NL2031262A

NL2031262A - Method and system for calibrating one-dimensional force sensor for legged robot's leg

Info

Publication number: NL2031262A
Application number: NL2031262A
Authority: NL
Inventors: Ma Guoliang; Yu Bin; Ba Kaixian; Wang Mingming; Kong Xiangdong; Song Yanhe; Wang Chunyu
Original assignee: Univ Yanshan
Priority date: 2021-04-01
Filing date: 2022-03-11
Publication date: 2022-10-06
Also published as: CN113091999A; NL2031262B1

Abstract

Disclosed are a method and system for calibrating a one— dimensional force sensor in a legged robot's leg. According to the present invention, the position of a leg driving system is controlled through an upper computer, a force is loaded, and one— dimensional force sensor actual detection values of each joint are obtained; then, a leg driving system virtual model is built by computer software, the actual detection values of the two— dimensional force sensor and the displacement sensor of each joint driving unit are inputted into the virtual model for simulation, and theoretical detection values of a one—dimensional force sensor in the virtual model are obtained; finally, a calibration curve is obtained by a least square method, and a calibration correction coefficient of each joint is obtained, thereby realizing recalibration. of the one—dimensional force sensor through this coefficient. The method enables recalibration of the one— dimensional force sensor with no need of removal.

Description

P1214 /NL METHOD AND SYSTEM FOR CALIBRATING ONE-DIMENSIONAL FORCE SENSOR FOR LEGGED ROBOT’ S LEG

TECHNICAL FIELD The present invention relates to the field of legged robot sensor detection equipment, in particular to a method and system for calibrating a one-dimensional force sensor in a legged robot's leg.

BACKGROUND ART Robots become more and more popular in the society because they can replace human beings to do some dangerous and repetitive works. The structural design of legged robots simulates the physi- ological characteristics of legged animals, so that legged robots are more adaptable to complicated environments with an obvious feature of discontinuous support.

Legged robots sometimes produce a large contact force between the foot end and the ground when walking. If the foot end contact force cannot be controlled well, electronic devices carried by a robot body may be damaged as a result of suffering from a large impact. Therefore, the real-time detection of the contact force between the foot end and the ground is required by a robot.

Usually, two methods are available to detect the contact force between the foot end and the ground: the first method is to add a multi-dimensional force sensor at a foot end, which can di- rectly detect the foot end contact force; the second method is to install a one-dimensional force sensor at a top end of each joint driving unit in a robot’s leg, so that the foot end contact force can be obtained indirectly in combination with statics. The multi- dimensional force sensor is expensive and easy to damage, while the one-dimensional force sensor has a high data stability and a long service life. Therefore, the second method is mostly used to indirectly solve the contact force between the robot’s foot end and the ground in a legged robot.

The one-dimensional force sensor has been calibrated upon de-

livery. Users can directly determine a calibration coefficient ac- cording to a linear relation between a range and an output voltage in a sample. However, after a long period of use, some errors will be produced between detected values and actual values of the one- dimensional force sensor due to changes of a friction force, a pretightening force and other relevant factors. Therefore, the one-dimensional force sensor is required to be recalibrated. A traditional calibration method is to remove the one-dimensional force sensor from the robot’s leg. However, an interference fit is used for most parts on the robot’s leg, and frequent removal will damage an assembly relation of the robot’s leg, thereby affecting a motion control performance of the robot; moreover, an improper installation process will affect the calibration coefficient and even damage the one-dimensional force sensor.

To sum up, in the technical field of legged robot sensor de- tection equipment, it is urgent to provide a method for recali- braing a one-dimensional force sensor with the one-dimensional force sensor not removed from the leg joint.

SUMMARY The present invention is intended to provide a method and system for calibrating a one-dimensional force sensor in a legged robot's leg, which can recalibrate a one-dimensional force sensor with the one-dimensional force sensor not removed.

To achieve the aforesaid purposes, the present invention pro- vides the following solution: A method for calibrating a one-dimensional force sensor in a legged robot's leg includes: in a process of loading and unloading a force to a foot end, using an upper computer to record two-dimensional force sensor de- tection values for loading, displacement sensor detection values of each joint driving unit and one-dimensional force sensor detec- tion values of each joint driving unit; inputting the displacement sensor detection values and the two-dimensional force sensor detection values into a leg driving system virtual model built based on a leg mechanical structure for simulation, and recording one-dimensicnal force sensor simulated detection values of each joint driving unit in the virtual model; obtaining actual detection values of each joint and simulated detection values of each joint, by subtracting the one-dimensional force sensor detection values of each joint (with the foot end placed in an initial position) from the one-dimensional force sen- sor detection values of each joint and the one-dimensional force sensor simulated detection values of each joint, wherein the ini- tial position refers to no force loaded to the foot end; according to a linear relation between each actual detection value and each simulated detection value, solving a one- dimensional force sensor calibration curve of each driving unit by a least square method; according to the calibration curve of each driving unit and an original calibration coefficient of each driving unit, obtain- ing a calibration correction coefficient of each driving unit, wherein the original calibration coefficient is a one-dimensional force sensor calibration coefficient upon delivery; and according to the calibration correction coefficient of each driving unit, calibrating the one-dimensional force sensor.

The present invention further provides a system for calibrat- ing a one-dimensional force sensor in a legged robot's leg, in- cluding: a sensor detection value acquisition module, configured to use an upper computer to record two-dimensional force sensor de- tection values for loading, displacement sensor detection values of each joint driving unit and one-dimensional force sensor detec- tion values of each joint driving unit in a process of loading and unloading a force to a foot end; a simulation module, configured to input the displacement sensor detection values and the two-dimensional force sensor de- tection values into a leg driving system virtual model built based on a leg mechanical structure for simulation, and record one- dimensional force sensor simulated detection values of each joint driving unit in the virtual model; a variable value and sampling value acquisition module, con- figured to obtain actual detection values of each joint and simu- lated detection values of each joint, by subtracting the one-

dimensional force sensor detection values of each joint (with the foot end placed in an initial position) from the one-dimensional force sensor detection values of each joint and the one- dimensional force sensor simulated detection values of each joint, wherein the initial position refers to no force loaded to the foot end; a calibration curve acquisition module, configured to solve a one-dimensional force sensor calibration curve of each driving unit by a least square method according to a linear relation be- tween each actual detection value and each simulated detection value; and a calibration module, configured to obtain a calibration cor- rection coefficient of each driving unit according to the calibra- tion curve of each driving unit and an original calibration coef- ficient of each driving unit, wherein the original calibration co- efficient is a one-dimensional force sensor calibration coeffi- cient upon delivery, and calibrate the one-dimensional force sen- sor according to the calibration correction coefficient of each driving unit.

