Skip to main content
Log in

Kinetostatic analysis for compliant legged robots with ground contact forces evaluation

  • Machines
  • Published:
Meccanica Aims and scope Submit manuscript

Abstract

This paper presents a systematic proposal for the kinetostatic analysis in compliant legged robots. The proposal applies the screw theory, Assur virtual chains, and Davies method as an integrated approach, including kinematics and statics during the quasi-static movement of n-legged robots. The general proposed procedure applies to robots with any number of legs and any degree of stiffness to evaluate the ground contact forces that emerge from the addition of compliant elements. The external forces on the torso and feet on the ground are analyzed, including the compliant effects. Numerical simulations validate the proposal, comparing the performance of planar legged robots during quasi-static movements, including two-legged robots without compliant elements (rigid robots) and composed of compliant elements on every joint. The analysis yields an index that features the kinetostatic performance and indicates a path to developing compensation strategies.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10
Fig. 11
Fig. 12

Similar content being viewed by others

Notes

  1. For generalization purposes, other types of joints can always be expanded in a set of rotary (revolution joints) or linear (prismatic joints) displacements.

  2. Contact Points (CP) are the coordinates of the ball-foot of each leg with respect to the reference frame.

References

  1. Bruzzone L, Quaglia G (2012) Review article: locomotion systems for ground mobile robots in unstructured environments. Mech Sci (Gött) 3(2):49–62. https://doi.org/10.5194/ms-3-49-2012

    Article  Google Scholar 

  2. Gehring C, Coros S, Hutter M, Dario Bellicoso C, Heijnen H, Diethelm R, Bloesch M, Fankhauser P, Hwangbo J, Hoepflinger M, Siegwart R (2016) Practice makes perfect: an optimization-based approach to controlling agile motions for a quadruped robot. IEEE Robot Autom Mag 23(1):34–43. https://doi.org/10.1109/MRA.2015.2505910

    Article  Google Scholar 

  3. Poulakakis I, Smith JA, Buehler M (2005) Modeling and experiments of untethered quadrupedal running with a bounding gait: the scout II robot. Int J Robot Res 24(4):239–256. https://doi.org/10.1177/0278364904050917

    Article  Google Scholar 

  4. Hutter M, Gehring C, Bloesch M, Hoepflinger MA, Remy CD, Siegwart R (2012) STARLETH: a compliant quadrupedal robot for fast, efficient, and versatile locomotion. Autonomous Systems Lab, ETH Zurich, Zürich. 15th international conference on climbing and walking robot (CLAWAR 2012); Conference Location: Maryland, USA; Conference Date: July 23-26, 2012. https://doi.org/10.3929/ethz-a-010034688

  5. Hutter M, Gehring C, Jud D, Lauber A, Bellicoso CD, Tsounis V, Hwangbo J, Bodie K, Fankhauser P, Bloesch M, Diethelm R, Bachmann S, Melzer A, Hoepflinger M (2016) ANYmal—a highly mobile and dynamic quadrupedal robot. In: 2016 IEEE/RSJ international conference on intelligent robots and systems (IROS), pp 38–44. https://doi.org/10.1109/IROS.2016.7758092

  6. Spröwitz A, Tuleu A, Vespignani M, Ajallooeian M, Badri E, Ijspeert AJ (2013) Towards dynamic trot gait locomotion: design, control, and experiments with Cheetah-cub, a compliant quadruped robot. Int J Robot Res 32(8):932–950. https://doi.org/10.1177/0278364913489205

    Article  Google Scholar 

  7. Weihmann L (2011) Modelagem e otimização de forças e torques aplicados por robôs com redundância cinemática e de atuação em contato com o meio (in portuguese). PhD thesis, Post-Graduate Program in Mechanical Engineering, Federal University of Santa Catarina, Florianópolis, Brasil. http://www.tede.ufsc.br/teses/PEMC1303-T.pdf

