Fig. Figure 2 is an enlarged view of a beam 16, which undergoes deformation due to a force F applied to the TAP 12 relative to the MAP 14. This force deforms the beam 16 slightly to the left (the figure is not to scale). A compressive force is exerted on the left surface of the beam 16, which is detected by the load measuring instrument 18, mounted on the side; the measuring instrument 18 outputs a signal with a polarity (sign), the value of which is proportional to the detected deformation. Simultaneously, a tensile force is induced on the right surface, which is detected by the load measuring instrument 18, mounted on the side; the right measuring instrument 18 outputs a signal with the opposite polarity (sign), the value of which is proportional to the detected deformation.The load measuring instrument 18 on the upper surface (just like the measuring instrument on the lower surface) experiences little or no deformation because it lies along the neutral axis of that side of the beam 16; this measuring instrument 18 outputs little or no signal. The neutral axis is a line that generally runs downwards along the center of each surface of the beam 16, whereby a compressive load applied to the left side of the beam 16 is transformed into a tensile load on the right side. Accordingly, the beam 16 is not subjected to any load at the neutral axis.

[0020] Fig. Figure 3 represents a cross-section of an instrumented support 16 with a deformation measuring instrument 18 attached to each surface thereof, wherein the support 16 is subjected to a stress V EXCis operated. The measuring instruments 18 are wired in a half-bridge configuration between a supply voltage and ground, generating two outputs which are referred to as vertical and horizontal for reference. In the circuit of the Fig. A force or torque deforming the support 16 in the vertical direction would produce a large signal at output 1 and a small or no signal at output 2. A force or torque deforming the support 16 in the horizontal direction would produce opposite signals at the outputs (i.e., a large signal at output 2 and a small or no signal at output 1).

[0021] The measuring instruments 18 on each support 16a, 16b, 16c thus generate six signals – two signals from each support, with each signal being generated by the measuring instruments 18 wired as a half-bridge and arranged on opposite sides of the support. By multiplying the six signals by a calibration matrix, six forces and torques (Fx, Fy, Fz, Tx, Ty, Tz) can be provided. The force / torque sensor 10 therefore has a compact, efficient design, employing the minimum number of supports 16 and load measuring instruments 18 (for the half-bridge configuration) to uniquely provide or calculate the six force and torque values.

[0022] Sensor 10, which is located in Fig. The system shown in Figure 1 exhibits a "single point of failure" (i.e., a fault at only one location leads to a failure), such as a fault in a support 16, one or more load measuring instruments 18, a wiring fault, or similar. In fact, in the event of such a fault, the sensor 10 might not be able to accurately determine all applied force / torque loads, but depending on the specific fault, the sensor 10 might not even be able to detect that a fault exists. In this case, a robot lacking the ability to accurately detect and determine the applied loads could operate carelessly, which poses an unacceptable safety risk.

[0023] A straightforward, well-known approach to fault detection and robust operation is redundancy. Therefore, one could simply add a fourth instrumented, deformable support to the sensor design, which is integrated into... Fig. As shown in Figure 1, add the following. By separately providing the signals from different combinations of three of the four supports (16) and comparing the results, a fault in each support, each measuring instrument, each instrument wiring, etc., could then be detected. The force / torque sensor could then operate using only the three remaining, functional, instrumented supports. One problem with this approach is maintaining consistent stiffness in all radial directions, as the deformable supports of a force / torque sensor should be evenly spaced radially around the central section. With four supports, this results in a configuration characterized by collinearity of the supports (e.g., a "+" or an "x" configuration). If one support or its instrumentation has a fault, the remaining supports would form a "T" configuration. With this orientation, or rather,In this case, some applied torques and forces on the main leg or middle leg could not be distinguished. Accordingly, although a solution with four supports is part of the scope of the present invention, it is not a preferred embodiment.

[0024] According to embodiments of the invention, a force / torque sensor comprises n deformable supports, where n ≥ 4 (preferably n ≥ 5), which are generally spaced evenly radially apart around a central part. Of these n supports, at least four are instrumented by having load measuring instruments attached to them. If a support or its instrumentation has a fault, at least three instrumented supports remain operational and are at least capable of detecting a fault. Preferably, the three operational instrumented supports are capable of measuring and reporting all forces and torques between the TAP and the MAP. The supports to be instrumented (designed) should be selected such that, at least generally, there is coverage around the radial circumference of the sensor without the supports being collinear (e.g.,parallel) to other instrumented carriers, if possible.