According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects: The position of a leg driving system is controlled through an upper computer, a force is loaded to the foot end by means of a two-dimensional force sensor, and one-dimensional force sensor ac- tual detection values of each joint are obtained; then, a leg driving system virtual model is built by computer software, the actual detection values of the two-dimensional force sensor and the displacement sensor of each joint driving unit are inputted into the virtual model for simulation, and simulated detection values of one-dimensional force sensor of each joint in the virtu- al model are obtained; finally, a calibration curve between each actual detection value and each simulated detection value of one- dimensional force sensor of each joint is obtained by a least square method, and a calibration correction coefficient of each Joint is obtained according to the calibration curve, thereby re- alizing calibration of the one-dimensional force sensor through this coefficient. Compared with previous methods, the method ena- bles completion of calibration with the one-dimensional force sen- sor not removed from the robot's leg, thereby avoiding a damage to the assembly relation of the robot's leg in a removal process and 5 an influence on the force sensor calibration coefficient in an in- stallation process.

BRIEF DESCRIPTION OF THE DRAWINGS In order to explain the technical solutions in the embodi- ments of the present invention or in the prior art more clearly, the accompanying drawings required in the embodiments will be de- scribed below briefly. Apparently, the accompanying drawings in the following description show merely some embodiments of the pre- sent invention, and other drawings can be derived from these ac- companying drawings by those of ordinary skill in the art without creative efforts.

FIG. 1 is a flow chart of a method for calibrating a one- dimensional force sensor in a legged robot's leg provided in exam- ple 1 of the present invention; FIG. 2 is a diagram of force loading experiment plans in ex- ample 1 of the present invention; FIG. 3 is a sketch of positive kinematics model of a leg driving system provided in example 1 of the present invention; FIG. 4 is a sketch of an inverse kinematics model of a leg driving system provided in example 1 of the present invention; FIG. 5 shows a relation between a joint angle of a legged ro- bot and a length of each joint driving unit provided in example 1 of the present invention; FIG. 6 is a schematic diagram of a method for calibrating a one-dimensional force sensor in a legged robot's leg in example 1 of the present invention; FIG. 7 is a one-dimensional force sensor calibration curve of each joint driving unit of a leg driving system provided in exam- ple 1 of the present invention; FIG. 8 is a sketch of a statics model of a leg driving system provided in example 1 of the present invention; FIG. 9 shows a relation between a torque on a joint and a force on a joint driving unit provided in example 1 of the present invention; FIG. 10 is a profile of a foot end contact force in loading direction of plan 1 in example 1 of the present invention; FIG. 11 is a profile of a foot end contact force in loading direction of plan 2 in example 1 of the present invention; FIG. 12 is a profile of a foot end contact force in loading direction of plan 3 in example 1 of the present invention; FIG. 13 is a profile of a foot end contact force in loading direction of plan 4 in example 1 of the present invention; FIG. 14 is a profile of a foot end contact force in loading direction of plan 5 in example 1 of the present invention; FIG. 15 is a profile of a foot end contact force in loading direction of plan 6 in example 1 of the present invention; FIG. 16 is a profile of a foot end contact force in loading direction of plan 7 in example 1 of the present invention; FIG. 17 is a profile of a foot end contact force in loading direction of plan 8 in example 1 of the present invention; FIG. 18 is a structure diagram of a system for calibrating a one-dimensional force sensor in a legged robot's leg provided in example 2 of the present invention.

Symbol description: O -- hip joint; E -- knee joint; G -- an- kle joint; A -- connection position between hip joint driving unit and base; B -- connection position between hip joint driving unit and thigh; C - connection position between knee joint driving unit and thigh; D -- connection position between knee joint driving unit and shank; F -- connection position between ankle joint driv- ing unit and shank; H -- connection position between ankle joint driving unit and foot end; I -- top point of semi-cylindrical foot end; OE -- thigh length; EG -- shank length; GI -- foot end length; 8; -- hip joint rotation angle: angle between thigh and positive x;-axis direction; 8: —- knee joint rotation angle: angle between shank EG and thigh OE extension line; 983 —- ankle joint rotation angle: angle between shank EG extension line and leg com- ponent GI; a —- angle between EG and ED; Bp —-- angle between OA and positive xg—axis direction; AB -- total length of hip joint driving unit; CD -- total length of knee joint driving unit; FH -- total length of ankle joint driving unit; AF,; -- force sensor detection signal of hip joint driving unit; AF.; -- force sensor detection signal of knee joint driving unit; AF; —- force sensor detection

F X signal of ankle joint driving unit; boi __ force on the foot end Fi of the leg in x-axis direction; fel __ force on the foot end of the leg in y-axis direction; T: -- torque on the hip joint; T:2 --

AX HDU torque on the knee joint; 13; -- torque on the ankle joint; nl - - length of hip joint driving unit of leg driving system test

AY U platform; p2 —— length of knee joint driving unit of leg driv-

AX PV ing system test platform; P3 —— length of ankle joint driving

AFDE unit of leg driving system test platform; -- one-dimensional force sensor detection value of hip joint driving unit of leg

AFDE driving system test platform; s2 —- one-dimensional force sen- sor detection value of knee joint driving unit of leg driving sys-

AJ tem test platform; s3 —— one-dimensional force sensor detection value of ankle joint driving unit of leg driving system test plat-

AF FHL form; sl —— force sensor simulated detection value of hip

AFDE Joint driving unit in virtual model; s2 -- force sensor simu- lated detection value of knee joint driving unit in virtual model;

AFEPY 53 —— force sensor simulated detection value of ankle joint

F SCHSOF driving unit in virtual model; ~~ —— two-dimensional force sen- F sensor sor detection value in x-axis direction; * -- two-dimensional F Sensory force sensor detection value in y-axis direction; = / -- actual contact force between the foot end and the ground in x-axis direc- F SERSOry tion directly detected by two-dimensional force sensor; * —- actual contact force between the foot end and the ground in y-axis F rx direction directly detected by two-dimensional force sensor; 2% - — contact force between the foot end and the ground in x-axis di-

rection solved by uncorrected force sensor calibration curve of Fy each joint driving unit; =~ -- contact force between the foot end and the ground in y-axis direction solved by uncorrected force FE correct edx sensor calibration curve of each joint driving unit; 4 —- contact force between the foot end and the ground in x-axis direc- tion solved by corrected force sensor calibration curve of each Jrcornectedy joint driving unit; #4 -- contact force between the foot end and the ground in y-axis direction solved by corrected force sen- sor calibration curve of each joint driving unit.

DETAILED DESCRIPTION OF THE EMBODIMENTS The technical solutions in the embodiments of the present in- vention will be described below clearly and completely in combina- tion with the accompanying drawings. Apparently, the embodiments described herein only constitute a part rather than all of the em- bodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts should fall within the protection scope of the present invention.

The present invention is intended to provide a method and system for calibrating a one-dimensional force sensor in a legged robot's leg, which can recalibrate a one-dimensional force sensor with the one-dimensional force sensor not removed.

To make the aforesaid purposes, features and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below in combination with the accompanying drawings and specific embodiments.