  8. Nokleby SB, Fisher R, Podhorodeski RP, Firmani F (2005) Force capabilities of redundantly-actuated parallel manipulators. Mech Mach Theory 40(5):578–599. https://doi.org/10.1016/j.mechmachtheory.2004.10.005

    Article  MATH  Google Scholar 

  9. Zhou Q, Yu Z, Zhang S, Chen X, Qin M, Zhang W, Huang Q (2018) Simultaneous prevention of rotational and translational slip for a humanoid robot. Appl Sci. https://doi.org/10.3390/app8091554

    Article  Google Scholar 

  10. Hou W, Zhang T, Chen Y, Ma H (2016) Compliant biped walking on uneven terrain with point feet. Int J Adv Robot Syst 13(2):51. https://doi.org/10.5772/62330

    Article  Google Scholar 

  11. Takemura H, Deguchi M, Ueda J, Matsumoto Y, Ogasawara T (2005) Slip-adaptive walk of quadruped robot. Robot Auton Syst 53(2):124–141. https://doi.org/10.1016/j.robot.2005.07.002

    Article  Google Scholar 

  12. Kwon O, Park JH (2002) Reflex control of bipedal locomotion on a slippery surface. Adv Robot 16(8):721–734. https://doi.org/10.1163/15685530260425710

    Article  Google Scholar 

  13. Brandão M, Hashimoto K, Santos-Victor J, Takanishi A (2016) Footstep planning for slippery and slanted terrain using human-inspired models. IEEE Trans Robot 32(4):868–879. https://doi.org/10.1109/TRO.2016.2581219

    Article  Google Scholar 

  14. Kaneko K, Kanehiro F, Kajita S, Morisawa M, Fujiwara K, Harada K, Hirukawa H (2005) Slip observer for walking on a low friction floor. In: 2005 IEEE/RSJ international conference on intelligent robots and systems, pp 634–640. https://doi.org/10.1109/IROS.2005.1545184

  15. Bellicoso CD, Bjelonic M, Wellhausen L, Holtmann K, Günther F, Tranzatto M, Fankhauser P, Hutter M (2018) Advances in real-world applications for legged robots. J Field Robot 35(8):1311–1326. https://doi.org/10.1002/rob.21839, https://onlinelibrary.wiley.com/doi/pdf/10.1002/rob.21839

  16. Ambe Y, Matsuno F (2016) Leg-grope walk: strategy for walking on fragile irregular slopes as a quadruped robot by force distribution. ROBOMECH J 3(7):2197–4225. https://doi.org/10.1186/s40648-016-0046-2

    Article  Google Scholar 

  17. Motoi N, Ikebe M, Ohnishi K (2007) Real-time gait planning for pushing motion of humanoid robot. IEEE Trans Ind Inform 3(2):154–163. https://doi.org/10.1109/TII.2007.898469

    Article  Google Scholar 

  18. Zhao F, Gao J (2019) Anti-slip gait planning for a humanoid robot in fast walking. Appl Sci. https://doi.org/10.3390/app9132657

    Article  Google Scholar 

  19. Werner A, Turlej W, Ott C (2017) Generation of locomotion trajectories for series elastic and viscoelastic bipedal robots. In: 2017 IEEE/RSJ international conference on intelligent robots and systems (IROS), pp 5853–5860. https://doi.org/10.1109/IROS.2017.8206476

  20. Bowling A (2011) Impact forces and agility in legged robot locomotion. J Vib Control 17(3):335–346. https://doi.org/10.1177/1077546309357855

    Article  MathSciNet  MATH  Google Scholar 

  21. Corral E, García MJG, Castejon C, Meneses J, Gismeros R (2020) Dynamic modeling of the dissipative contact and friction forces of a passive biped-walking robot. Appl Sci. https://doi.org/10.3390/app10072342

    Article  Google Scholar 

  22. Erden MS, Leblebicioglu K (2007) Torque distribution in a six-legged robot. IEEE Trans Robot 23(1):179–186. https://doi.org/10.1109/TRO.2006.886276