[0025] A configuration that provides good stiffness and inherently avoids the collinearity problem is a sensor comprising five deformable supports connecting the TAP to the MAP, four of which are equipped with load measuring instruments. By arranging the five supports at uniform radial intervals around the TAP, it is ensured that they are not collinear (parallel) to one another. The number of four supports corresponds to the minimum number of instrumented supports required to guarantee that a fault in one support or in the instrumentation circuit of a support will at least be detected. Accordingly, the further discussion here refers to a sensor with five supports, four of which are instrumented. However, the present invention is not limited to this configuration.

[0026] In one embodiment, such a five-beam sensor has eight load-measuring instrument signals from four instrumented beams and can accurately measure applied loads using only six signals from any three of these beams. Therefore, the sensor has five possible calibration matrices for determining the loads: an 8x6 matrix that uses all signals, and four 6x6 matrices, each using signals from a different combination of three beams. For computational efficiency, the first matrix (i.e., the 8x6 matrix) is usually omitted.

[0027] When the sensor is fully functional, the load outputs provided by each of the four matrices produce identical outputs (within the calibration tolerance). If any fault damages a carrier or its instrumentation, the outputs will not match. These four outputs are continuously determined and compared during operation, and if they do not match, the system can report a fault condition. Depending on the type of fault, it may be possible to determine which carrier or instrumentation has been damaged. In this case (if the faulty carrier can be identified), the loads can still be accurately determined despite the fault by selecting the matrix that does not use the signal associated with the fault (the faulty carrier).

[0028] Fig. 4 and Fig. Figure 5 represents a force / torque sensor 50 according to an embodiment of the invention. In the sensor 50, a TAP 52 is connected to a MAP 54 by means of five supports 56a, 56b, 56c, 56d, and 56e, which are spaced evenly radially apart around the TAP 52. Only four of the supports 56a, 56b, 56c, and 56d are instrumented, with support 56e not having a load measuring instrument 58. This provides a non-symmetrical configuration of supports 56 while maintaining the same stiffness in all directions.

[0029] Each support 56 is directly connected to the TAP 52 and to the MAP 54 via bends 57, which support the deformation of the supports 56 under mechanical load. The TAP 52 is designed to be connected to a first object, such as a robot tool. The MAP 54 is designed to be connected to a second object, such as a robot arm. The TAP 52 and the MAP 54 are connected only by the supports 56.

[0030] Although it is not in the Fig. 4 and Fig. As shown in Figure 5, the force / torque sensor 50 also includes a processing circuit that is ready to receive electrical signals from each pair of load measuring instruments 58 and process the signals to determine the magnitude and direction of the force(s) and torque(s) applied between the MAP 54 and the TAP 52. The processing circuits determine these forces and torques by using each of four combinations of six signals (from three pairs of load measuring instruments 58, with the load measuring instruments 58 of each pair attached to opposite sides of a respective support 56) and comparing the results. A difference in the results (beyond calibration error) indicates a fault. The processing circuits can, for example,comprise a microprocessor coupled with a memory that is ready to store program code and sensor calibration data.

[0031] To account for the case in which a perfectly aligned overload with respect to Fz or Tz damages all supports 56 in the same way, the four instrumented supports can be measured with two different types of load measuring instruments 58, which have significantly different lengths, as shown in Fig. Figure 6 illustrates this. In this example, long measuring instruments 58l are attached to the instrumented support 56a, and short measuring instruments 58s are attached to the instrumented support 56b. Since silicon stress measuring instruments 58 are damaged by a peak load at any single point along their active length, it is ensured that the longer measuring instruments 58b will fail before the shorter measuring instruments 58s fail, as they extend further inward toward the support end into the region of higher stress concentration. The outputs of the stress measuring instruments 58l and 58s follow the mean load over their entire active length, so the two different lengths do not cause any complications regarding the design of the geometry of the supports 56 when the two types of measuring instruments 58l and 58s are centered at the same location.These different fault occurrence points for two of the carriers 56, in comparison with the other two carriers, ensure that no overload condition damages all measuring instruments 58 in exactly the same way and that the multimatrix fault detection method is misled.