Hydraulic driving has the advantages of a large power-weight ratio and a fast response, so the legged robot mentioned herein can be a hydraulically-driven legged robot, but not limited to a hydraulically-driven legged robot.

Example 1 The example provides a method for calibrating a one- dimensional force sensor in a legged robot's leg, as shown in FIG.

1. The method includes the following steps:

Sl, in a process of loading and unloading a force to a foot end, use a DSpace controller to record two-dimensional force sen- sor detection values for loading, displacement sensor detection values of each joint driving unit and one-dimensional force sensor detection values of each joint driving unit.

Specifically, a method for loading a force to the foot end may include the following steps: slowly pulling by hand a wood cu- boid with the two-dimensional force sensor fixed to load a force to the foot end, slowly unloading the force after a preset tensile force is reached, and recording the two-dimensional force sensor Fres” [sensor | detection values ({ ~ and + }; displacement sensor detection AYU Ax AY EDE values ( PL m2 and £2) of each joint driving unit and one-dimensional force sensor detection values AF , AF and

AFDE $3) of each Joint driving unit in the process of loading and unloading.

FIG. 2 is a diagram of force loading experiment plans. As shown in FIG. 2, a leg driving system consists of a leg mechanical structure, a hip joint driving unit, a knee joint driving unit and an ankle joint driving unit. The one-dimensional force sensor is installed at a tail end of a piston rod of each joint driving unit, configured to detect a force on each joint driving unit.

As an optional implementation mode, before S1, the method further includes the following steps: according to a dimension and geometric relation of the leg mechanical structure, determining a mapping relation between a length of each joint driving unit and a rotation angle of each joint; based on the mapping relation and a kinematics model obtained from the leg mechanical structure, using an upper computer to control the leg driving system in an initial position.

Wherein a method for obtaining a kinematics model includes the following steps: simplifying the leg mechanical structure of a legged robot and obtaining parameters of a single-leg mechanical structure; determining a positive solution of the leg mechanical structure according to the parameters of the single-leg mechanical structure, wherein the positive solution is solving a function of a position of the foot end relative to the hip joint © according to a rotation angle 6; of hip joint ©, rotation angle 8: of knee joint E, rotation angle 8: of ankle joint G, thigh length OE, shank length EG and foot end length GI; and determining an inverse solution of the leg mechanical struc- ture according to the parameters of the single-leg mechanical structure, wherein the inverse solution is solving functions of the hip joint rotation angle ©;, knee joint rotation angle 8; and ankle joint rotation angle 6; according to the position of the foot end relative to the hip joint O, thigh length OE, shank length EG and foot end length GI; and obtaining the kinematics model accord- ing to the positive solution and inverse solution.

A calculation method for obtaining a kinematics model will be specifically introduced below.

FIG. 3 is a sketch of a positive kinematics model of a sim- plified leg of a leg driving system . The D-H parameters of the mechanical structure obtained from FIG. 3 are as shown in Table 1. Table 1 D-H Parameters of the Mechanical Structure of the Leg Driving System con- a, of, d g necting 1 oo oa 2 / 0 0 o, 3 l, 0 0 0, 4 lk 0 0 a, measured along x;.i, Oi represents an angle between z;.; and z; ro- tating around x; 1, di represents a distance between x; 1 and x; measured along z;, 6; represents an angle between x;.; and x; rotat- ing around z;, and z; represents a z-axis vertical to both x;-axis and y;-axis in x; and y; coordinates, and i=1, 2, ..., 4. In the leg D-H coordinates shown in FIG. 3, a position rela-

it tion can be represented by a connecting rod transformation 7 . cosó, —sinG 0 a, ip sind cosa, cosd cosa, —sina,, -—d sing, © |singsina,, cosé sina, , cosa, d cosa, (1) 0 0 0 1 With D-H parameters of the main mechanical structure of the leg driving system in Table 1 put into formula (1), transformation matrixes between the adjacent connecting rods can be obtained as follows: cos, —sing, 0 0 sing, cos&, 0 d= 0 0 1 (2) 0 0 0 cos, -—smó, 0 … | sing, cos@, 0 Ad = : 0 0 1 (3) 0 0 0 cos@, —sing, 0 > sin6, cos, 0 d= : : 0 0 1 (4) 0 0 0 cosd, —sinf, 0 s sind, cost, 0 d= 0 0 1 (5) 0 0 0 Through successive multiplication of formula (2), formula Tr (3), formula (4) and formula (5), * can be obtained as follows:

A relative position relation between the foot end position I d= I, cos@, +1, cos(6, +6, )+ cos(6, +6, +6,+6,) —sin(6,+6,+6,+6,) 0 l,cos(6, +6, +6,) | Lsin8, +L sin(8, +6, )+ (6) =| sin(6,+6,+6,+6,) cos(6,+6,+6,+6,) 0 I;sin(6, +6, +6, ) 0 0 1 0 0 0 0 1 (Xs Xoor) Por" Jet} and the hip joint O can be obtained from formula (6), where 984=0, representing that there is no 84. The kinematics posi- tive solution is as follows: |X, = 40088, +1, cos(8, +0, +1, cos(8, +8, +6,) ‚ on =h5in8, +1, sin (8, +6, +1, sin (9, +6, +0,) HY where 1, represents a length of an thigh component, 1, repre- sents a length of a shank component, 1; represents a length of a foot component, 6, represents a rotation angle of a hip joint O, 6, represents a rotation angle of a knee joint E, and 98; represents a rotation angle of an ankle joint G; FIG. 4 is a sketch of an inverse kinematics model of a sim- plified leg of a leg driving system. In order to avoid indetermi- nate solution problems that may appear when the kinematics inverse solution is performed, the present invention keeps the hip joint 0, ankle joint G and foot end I of the leg driving system colline- ar.

In FIG. 4, the angle 8; between the straight line where OGI locates and the positive x;-axis direction is as follows: 6, = arctan A Xp Xo ) (8) In AOEG, ZEOG is as follows:

jn OF’ +0G EG ZEOG = arceos ___ 20E OG . ” - > 2 9 OE? (a -q1) EG? (9)

2.0F- (v Koor + Xo 7 Gi) The hip joint rotation angle 8; can be obtained from FIG. 2 as follows: 0, =0,+ ZEOG 7 roa - 2 pr OE +{ |X 24 XV 2 GI) EG ¥ Y Joot foot 10 =arctan 2(X ops Ä or) +arccos| ————r—e ee (10) me 2-0F |. [xT ix? ~Gi | In AOEG, ZOEG is as follows: ps EK + EG OG ZOEG = arccos OE + EG 00 20E EG OF + EG? <{ + -ar) (11) SI — 20E. EG The knee joint rotation angle 6, can be obtained from FIG. 2 as follows: 0, = ZOEG-z 2 12 2 Ta OE + EG A A or +X, 7 Gl | (12) =drccos| TTT | = J] 20E EG In AOEG, ZOGE is as follows: vd 2 pa? pe + EG” —OE LOGE = arccos OG + EG OE

2.0G- EG > 9 CD - 2 2 i. (Ja +x, -q1) +EG2-0E? (13) = Areeos| ~———— me

2. t [vr ext ro Gi) EG The ankle joint rotation angle 6; can be obtained from FIG. 4 as follows:

0, = LOGE 2 Vie +57 - GIJ +EG 0E a, FE AMCCOS | ee 2, [xt exit; ~G1)-EG A method for calculating a mapping relation between a length of each joint driving unit and a rotation angle of each joint will be specifically introduced below: FIG. 5a)-c) show the relation between the angles of the hip Joint O, knee joint E and ankle joint G and the lengths of the driving units respectively.