    Article  Google Scholar 

  23. Machado M, Moreira P, Flores P, Lankarani HM (2012) Compliant contact force models in multibody dynamics: evolution of the Hertz contact theory. Mech Mach Theory 53:99–121. https://doi.org/10.1016/j.mechmachtheory.2012.02.010

    Article  Google Scholar 

  24. Hu Y, Guo W (2021) A new concept of contact joint to model the geometric foot-environment contacts for efficiently determining possible stances for legged robots. Mech Mach Theory 162:104327. https://doi.org/10.1016/j.mechmachtheory.2021.104327

    Article  Google Scholar 

  25. Hunt KK (2000) Don’t Cross-thread the Screw. In: a symposium commemorating the legacy, works, and life of Sir Robert Stawell Ball upon the 100th anniversary of a treatise on the theory of screws, University of Cambridge, Trinity College, Ball 2000 Conference. pp 1–37. https://doi.org/10.1002/rob.10095

  26. Ceccarelli M (2000) Screw axis defined by Giulio Mozzi in 1763 and early studies on helicoidal motion. Mech Mach Theory 35(6):761–770. https://doi.org/10.1016/S0094-114X(99)00046-4

    Article  MATH  Google Scholar 

  27. Dai JS (2006) An historical review of the theoretical development of rigid body displacements from Rodrigues parameters to the finite twist. Mech Mach Theory 41(1):41–52. https://doi.org/10.1016/j.mechmachtheory.2005.04.004

    Article  MathSciNet  MATH  Google Scholar 

  28. Campos A, Guenther R, Martins D (2005) Differential kinematics of serial manipulators using virtual chains. J Braz Soc Mech Sci Eng XXVII(4):345–356. https://doi.org/10.1590/S1678-58782005000400002

    Article  Google Scholar 

  29. Simas H, Guenther R, da Cruz DFM, Martins D (2009) A new method to solve robot inverse kinematics using Assur virtual chains. Robotica 27(7):1017–1026. https://doi.org/10.1017/S0263574709005426

    Article  Google Scholar 

  30. Davies TH (1981) Kirchhoff’s circulation law applied to multi-loop kinematic chains. Mech Mach Theory 16(3):171–183. https://doi.org/10.1016/0094-114X(81)90033-1

    Article  Google Scholar 

  31. Davies TH (1995) Couplings, coupling networks and their graphs. Mech Mach Theory 30(7), 991–1000. https://doi.org/10.1016/0094-114X(95)00023-R. Graphs and mechanics first international conference

  32. Nise NS (2011) Control systems engineering, 6th edn. John Wiley & Sons Inc, United States of America

    MATH  Google Scholar 

  33. Acevedo MLR (2012) Mecanismos com desacoplamento cinetoestático para substituição de atuadores robóticos. Master’s thesis, Universidade Federal de Santa Catarina, Florianópolis, Brasil (fevereiro 2012). https://repositorio.ufsc.br/bitstream/handle/123456789/101044/307711.pdf

  34. Beer F, Russell Johnston EJ, Eisenberg E, Mazurek D (2009) Vector mechanics for engineers: statics. McGraw-Hill Education, New York

    Google Scholar 

  35. Popov VL (2010) Contact mechanics and friction: physical principles and applications. Mechanisms and Machine Science. Springer, Berlin

  36. Sliney HE, Spalvins T (1991) The Effect of Ion Plated Silver and Sliding Friction on Tensile Stress-induced Cracking in Aluminum Oxide. In: Annual Meeting of the Society of Tribologists and Lubrication Engineers, Philadelphia, PA, National Aeronautics and Space Administration. pp 1–15

  37. Campos A (2004) Cinemática diferencial de manipuladores empregando cadeias virtuais. PhD thesis, Universidade Federal de Santa Catarina, Florianópolis-SC (março 2004). http://www.tede.ufsc.br/teses/PEMC0776.pdf