[0032] The measurement electronics for sensor 50 require eight input channels, which are sampled simultaneously. After each sample, the eight values ​​are processed through four different calibration matrices, each generating six resolved forces and torques. The six outputs from each matrix are then compared. An error condition occurs if the outputs deviate by more than the allowed tolerance. The outputs from each individual matrix, or the average of all four matrices, can be used to generate the system's reported force / torque outputs.In embodiments where this comparison operation strains computing resources and limits performance, this comparison can be performed periodically at a rate lower than the system's normal sampling rate to reduce the computational load while still providing error detection at a speed sufficient to prevent injury to an operator.

[0033] If an overload occurs that exerts sufficient force to cause the metal body of sensor 50 to yield (without damaging the measuring instruments 58), a permanent deviation results in one or more pairs of measuring instruments 58. This causes the results of the three matrices using the signal to be inconsistent. If a test indicates that in such a case the same static force is generated on all supports (which is correctly measured), the robot can be programmed to perform a maneuver that, upon startup, applies one or more known loads to the sensor to detect the permanent deviation. In this fault scenario, it is assumed that the damage occurred while the power was off, since any overload during active operation is detected by the system and recognized as a fault.

[0034] If a catastrophic failure of measuring instrument 58 occurs due to an overload or electrical damage, a circuit break results and the signal becomes saturated (e.g., the signal assumes a maximum value). This signal immediately leads to an error in the matrix comparison test. Wiring faults that create a circuit break or short circuit lead to the same result (in particular, they are detected as errors).

[0035] If a partial failure of measuring instrument 58 occurs, generating a deviation, the matrix comparison test will result in an error. The matrix comparison test could only fail to result in an error if all measuring instruments 58 were damaged in exactly the same way. If the probability of this failure is sufficiently high, the probability of the failure going undetected can at least be reduced by using measuring instruments of different lengths, as previously discussed with reference to Fig. 6 is described.

[0036] If the load measuring instrument 58 becomes detached, the faulty measuring instrument no longer reports the applied load, which immediately leads to an error in the matrix comparison test.

[0037] The forces and torques applied to the sensor 50 are determined from signals of the measuring instrument 58 using a group of calibration matrices. The coefficients of each matrix can be determined by a calibration procedure in which one or more known forces or torques are applied to the sensor 50 and the corresponding measuring instrument outputs are sampled. The calibration matrix coefficients are then derived so that the actual measuring instrument outputs are mapped to the known force / torque values. Such calibration procedures are known to those skilled in the art and are therefore not described in further detail. The calibration matrices can be stored in a memory on the sensor 50 and are accessible to the processing circuits. Alternatively, they can be stored by a robot controller or another processing system outside the sensor 50.

[0038] A matrix for linear calibration using all eight instrument signals is possible, as shown below. Instrument signals are denoted by g0 to g7, and the resolved forces and torques are denoted by f. x , f y , t x etc. The matrix coefficients are denoted by k. 01 for the coefficient of g0 for f x , with k 12 for the coefficient of g1 for f y , etc. [g0g1g2g3g4g5g6g7]×[k01k11k21k31k41k51k61k71k02k12k22k32k42k52k62k72k03k13k23k33k43k53k63k 73k04k14k24k34k44k54k64k74k05k15k25k35k45k55k65k75k06k16k26k36k46k56k66k76]=[fxfyfztxtytz]

[0039] An analogous equation for a linear calibration of a standardized non-redundant sensor 10 with three carriers, using only six measuring instrument signals, is shown below, using the same naming conventions. [g0g1g2g3g4g5]×[k01k11k21k31k41k51k02k12k22k32k42k52k03k13k23k33k43k53 k04k14k24k34k44k54k05k15k25k35k45k55k06k16k26k36k46k56]=[fxfyfztxtytz]

[0040] Since only six discrete signals are required to determine the loads, a sensor 50 with a redundant carrier according to an embodiment of the invention has four different 6x6 matrix options, each of these matrices not using the two signals from one carrier 56. That is, the signals from measuring instruments 58 attached to three of the four instrumented carriers 56 are sufficient to determine the force / torque. There are four ways to combine three of the four carriers 56. For example, if the four instrumented carriers are designated A, B, C, and D, there are four possible combinations of three of these four carriers: ABC, BCD, ACD, and ABD. All four combinations are calculated and compared. If the result of any combination deviates (by more than a predetermined tolerance) from the result of the other combinations, the other three combinations have a faulty input signal.The four matrix operations with the measuring instrument signal inputs, the respective calibration matrices and force / torque outputs are in . Fig. 7 shown.