The total length AB of the hip joint driving unit, the total length CD of the knee joint driving unit, and the total length FH of the ankle joint driving unit are as follows: AB =I, + Ax, CD =1, + Ax, (15) FH =I, + Ax where lx: represents an initial length of the hip joint driv- ing unit, ly; represents an initial length of the knee joint driv- ing unit, and li; represents an initial length of the ankle joint driving unit. In FIG. ba) AAOB, ZAOB is obtained from a law of cosines as follows: ZAOB = arccos EE (16) 2-04-0OB According to the geometric relation in FIG. 5a), the hip Joint rotation angle 8; can be obtained as follows: | = +OB (ly +Ax,,) | 9 = pf —- LA0B —- LBOE=f} —arccos| — | /BOE (17) 2-04-0B The length Ax,, of the hip joint driving unit can be obtained from the above formula as follows: 2 2 (18) Ax, = JOA’ + OB’ -2:04-0B-cos(B- ZBOE-0,) ~1,, In FIG. 5b) ACDE, ZCED can be obtained from a law of cosines as follows:

ye EC? + ED* -CD° ZCED = arccos OOO

2. EC-ED (19) According to the geometric relation in FIG. 5b), the knee Joint rotation angle ©; can be obtained as follows: EC +ED {lg + Ax, ) 0, = LCED ~ 7 — a + ZOEC=arccos ot a + LOEC (20) ë 2. EC-ED The length Ax,; of the knee joint driving unit can be obtained from the above formula as follows: 2 nD A rr pe Jy. (21) Ar, = EC? + ED? -2-EC-ED-cos(LOEC -6, ~ 1 - a) = 1, In FIG. 5c) AFGH, ZFGH can be obtained from law of cosines as follows: GF? +GH® FH? ZFGH = arccos | =e (22)

2.-GF-GH According to the geometric relation in FIG. 5c), the ankle joint rotation angle 8; can be obtained as follows: 0, =2— ZFGH — ZIGH — ZEGF - 2 GF? + GH? (ly + Ax) . (23) =7 —arccos| | JGH — ZEGE 2-GF-GH The length Ax,; of the ankle joint driving unit can be ob- tained from the above formula as follows: ee ee ; == (24) Ax =\JGF* +GH° =2-GF -GH -cos(6, ~ n+ ZIGH + ZEGF) ~,; After the kinematics model of the leg mechanical structure and the mapping relation between a length of each joint driving unit and a rotation angle of each joint are obtained, the kinemat- ics model of the leg mechanical structure and the mapping relation between a rotation angle of each joint and a length of each joint driving unit obtained from the above steps are programmed specifi- cally in the DSpace controller, the length of each driving unit of the leg is controlled through the program to keep the foot end po- LL ‚ Le ‚Le een Xi) . sition in the initial position ’ “+, then the wood cuboid with the two-dimensional force sensor fixed is horizontally and slowly pulled towards the left by hand to load a horizontal force to the foot end, the force is unloaded slowly after a preset ten- sile force is reached and the above steps are repeated for multi-

AX IDE ple times; and the displacement sensor detection values ( LE

HOU DU AX and AX ) of each joint driving unit, one-dimensional ns ns 3 force sensor detection values AF ; AFG and AFS, of each Joint driving unit and two-dimensional force sensor detection val- ues (°F and + ) are recorded in the process of loading and unloading.

It should be noted that the initial position mentioned herein is the position with no force loaded to the foot end. The position could be anywhere as long as the position is corresponding to the leg driving system virtual model. At the same time, the aforesaid loading of horizontal force to the foot end shows merely one con- dition of the embodiments. The application does not limit the loading of horizontal force to the foot end. The direction of the force can be arbitrary.

S2,; inputting the displacement sensor detection values of each joint driving unit and the two-dimensional force sensor de- tection values into a leg driving system virtual model built based on a leg mechanical structure for simulation, and recording one- dimensional force sensor simulated detection values of each joint driving unit in the virtual model; wherein the leg driving system virtual model is based on the leg mechanical structure of a legged robot in Sim- ulink/SimMechanics.

According to the example, SimMechanics Second Generation is selected from SimMechanics Link tools in SolidWorks, an assembly file of the leg mechanical structure is exported to a XML file and several STL files, and then a smimport command is run in a MATLAB command window. MATLAB will automatically read the corresponding XML file and STL files, and automatically build a leg driving sys- tem virtual model of a legged robot.

A Prismatic module in the virtual model: a prismatic pair module, configured to connect the piston rod to a driving unit cylinder, so that the two rigid bodies can be shifted relative to each other. A sliding position of the prismatic pair can be set through Actuation attribute, and a force on the prismatic pair can be measured through sensing attribute, that is, the lengths (

AX DU AY IDE AYE eto P2 and P3 } of each joint driving unit can be set through Actuation attribute, and the force sensor simulated detec- tion values AE AFL and AFG of each joint driving unit in the virtual model can be detected through sensing attribute; External Force and Torque module: an external force and torque module, configured to directly apply a disturbing force to the foot end through Force attribute, that is, the two-dimensional Freer FT force sensor detection values (°F and + ) can be inputted into the leg driving system virtual model for simulation through the External Force and Torque module.

S3, obtaining actual detection values and simulated detection values of each joint, by subtracting the one-dimensional force sensor detection values of each joint (with the foot end placed in an initial position) from the one-dimensional force sensor detec- tion values of each joint and the one-dimensional force sensor simulated detection values of each joint, wherein the initial po- sition refers to no force loaded to the foot end; S4, according to a linear relation between each actual detec- tion value and each simulated detection value, solving a one- dimensional force sensor calibration curve of each driving unit by a least square method; the least square method is a classical and the most basic pa- rameter identification method, and also the most widely used iden- tification method, with a fundamental formula as follows: Z,=HJ+V, 4, where Z, represents a sampling value, Hs represents a variable value, A represents a parameter value, and V, represents a sampling noise.