  38. dos Santos CHF, Guenther R, Martins D, Pieri ERD (2006) Virtual kinematic chains to solve the underwater vehicle-manipulator systems redundancy. J Braz Soc Mech Sci Eng 28(3):354–361. https://doi.org/10.1590/S1678-58782006000300014

    Article  Google Scholar 

  39. Guenther R, Cruz D, Martins D, Simas H (2008) A new integration method for differential inverse kinematics of closed-chain robots. In: Proceedings of the ABCM symposium series in mechatronics, ABCM, Brasilia-DF. vol. 3, pp 225–235

  40. Rocha CR, Tonetto CP, Dias A (2011) A comparison between the Denavit-Hartenberg and the screw-based methods used in kinematic modeling of robot manipulators. Robotics and Computer-Integrated Manufacturing 27(4), 723–728. https://doi.org/10.1016/j.rcim.2010.12.009. Conference papers of Flexible Automation and Intelligent Manufacturing

  41. Cazangi HR (2008) Aplicação do método de Davies para analise cinemática e estática de mecanismo com múltiplos graus de liberdade. Master’s thesis, Universidade Federal de Santa Catarina, Florianópolis. http://www.tede.ufsc.br/teses/PEMC1080-D.pdf

  42. Tonetto CP, Simas H (2021) Analysis of compliant joints for planar legged robots. In: proceedings of the 26th international congress of mechanical engineering—COBEM 2021, Florianópolis, Brazil. https://doi.org/10.26678/ABCM.COBEM2021.COB2021-0862

  43. Tsai L-W (1999) Robot analysis: the mechanics of serial and parallel manipulators. Jonh Wiley & Sons, INC, New York

    Google Scholar 

  44. Tonetto CP, Bento Filho A, Dias A (2018) Modeling of a four-legged robot with variable center of mass as a cooperative multirobot system. In: Carvalho JCM, Martins D, Simoni R, Simas H (eds) Multibody mechatronic systems. Springer, Cham, pp 116–127

    Chapter  Google Scholar 

  45. Toscano GS, Simas H, Castelan EB, Martins D (2018) A new kinetostatic model for humanoid robots using screw theory. Robotica 36(4):570–587. https://doi.org/10.1017/S0263574717000595

    Article  Google Scholar 

Download references

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Cristiane Pescador Tonetto.

Ethics declarations

Conflict of interest

The authors declare that they have no conflict of interest.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Appendices

Appendix A: fundamentals tools

In this work, the following fundamental tools were applied to develop the kinematics, statics, and compliant elements integration: graph theory, screw theory, Assur virtual chains, and Davies method. These theories have a good track of applications and developments, such as in the research presented on [25,26,27,28,29,30,31, 37,38,39,40,41,42].

The graph theory is applied to represent the topological structure of the mechanisms (coupling graph, for kinematics) and the actions graph (for statics analysis) [30, 31].

Screw theory provides a mathematical representation of the movements and constraints of the system in which the mechanism is included. The use of screws allows a single representation for both kinematics and statics of bodies in space. It has the unique feature of being a geometric element that can represent mechanical properties such as angular, translation speeds and torques, moments over a body [25,26,27, 40].

The instantaneous movement of a rigid body related to an inertial coordinate system is here described by screws, called twists. This \({\$}^M\) twist is presented in Eq. (13). The instantaneous displacement \({\$}^M\) is decomposed in the magnitude \(\varphi\) and the normalized axis \(\hat{\$}{^M}\).

$$\begin{aligned} \$^M= \begin{bmatrix} \vec {\omega }\\ \vec {S_o}\times \vec {\omega } + h\vec {\omega } \end{bmatrix}= \begin{bmatrix} \vec {S}^M\\ \vec {S_o}\times \vec {S}^M + h\vec {S}^M \end{bmatrix}\varphi = \hat{\$}{^M}\varphi \end{aligned}$$
(13)

in which \(\omega\) is a vector with the angular speeds, \(\vec {S}_o\) is the position vector for one point in the screw axis, and the unitary vector \(\vec {S}^M\) represents the screw axis. The revolution and translation are related according to a screw pitch h [25, 41, 43].