[0041] Each of these matrices generates its own set of force and torque outputs, and the outputs of all matrices should normally be nearly identical if all parts of sensor 50 are functioning correctly. If the signal from a measuring instrument 58 or a carrier 56 becomes faulty, the three matrices using the faulty element will produce incorrect results, while the result from the fourth matrix will still be correct. For example, if one or more measuring instruments generating the signal g1 are faulty, only the second matrix will produce correct results, since only the second matrix does not use the signal g1. In this case, the comparison software detects a faulty match between the specific outputs of the second matrix and the specific outputs of all other matrices and reports an error to the main system.

[0042] During the preliminary discussion, detailed representations of linear calibration matrix equations were presented to assist those skilled in the art in understanding the embodiments of the invention. However, the invention is not limited to linear calibration. Nonlinear calibration equations employ the same load-measuring instrument outputs but can accommodate higher-order values, such as g0. 2g1 * g2, |g0| and similar terms, along with larger calibration matrices that map these nonlinear terms, are used to calculate the forces and torques. Interested readers can find additional information on nonlinear calibration of load-measuring instruments in the American Institute of Aeronautics and Astronautics (AIAA) standard R-091-2003e, entitled "AIAA Recommended Practice for Calibration and Use of Internal Strain-Gage Balances with Application to Wind Tunnel Testing." The term "measuring instrument signals," as used here in a mapping operation to determine force and torque values ​​from the load-measuring instrument outputs, includes the signal outputs of, for example, the half-bridge circuit described in Fig. Figure 3 shows values ​​derived from these output signals.

[0043] Fig.Figure 8 describes a method for detecting a fault in a calibrated force / torque sensor 50, which has a number n of deformable supports 56, where n ≥ 5, and where four of the n deformable supports are instrumented to generate two instrument signals each. Eight instrument signals are sampled from the four instrumented supports (two instrument signals per support 56) (step 102). The eight sampled instrument signals are grouped into four individual groups of six signals each, such that each group contains the instrument signals from three of the four instrumented supports 56 (step 104). A group of six sampled instrument signals is multiplied by a calibration matrix assigned to the corresponding group to calculate force and torque values ​​(step 106).This procedure is repeated for all four groups of sampled instrument signals (step 108). The force and torque values ​​are output (step 110), which includes outputting one of the four groups of force and torque values ​​or outputting the average of all four groups of values. The four groups of force and torque values ​​are compared (step 112). If any of the groups of force and torque values ​​differs from the other three groups by more than a predetermined tolerance or threshold (block 114), an error is reported (step 116). If the force and torque values ​​of all four groups are identical (step 114), the procedure is repeated after a predetermined interval, starting again at step 102 to sample new instrument signals.

[0044] The term “deformable support” or simply “support” as used here is to be interpreted broadly as any mechanical connection between the TAP and MAP of a force / torque sensor, wherein the deformation of this connection is detected and / or quantified by sensors or transducers, such as load-measuring instruments or strain gauges, attached to surfaces of the connection. In particular, the term “support” is not limited to spoke-like structures depicted in the figures of the present application. For example, the bridge elements 17 of U.S. Patent No. 4,823,618 are supports in the context of embodiments according to the invention, as are the curved segments 13-16 of U.S. Patent No. 4,488,441. These examples illustrate the broad interpretation of the term “support” as used here and are not intended to be restrictive.

[0045] The embodiments according to the invention offer numerous advantages over prior art force / torque sensors. The provision of redundant instrumented deformable carriers enables fault detection by calculating force / torque solutions using inputs from various combinations of instrumented carriers and comparing the results. Furthermore, the force / torque sensor 50 can remain operational even in the event of a fault and can continue to measure and report forces and torques. By providing a total of n ≥ 4 deformable carriers, instrumentation of at least four carriers is ensured.

[0046] The present invention can, of course, also be implemented in a different manner than specifically described here, without any deviation from important features of the invention. The present embodiments are to be interpreted in all respects as clarifying and not as limiting, and all modifications which fall within the meaning and equivalence of the appended claims shall be covered by the scope of protection of the claims.