The concept of the least square method is to find an estimat- ed value 4 of A, so as to minimize a quadratic sum of a difference between each measured value Z;i(i=1, .., m) and a measured value es- timate Z=Hi determined by the estimated value 4, as follows: J(2)-(Z, H,Â) (Z, H,Â)-min((Z, H,2) (Z, H,À € (4)=( > 4) | > 4) min H,£) ( > 4) (06) According to an extreme value theorem, we can get: ie =2H;{(Z, H,Â)=0 J=á (27) Formula {27} can be simplified as follows: H)H,i=HZ, when the number m of samples is no less than the number n of I + identification parameters, HH, is nonsingular, then (H.H,) exists. At this time, the least square estimation of A is as fol- lows: i=(H!H,)) HZ, (26) the parameter obtained from estimation minimizes the quadrat- ic sum of the deviation of the formula, and minimizes an error of the overall parameter estimate, which is beneficial to reducing the influence from actual noise interference and measurement er- ror. After a batch of samples is obtained, the parameter identifi- cation is performed according to the samples obtained. This batch processing method is very suitable for off-line identification. When the sample size is enough and relatively accurate, the corre- sponding parameter estimates can be estimated.

FIG. 6 is a schematic diagram of a method for calibrating a one-dimensional force sensor in a legged robot's leg.

In order to avoid the influence of a gravity item in leg dy- namics on the calibration curve, the one-dimensicnal force sensor detection values of each joint driving unit with the foot end placed in an initial position and unloaded are subtracted from the one-dimensional force sensor detection values AE ARS and

AF PU == == | | $3} of each joint driving unit obtained from the robot leg driving system test platform in S1, and the one-dimensional force sensor simulated detection values AE AF and AFL of each joint driving unit obtained from the leg driving system vir- tual model in S2. The calibration curves of the hip joint ©, knee joint E and ankle joint G are fitted by formula (29), with the force sensor detection values ART AF” and AFG of each Joint driving unit obtained from the robot leg driving system test platform after the initial value subtracted as H,, and the force ID 3 IDO sensor simulated detection values (AF ; AFL and AFL of each joint driving unit obtained from the leg driving system vir- tual model after the initial value subtracted as Z,.

S5, according to the calibration curve of each driving unit and an original calibration coefficient of each driving unit, ob- taining a calibration correction coefficient of each driving unit, wherein the original calibration coefficient is a one-dimensional force sensor calibration coefficient upon delivery; according to the calibration correction coefficient of each driving unit, cali- brating the one-dimensional force sensor.

Specifically, the original calibration coefficient of the one-dimensional force sensor of each joint driving unit is multi- plied by a slope value of the corresponding calibration curve, and then an intercept of the corresponding calibration curve is added, thereby completing a calibration for the one-dimensional force sensor of each joint driving unit.

As an optional implementation mode, the calibration method of the present invention further includes the following step: verify- ing an effectiveness of the calibration method after calibrating the one-dimensional force sensor.

FIG. 7a)-c¢) are the one-dimensional force sensor calibration curves of the hip joint O, knee joint E and ankle joint G driving units of the legged robot leg driving system. The black solid lines are the curves formed by each successive point from the force sensor detection values in the leg driving system perfor- mance test platform as the abscissas, and the force sensor detec-

tion values of each joint driving unit in the virtual model as the ordinates; the dotted lines are the fitted curves obtained from black curves processed by a least square method, and defined as the force sensor calibration curves.

As shown in FIG. 7, the slope and intercept of the hip joint calibration curve are 2.463 and -7.452 respectively; the slope and intercept of the knee joint calibration curve are 2.147 and 2.703 respectively; the slope and intercept of the ankle joint calibra- tion curve are 2.338 and -0.4568 respectively; the linearity of fitting of each joint force sensor calibration curve is good, in- dicating that the original calibration coefficient of force sensor has some deviation. The original calibration coefficient of force sensor can be compensated through the calibration method proposed in the present invention.

As an optional implementation mode, the step of verifying an effectiveness of the calibration method further includes the fol- lowings: with the foot end placed in an initial position, verifying the effectiveness of the calibration method in x-axis direction and y-axis direction respectively, in the process of fast load- ing/unloading and slow loading/unloading; with the foot end placed in a non-initial position, verifying the effectiveness of the calibration method in x-axis direction and y-axis direction respectively, in the process of fast load- ing/unloading and slow loading/unloading.

Specifically, in order to test whether the force sensor com- pensation strategy proposed in the present invention is correct or not, the experiment plans are prepared as shown in Table 2.

Table 2 Test Plans Experimental 0 DirectionofForceon Leg Condition Foot End Position Foot End Change Plans the Foot End 1 Elevatedphase (0,0) OO caxsdirection Fast 2 Grounded phase (0, 0) y-axis direction Fast 3 Elevated phase (100, 100) x-axis direction Fast 4 Grounded phase (100, 100) y-axis direction Fast

Elevatedphase (0,0) xaxisdirection Slow 6 Grounded phase (0, 0) y-axis direction Slow 7 Elevated phase (109, 100) x-axis direction Slow 8 Grounded phase {100, 100) y-axis direction Slow © In Table 2, plans 1, 2, 5 and 6 are prepared to test the cor- rection effect of the correction strategy proposed in the present invention, in the process of fast loading/unloading and slow load- ing/unloading in the x/y-axis direction, with the foot end placed 5 in an initial position in the leg driving system; plans 3, 4, 7 and 8 are prepared to test whether the calibration method proposed in the present invention is accurate, in the process of fast load- ing/unloading and slow loading/unloading in the x/y-axis direc- tion, with the foot end placed in other position in the leg driv- ing system.

The statics model of the leg mechanical structure and the mapping relation between a force on each joint driving unit with a rotation angle and torque of each joint are required in calcula- tion of the contact force between the foot end and the ground.

Therefore, the calculation methods for the statics model of the leg mechanical structure and the mapping relation between a force on each joint driving unit with a rotation angle and torque of each joint will be specifically introduced below.

only the positive solution of the statics model is used in the application, therefore only the calculation method for statics positive solution will be introduced herein.

FIG. 8 is a sketch of a statics model of a simplified leg of a leg driving system .

According to the principle of virtual work, we can get: . r=J"(q)F (30) where F represents a force (N) on the foot end of the leg, J (g) represents a force Jacobian on the leg, and T represents a torque on each joint.

In formula (30), the force Jacobian J (gq) on the leg is as follows:

ox, ox, Joot “<t foot 29, 26 av avi jt (9) _ OA oen eK foot 00, 00, x f ox foot eK foot 79, 6, <t sin(68,)-4 sin(9, +6,) - t cos(9,) +1, cos(9, +8, ) + I,sin(6, +6, +8, ) | cos(6, +6, +0,) =| sin(8 +6, )-L, sin(9, +6, +6,) 1, cos(8, +6, ) +1, cos (4, +8, +6, ) —1,sin(6, +6, +6, ) I, cos(6, +9, +6, } (31) With formula (31) put into formula (30), we can get: pn Lsin(8,) +1, sin (6, +8, } + Lp L cos (6, )+1, cos(6, +6,)+ 7 =f foo . ’ foc | 1,sin(6, +6, +6,) | Lcos(8, +6, +6,) ox |g (6+ ar 0, + 7, ==F; | I,sin(6,+86,)+1 sin +15 | I cos(6, +6,) +1, cos ’ ) 0, +0, 8, +0, 7, == Lsin(6,+6,+6,)+F, I cos(6, +6, +6, ) (32) The force on the foot end is obtained through the torques on the knee joint E and ankle joint G. The statics positive solution can be obtained as follows: == rl, cos(6, +9, +0,) 1, (4, cos(9, +9, ) +1, cos (6, +6, +6,)) Soot = . Ll sin(8,) | a. 71; sin(8, +6, +6,) 7, (1, sin (6, +6,) +1, sin(9, +9, +6,)) foot .