The screw applied on statics analysis is called wrench \(\$^A\) and is composed by the action state over a body that is referenced to an inertial coordinate system, as in Eq. (14). By normalizing the wrench \(\$^A\), the magnitude \(\psi\) and normalized axis \(\hat{\$}{^A}\) are defined.

$$\begin{aligned} \$^A= \begin{bmatrix} \vec {S_o}\times \vec {R} + h\vec {R}\\ \vec {R}\\ \end{bmatrix}= \begin{bmatrix} \vec {S_o}\times \vec {S}{^A} + h\vec {S}{^A}\\ \vec {S}{^A} \end{bmatrix}\psi = \hat{\$}{^A}\psi \end{aligned}$$
(14)

in which \(\vec {R}\) is the resultant force, \(\vec {S_o}\) is the position vector for the point in the screw axis, the unitary vector \(\vec {S}{^A}\) represents the screw axis, and h is the pitch [25, 41].

The virtual chains are added to the real kinematic chains to close the kinematic circuit. As such, the legged robots is kinematically modelled as a parallel mechanism (closed kinematic chain) [29]. According to the definition, a virtual chain is an open kinematic chain composed of links and joints. Twists that represent movement in virtual chains are linearly independent. With the addition of such a virtual chain, the degrees of freedom of the real kinematic chain are unmodified. One example for planar virtual chains is PPR (two prismatic and one revolute joint) and for spatial virtual chains is 3P3R (three prismatic and one spherical joint) [28, 29, 37, 38].

The Davies method relates the links and kinematic pairs of the mechanism by applying the Kirchhoff laws, so building a set of equations that provide a solution for kinematics and statics in mechanisms [29, 30, 39, 41]. In general, the constraint equation is given by Eq. (15).

$$\begin{aligned} N\dot{q}= & {} \overrightarrow{0} \qquad (\text {kinematics})\nonumber \\ \hat{A}_N\Psi= & {} \overrightarrow{0} \qquad (\text {statics}) \end{aligned}$$
(15)

in which N is the network matrix that includes the normalized twists (\(\hat{\$}^M\)) and \(\dot{q}\) is the vector with the speeds magnitudes. \(\hat{A}_N\) is the action matrix for network unitary actions twists (\(\hat{\$}^A\)) and \(\Psi\) is the action magnitudes vector.

The partitioning of the matrices and isolating the components related to secondary joints (with unknown magnitudes) from the ones related to primary joints (with known magnitudes) leads to Eq. (16).

$$\begin{aligned} N_s\dot{q}_s= & {} -N_p\dot{q}_p \qquad (\text {kinematics}) \nonumber \\ \hat{A}_{NS}\Psi _{S}= & {} -\hat{A}_{NP}\Psi _{P} \quad (\text {static}) \end{aligned}$$
(16)

If the \(N_s\) and \(\hat{A}_{NS}\) are invertible, the magnitudes of the speeds and actions of the secondary joints \(\dot{q}_s\) and \(\Psi _{S}\) are computed by applying Eq. (17).

$$\begin{aligned} \dot{q}_s= & {} -N_s^{-1}N_p\dot{q}_p \qquad (\text {kinematics})\nonumber \\ \Psi _{S}= & {} -\hat{A}_{NS}^{-1}\hat{A}_{NP}\Psi _{P} \quad (\text {static}) \end{aligned}$$
(17)

Appendix B: kinematic model

The torso is the legged robot’s main part, and its displacement is defined as an input for the modeling. The model is based on screw theory as presented in [44]. Figure 13 presents the graph representation for a general legged robot with n legs. In this approach, the movement of the torso is defined by a virtual Assur chain named \(\$^M_{V_0}\), which imposes kinematic conditions (positions and orientations) on the torso. This kinematic chain is one edge in the graph representation of multiple joints (3 revolute and 3 prismatic joints (3P3R), in this case).