Claims

[1] Force / torque sensor comprising: a tool adapter plate (52) which is ready to be connected to a first object; a mounting adapter plate (54) which is ready to be connected to a second object; a number n of deformable supports (56a-e) which connect the tool adapter plate (52) to the mounting adapter plate (54), where n ≥ 4; an instrumentation comprising load measuring instruments (58l, 58s) attached to surfaces of four of the n deformable supports (56a-e); and a measuring circuit that is ready for operation, to separately determine the direction and magnitude of a force and torque between the first and second objects depending on electrical signals (g). 0-7 ) from the load measuring instruments (58l, 58s) of four different combinations of three of the four instrumented carriers (56a-d), to determine the specific force and torque outputs (f x , f y , f z , t x , t y , t z ) to compare the four provisions; and to report an error if at least one of the four force / torque outputs (f x , f y , f z , t x , t y , t z ) differs from the others by more than a predetermined tolerance. [2] Sensor according to claim 1, where the first object is a robot tool or a mechanical coupling with a robot tool; and where the second object is a robot arm or a mechanical coupling with a robot arm. [3] Sensor according to claim 1 or 2, wherein the measuring circuit is furthermore ready to operate in order to provide the respective determined force and torque outputs (f x , f y , f z , t x , t y , t z) to be output if, in the comparison step, three of the four determinations match within a predetermined tolerance and one determination does not match. [4] Sensor according to any of the preceding claims, wherein the instrumentation comprises four load measuring instruments, wherein a measuring instrument is attached to each surface of an instrumented carrier (56a-d), wherein the measuring instruments are wired together in a half-bridge topology on opposite sides of each carrier (56a-d) such that each instrumented carrier (56a-d) outputs two measuring instrument signals (g0, g1; g2, g3; g4, g5; g6, g7). [5] Sensor according to any of the preceding claims, wherein determining the direction and magnitude of the force and torque between the first and second object for each combination of three instrumented carriers comprises multiplying six measuring instrument signals with a 6x6 calibration matrix, from which six force and torque values ​​(f x , f y , f z , t x , t y , t z ) result. [6] Sensor according to any of the preceding claims, wherein the load measuring instruments (58l) attached to two of the four instrumented supports (56a-d) are longer than the load measuring instruments (58s) attached to the other two of the four instrumented supports (56a-d), and wherein the longer measuring instruments (58l) fail earlier under the same excessive force or torque than the shorter load measuring instruments (58s). [7] Method for detecting a fault in a calibrated force / torque sensor (50) having a number n of deformable supports (56a-e), where n ≥ 4, and wherein four of the n deformable supports (56a-e) are instrumented to generate two instrument signals (g0, g1; g2, g3; g4, g5; g6, g7), comprising: iterative sampling of the eight measuring instrument signals (g 0-7 ) of the four instrumented carriers (56a-d); Grouping the eight measuring instrument signals (g 0-7 ) into four individual groups of measuring instrument signals, each group having measuring instrument signals from three of the four instrumented carriers (56a-d); Multiply each group of sampled measuring instrument signals by a separate calibration matrix assigned to the group to obtain force and torque values ​​(f). x , f y , f z , t x , t y , t z ) to calculate four groups of force and Torque values ​​(f x , f y , f z , t x , t y , t z ) to win; Outputting one of the groups of force and torque values ​​(f x , f y , f z , t x , t y , t z ) or the mean of the four groups of force and torque values ​​(f x , f y , f z , t x , ty , t z ); Comparing the four groups of force and torque values; and if one of the four groups of force and torque values ​​(f x , f y , f z , t x , t y , t z ) differs from the other three groups by more than a predetermined tolerance, reporting a sensor error. [8] Method according to claim 7, wherein the grouping of the eight measuring instrument signals (g 0-7 ) into four individual groups of measuring instrument signals, a grouping of the eight measuring instrument signals (g 0-7 ) comprises four individual groups of six signals each; and wherein multiplying each group of sampled instrument signals by a separate calibration matrix comprises multiplying each group of six sampled instrument signals by a separate calibration matrix. [9] The method of claim 7 or 8, further comprising the comparison step: when three of the four groups of force and torque values ​​(f x , f y , f z , t x , t y , t z ) within the predetermined tolerance, output of the group of force and torque values ​​(f x , f y , f z , t x , t y , t z ), which does not match the other three groups. [10] Method according to one of claims 7-9, wherein the comparison step is performed less frequently than the sampling, grouping, multiplying and output step. [11] Method according to any one of claims 7-10, further comprising: During initialization, apply a known force or torque to the sensor (50) and check whether the sensor (50) displays the correct force and torque values ​​(fx , f y , f z , t x , t y , t z ) outputs to ensure that the sensor (50) was not damaged during a shutdown.

Images