Ll, sin(6,) (33) The calculation method for the mapping relation between a force on each joint driving unit with a rotation angle and torque of each joint will be specifically introduced below: FIG. %a)-c) show the relation between the torques on the hip joint O, knee joint E and ankle joint G and the forces on the driving units respectively.

In FIG. 9a) AAOB, ZABO can be obtained from a law of cosines combined with formula (15) as follows:

(hy +Ax,) +0B 04 ZABO = arccos| ——mFF———— 2-(l, + Ax, )-OB p (34)

From formula (34), the mapping relation between the force sensor detection signal AF, of the hip joint driving unit and the torque T; on the hip joint can be obtained as follows:

] (hy +Ax,,) +0B°-04

7, =AF OB -8In| arccos| =r

2-(I,,+Ax,,)-OB pl (35) In FIG. 9b)}) ACDE, ZDCE can be obtained from a law of cosines as follows: 2 ~2 2 (hy +Ax,,) + EC? —ED DCE = arccos ee TT TTT 2-(1, + Ax, )- EC (36)

From formula (36), the mapping relation between the force sensor detection signal AF: of the knee joint driving unit and the torque 1, on the knee joint can be obtained as follows:

2 132 Co (hy +Ax,,) + EC? ~ED r, =-AF -EC-sin| arccos| ———————— 2-(hy + Ax, EC (37)

In FIG.

Sc) AFGH, ZGFH can be obtained from law of cosines as follows: 2 ~~ ~ Bn (hs +Ax,,) +GF? GH"

ZGFH = arceos| ~—F—F—

2-(hs + Ax, )- GF

(38)

From formula (38), the mapping relation between the force sensor detection signal AF,3 of the ankle joint driving unit and the torque 13; on the ankle joint G can be obtained as follows:

2 7 Y + oo (4s +Ax,,) +GF° ~GH® r, = AF, GF -sin| arccos| ~——F———— 2: (Is + Ax} GF (39)

The contact force curves in loading direction in plans 1-8 are shown in FIGS. 10-17 respectively.

The actual contact force (# SCHSOFX J sensory )

foot ' foot J phetween the foot end and the ground directly detected by the two-dimensional force sensor is dotted line of “....”; the

( Jeorrecteds pp correctedy ) contact force \ fo Th foot between the foot end and the ground solved according to a corrected force sensor calibration curve of each joint driving unit, the statics model of the leg mechanical structure, as well as the mapping relation between a force on each joint driving unit with a rotation angle and torque of each joint is black solid line; the contact force \ #7 PJ between the foot end and the ground solved according to an uncorrected force sensor calibration curve of each joint driving unit, the statics model of the leg mechanical structure, as well as the mapping relation be- tween a force on each joint driving unit with a rotation angle and torque of each joint is dotted line of “-----*%. The present inven- tion adjusts the contact forces detected and solved to zero, in order to prevent the dynamics of the leg mechanical structure and the zero drift of force sensor from causing a non-zero initial value of a contact force solved, meanwhile in order to easily ob- serve the effect of the compensation strategy proposed in the pre- sent invention. From FIGS. 10-17, it can be seen that after the calibration method proposed in the present invention is added, the deviation between the actual foot end contract force and the cor- rected foot end contract force in loading direction is almost ze- ro, indicating that the one-dimensional force sensor calibration method proposed in the present invention is applicable in differ- ent positions. Example 2 The example provides a system for calibrating a one- dimensional force sensor in a legged robot's leg, as shown in FIG.

18. The system includes: a sensor detection value acquisition module, configured to use an upper computer to record two-dimensional force sensor de- tection values for loading, and displacement sensor detection val- ues and one-dimensional force sensor detection values of each Joint driving unit in a process of loading and unloading a force to a foot end; a simulation module, configured to input the displacement sensor detection values and the two-dimensional force sensor de- tection values into a leg driving system virtual model built based on a leg mechanical structure for simulation, and record one- dimensional force sensor simulated detection values of each joint driving unit in the virtual model; a variable value and sampling value acquisition module, con- figured to obtain actual detection values of each joint and simu- lated detection values of each joint, by subtracting the one- dimensional force sensor detection values of each joint (with the foot end placed in an initial position) from the one-dimensional force sensor detection values of each joint and the one- dimensional force sensor simulated detection values of each joint, wherein the initial position refers to no force loaded to the foot end; a calibration curve acquisition module, configured to solve a one-dimensional force sensor calibration curve of each driving unit by a least square method according to a linear relation be- tween each actual detection value and each simulated detection value; and a calibration module, configured to obtain a calibration cor- rection coefficient of each driving unit according to the calibra- tion curve of each driving unit and an original calibration coef- ficient of each driving unit, wherein the original calibration co- efficient is a one-dimensional force sensor calibration coeffi- cient upon delivery, and calibrate the one-dimensional force sen- sor according to the calibration correction coefficient of each driving unit.

As an optional implementation mode, the system further in- cludes a verification module, configured to verify an effective- ness of the calibration method after obtaining the calibrated co- efficient of each driving unit.

The above-mentioned embodiments describe only the preferred implementation modes of the present invention, rather than limit- ing the scope of the present invention. On the basis of not devi- ating from the design spirit of the present invention, any modifi- cations or improvements of the technical solutions of the present invention made by those of ordinary skill in the art should fall within the protection scope determined by the claims of the pre- sent invention.

The principle and implementation mode of the present inven- tion are described with specific embodiments.

The description of the above-mentioned embodiments is only used to help understand the method of the present invention and its core idea.

Meanwhile, both specific implementation mode and scope of application will be changed by those of ordinary skill in the art based on the idea of the present invention.

To sum up, the content of the specification should not be understood as a limitation to the present invention.