Fig. 13
figure 13

Kinematics graph representation for a general legged robot with n legs

Each one of the legs is represented as a kinematic chain (\(\$^M_i, i=1,2,...,n\)), and the displacement of each point of contact related to the ground reference is defined by the respective virtual 3P3R Assur chain (\(\$^M_{V_i}, i=1,2,...,n\)). One kinematic chain, such as the leg’s or virtual Assur chain, is composed of many screws, representing each joint that produces movements on the bodies (or impose movement, in the sense of virtual chains).

There is a pattern on the modeling for legged robots: the torso movements are defined by one kinematic chain (\(\$^M_{V_0}\)), and from the torso, there is one kinematic loop for each one of the multiple legs, passing through \(\$^M_i\) and \(\$^M_{V_i}\).

The kinematic analysis presented here is made with all legs touching the ground without changing foot or point of contact. For kinematic, as shown, all the \(\$^M_{V_i}\) chains are null at this point of the analysis.

Appendix C: statics model

The representation of a compliant joint in statics must include the forces and moments of the elements. The coupling between two adjacent elements is composed of a wrench with six static actions. One of these six components is active (resulting from a motor-actuated action), and the others are passive. For the graph representation, the pairs of wrenches of the joint are simplified into one single edge. The graph elements for a compliant joint are presented in Fig. 14.

Figure 14 presents a schematic for a revolute joint and a prismatic joint with a compliant element. In each one of these joints, an input torque \(\tau _{ij}\) is defined due to the motor actuation of the revolute joint or a force \(F_{ij}\) that is imposed by the prismatic joint. The angle of the revolute joint is indicated by \(\theta _{ij}\) and the position of the prismatic joint is named \(d_{ij}\). The displacement due to the compliant element indicated by \(\delta _{ij}\) is defined by the ratio \(\delta _{ij} = \tau _{ij}/k\) (revolute joint) or \(\delta _{ij} = F_{ij}/k\) (prismatic joint), in which k is the value of the stiffness constant. The total angle or linear movement produced by the joint would be the sum of the actuator movement and the displacement of the compliant element (which can be positive or negative, as it is a function of the torques and forces over the compliant element).

Fig. 14
figure 14

Compliant joint schematics with graph representation for revolute and prismatic joints

For the whole robot, there is a wrench \(\$^A_{V0}\) representing the external forces on the torso, as well as one wrench \(\$^A_{Vi}\) for each point of contact (foot). The contact to the ground is made by a point in a rough surface, which means that the contact forces imply constraints, but no moment constraint is defined [24, 34]. The edge from the feet to the torso \(\$^A_{i}\) is the compact form of the wrench chain and includes all the compliant joints for each leg (Fig. 15).

The static computation is made by defining the fundamental cutsets (k-cuts) and building the network action matrix from these k-cuts. One k-cut is made including the external forces and moments (\(\$^A_{V0}\) and \(\$^A_{Vi}\)), and, for each leg, there is two k-cuts for each joint (one before the compliant element, and one after), with a total of 2m k-cuts, with m being the number of joints of each leg. Figure 15 presents the arrangement for this model, following the methodology proposed by [45].

Fig. 15
figure 15

Statics model (graph representation with k-cuts)

The statics model is applied to determine the value of the variables on the joints. The input for the model analysis is the imposed positions given by the kinematics and the external forces and moments. The joint position is essential in defining the torques distribution, so the kinematic outputs are also used as inputs on the statics model.

As a result of this dependence, the ordering of the operations is essential, as presented in [42, 45]: first, the kinematics model is evaluated, and afterward, the statics.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Tonetto, C.P., Simas, H. Kinetostatic analysis for compliant legged robots with ground contact forces evaluation. Meccanica 57, 2379–2393 (2022). https://doi.org/10.1007/s11012-022-01555-z

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11012-022-01555-z

Keywords

Navigation