Claims

CONCLUSIESCONCLUSIONS

1. Werkwijze voor het kalibreren van een eendimensionale kracht- sensor in het been van een robot met poten, omvattende: in een proces van het laden en lossen van een kracht naar een voe- teneinde, het gebruik maken een bovenste computer voor het opnemen van tweedimensionale krachtsensordetectiewaarden voor het laden, verplaatsingssensordetectiewaarden van elke gewrichtsaandrijfeen- heid en eendimensionale krachtsensordetectiewaarden van elke ge- wricht aandrijvende eenheid; het invoeren van de detectiewaarden van de verplaatsingssensor en de tweedimensionale krachtsensordetectiewaarden in een virtueel model van een beenaandrijfsysteem gebouwd op basis van een mecha- nische beenstructuur voor simulatie, en het opnemen van eendimen- sionale gesimuleerde krachtsensordetectiewaarden van elke gewricht aandrijvende eenheid in het virtuele model; het verkrijgen van werkelijke detectiewaarden van elk gewricht en gesimuleerde detectiewaarden van elk gewricht, door de eendimensi- onale krachtsensordetectiewaarden van elk gewricht (met het voe- teneinde in een beginpositie) af te trekken van de eendimensionale krachtsensordetectiewaarden van elk gewricht en de eendimensionale gesimuleerde krachtsensordetectiewaarden van elk gewricht, waarbij de beginpositie verwijst naar geen kracht die op het voeteneinde wordt uitgeoefend; het, in overeenstemming met een lineaire relatie tussen elke wer- kelijke detectiewaarde en elke gesimuleerde detectiewaarde, oplos- sen van een eendimensionale krachtsensorkalibratiecurve van elke aandrijfeenheid door middel van een kleinste-kwadratenmethode; het, in overeenstemming met de kalibratiecurve van elke aandrijf- eenheid en een oorspronkelijke kalibratieco&fficiént van elke aan- drijfeenheid, verkrijgen van een kalibratiecorrectiecoëfficiënt van elke aandrijfeenheid, waarbij de oorspronkelijke kalibratieco- efficiënt een eendimensionale krachtsensorkalibratiecoëfficiënt is bij levering; en het, in overeenstemming met de kalibratiecorrectiecoëfficiënt van elke aandrijfeenheid, kalibreren van de eendimensionale krachtsen- sor.A method of calibrating a one-dimensional force sensor in the leg of a robot with legs, comprising: in a process of loading and unloading a force to a feeder, using an upper computer to record two-dimensional force sensor detection values for loading, displacement sensor detection values of each joint driving unit and one-dimensional force sensor detection values of each joint driving unit; inputting the displacement sensor detection values and the two-dimensional force sensor detection values into a virtual model of a leg drive system built on the basis of a mechanical bone structure for simulation, and incorporating one-dimensional simulated force sensor detection values of each joint driving unit in the virtual model; obtaining actual detection values of each joint and simulated detection values of each joint, by subtracting the one-dimensional force sensor detection values of each joint (with the foot in an initial position) from the one-dimensional force sensor detection values of each joint and the one-dimensional simulated force sensor detection values of any joint, where the initial position refers to no force applied to the foot end; according to a linear relationship between each actual detection value and each simulated detection value, solving a one-dimensional force sensor calibration curve of each drive unit by a least squares method; obtaining, in accordance with the calibration curve of each drive unit and an original calibration coefficient of each drive unit, a calibration correction coefficient of each drive unit, the original calibration coefficient being a one-dimensional force sensor calibration coefficient upon delivery; and calibrating the one-dimensional force sensor in accordance with the calibration correction coefficient of each drive unit.

2. Werkwijze volgens conclusie 1, waarbij de stap van het gebrui- ken van een bovenste computer voor het registreren van tweedimen- sionale krachtsensordetectiewaarden voor het laden, en verplaat- singssensordetectiewaarden van elke gewricht aandrijvende eenheid en eendimensionale krachtsensordetectiewaarden van elke gewricht aandrijvende eenheid in een proces van het laden en lossen van een kracht op een voeteneinde verder de volgende stappen omvat: het langzaam met de hand trekken van een houten blok met de twee- dimensionale krachtsensor bevestigd om een kracht op het voeten- einde te laden, het langzaam lossen van de kracht nadat een vooraf ingestelde trekkracht is bereikt, en het opnemen van de tweedimen- sionale krachtsensordetectiewaarden, en de verplaatsingssensorde- tectiewaarden en eendimensionale krachtsensordetectiewaarden van elke gewricht aandrijvende eenheid tijdens het laden en lossen.The method of claim 1, wherein the step of using an upper computer to record two-dimensional force sensor detection values for loading, and displacement sensor detection values of each joint driving unit and one-dimensional force sensor detection values of each joint driving unit in a process of loading and unloading a force on a footboard further includes the following steps: slowly pulling a wooden block by hand with the two-dimensional force sensor attached to load a force on the footboard, slowly unloading the force after a preset tensile force is reached, and recording the two-dimensional force sensor detection values, and the displacement sensor detection values and one-dimensional force sensor detection values of each joint driving unit during loading and unloading.

3. Werkwijze volgens conclusie 1 of 2, waarbij de werkwijze verder de volgende stappen omvat: vóór het laden en lossen van een kracht op het voeteneinde, het, in overeenstemming met een afmeting en geometrische relatie van de mechanische beenstructuur, bepalen van een mappingrelatie tussen een lengte van elke gewrichtsaandrijf- eenheid en een rotatiehoek van elk gewricht; en het, in overeenstemming met de mappingrelatie en een kinematicamo- del verkregen uit de mechanische structuur van het been, gebruik maken van een bovenste computer om het beenaandrijfsysteem in een initiële positie te besturen.A method according to claim 1 or 2, wherein the method further comprises the steps of: before loading and unloading a force on the foot end, determining, in accordance with a size and geometric relationship of the mechanical leg structure, a mapping relationship between a length of each joint drive unit and a rotation angle of each joint; and in accordance with the mapping relationship and a kinematic model obtained from the mechanical structure of the leg, using an upper computer to control the leg drive system in an initial position.

4. Werkwijze volgens conclusie 3, waarbij een werkwijze voor het verkrijgen van een kinematicamodel verder de volgende stappen om- vat: het vereenvoudigen van de mechanische beenstructuur van een robot met poten en het verkrijgen van parameters van een mechanische structuur met één been; het bepalen van een positieve oplossing van de mechanische struc- tuur van het been volgens de parameters van de mechanische struc- tuur met één been, waarbij de positieve oplossing het oplossen van een functie van een positie van het voeteinde ten opzichte van het heupgewricht is volgens een rotatiehoek van het heupgewricht, ro- tatiehoek van het kniegewricht, rotatiehoek van het enkelgewricht, dijlengte, scheenbeenlengte en voeteindlengte; het bepalen van een inverse oplossing van de mechanische structuur van het been volgens de parameters van de mechanische structuur met één been, waarbij de inverse oplossing het oplossen van func- ties van de rotatiehoek van het heupgewricht, de rotatiehoek van het kniegewricht en de rotatiehoek van het enkelgewricht is vol- gens de positie van het voeteneinde ten opzichte van het heupge- wricht, dijlengte, scheenbeenlengte en voeteindelengte; en het verkrijgen van het kinematicamodel volgens de positieve oplos- sing en de inverse oplossing.The method of claim 3, wherein a method of obtaining a kinematics model further comprises the steps of: simplifying the leg mechanical structure of a legged robot and obtaining parameters of a single leg mechanical structure; determining a positive solution of the mechanical structure of the leg according to the parameters of the single-leg mechanical structure, wherein the positive solution is solving a function of a position of the foot end relative to the hip joint according to a hip joint rotation angle, knee joint rotation angle, ankle joint rotation angle, thigh length, shin length and foot end length; determining an inverse solution of the mechanical structure of the leg according to the parameters of the single-leg mechanical structure, the inverse solution solving functions of the angle of rotation of the hip joint, the angle of rotation of the knee joint, and the angle of rotation of the ankle joint is according to the position of the foot end relative to the hip joint, thigh length, shin length and foot end length; and obtaining the kinematics model according to the positive solution and the inverse solution.

5. Werkwijze volgens conclusie 1, waarbij het virtuele model van het beenaandrijfsysteem wordt gebouwd via SimMechanics Link-tools in Solidworks.The method of claim 1, wherein the virtual model of the leg drive system is built via SimMechanics Link tools in Solidworks.

6. Werkwijze volgens conclusie 1, waarbij de kalibratiemethode verder de volgende stap omvat: het verifiëren van de effectiviteit van de kalibratiemethode na het kalibreren van de eendimensionale krachtsensor.The method of claim 1, wherein the calibration method further comprises the step of: verifying the effectiveness of the calibration method after calibrating the one-dimensional force sensor.

7. Werkwijze volgens conclusie 6, waarbij de stap van het verifi- eren van een effectiviteit van de kalibratiemethode verder de vol- gende stappen omvat: het met het voeteinde in een beginpositie, verifiëren van de ef- fectiviteit van de kalibratiemethode in respectievelijk x- asrichting en y-asrichting, tijdens het proces van snel la- den/lossen en langzaam laden/lossen.The method of claim 6, wherein the step of verifying an effectiveness of the calibration method further comprises the steps of: with the foot end in an initial position, verifying the effectiveness of the calibration method in x- axis direction and y-axis direction, during the process of fast loading/unloading and slow loading/unloading.

8. Werkwijze volgens conclusie 6, waarbij de stap van het verifi- eren van een effectiviteit van de kalibratiewerkwijze verder de volgende stappen omvat: het met het voeteinde in een niet-initiële positie, verifiëren van de effectiviteit van de kalibratiemethode in respectievelijk x- asrichting en y-asrichting, tijdens het proces van snel la- den/lossen en langzaam laden/lossen.The method of claim 6, wherein the step of verifying an effectiveness of the calibration method further comprises the steps of: with the foot end in a non-initial position, verifying the effectiveness of the calibration method in x-axis direction, respectively and y-axis direction, during the process of fast loading/unloading and slow loading/unloading.

9. Systeem voor het kalibreren van een eendimensionale krachtsen-9. System for calibrating a one-dimensional force

sor in het been van een robot met poten, omvattende:sor in the leg of a robot with legs, comprising:

een sensordetectiewaarde-acquisitiemodule, geconfigureerd voor het gebruiken van een bovenste computer voor het opnemen van tweedi-a sensor detection value acquisition module configured to use a top computer to record two-dimensional

mensionale krachtsensordetectiewaarden voor het laden, verplaat-mensional force sensor sensing values for loading, moving

singssensordetectiewaarden van elke gewricht aandrijvende eenheid en eendimensionale krachtsensordetectiewaarden van elke gewricht aandrijvende eenheid in een proces van het laden en lossen van een kracht naar een voeteneinde; een simulatiemodule, geconfigureerd voor het invoeren van de de- tectiewaarden van de verplaatsingssensor en de detectiewaarden van de tweedimensionale krachtsensor iin een virtueel model van een beenaandrijfsysteem gebouwd op basis van een mechanische been-sing sensor detection values of each joint driving unit and one-dimensional force sensor detection values of each joint driving unit in a process of loading and unloading a force to a foot end; a simulation module configured to input the displacement sensor detection values and the two-dimensional force sensor detection values into a virtual model of a leg drive system built on the basis of a mechanical leg drive

structuur voor simulatie, en het opnemen van eendimensionale gesi- muleerde detectiewaarden van de krachtsensor van elke gewricht aandrijvende eenheid in het virtuele model; een module voor het verkrijgen van variabele waarde en bemonste- ringswaarde, geconfigureerd om werkelijke detectiewaarden van elk gewricht en gesimuleerde detectiewaarden van elk gewricht te ver- krijgen, door de eendimensionale krachtsensordetectiewaarden van elk gewricht (met het voeteinde in een beginpositie geplaatst) af te trekken van de eendimensionale krachtsensordetectiewaarden van elk gewricht en de gesimuleerde eendimensionale krachtsensor de-structure for simulation, and including one-dimensional simulated detection values of the force sensor of each joint driving unit in the virtual model; a variable value and sample value acquisition module configured to obtain actual detection values of each joint and simulated detection values of each joint by subtracting the one-dimensional force sensor detection values of each joint (with the foot end placed in an initial position) of the one-dimensional force sensor detection values of each joint and the simulated one-dimensional force sensor de-

tectiewaarden van elk gewricht, waarbij de initiële positie ver- wijst naar geen kracht die op het voeteinde wordt geladen; een kalibratiecurve-acquisitiemodule, geconfigureerd om een eendimensionale krachtsensor-kalibratiecurve van elke aandrijfeen- heid op te lossen door middel van een kleinste-kwadratenmethode in overeenstemming met een lineaire relatie tussen elke werkelijke detectiewaarde en elke gesimuleerde detectiewaarde; en een kalibratiemodule, geconfigureerd om een kalibratiecorrectiecoëfficiënt van elke aandrijfeenheid te ver- krijgen in overeenstemming met de kalibratiecurve van elke aan-tection values of each joint, where the initial position refers to no force being loaded on the foot end; a calibration curve acquisition module configured to resolve a one-dimensional force sensor calibration curve of each drive unit by a least squares method in accordance with a linear relationship between each actual detection value and each simulated detection value; and a calibration module configured to obtain a calibration correction coefficient from each drive unit in accordance with the calibration curve of each

drijfeenheid en een oorspronkelijke kalibratiecoëfficiënt van elke aandrijfeenheid, waarbij de oorspronkelijke kalibratiecoëfficiënt een eendimensionale kalibratiecoëfficiënt van de krachtsensor is bij levering, en het kalibreren van de eendimensionale krachtsen- sor in overeensteming met de kalibratiecorrectiecoëfficiënt van elke aandrijfeenheid.drive unit and an original calibration coefficient of each drive unit, the original calibration coefficient being a one-dimensional calibration coefficient of the force sensor upon delivery, and calibrating the one-dimensional force sensor in accordance with the calibration correction coefficient of each drive unit.

10. Systeem volgens conclusie 9, waarbij het systeem verder omvat: een verificatiemodule, geconfigureerd om de effectiviteit van de kalibratiemethode te verifiëren na het verkrijgen van de gekali- breerde coëfficiënt van elke aandrijfeenheid.The system of claim 9, wherein the system further comprises: a verification module configured to verify the effectiveness of the calibration method after obtaining the calibrated coefficient of each drive unit.