- Email: [email protected]
Design of walking machines - control aspects
DESIGN OF WALKING MACHINES - CONTROL ASPECTS ...
14th World Congress ofTFAC
B-le-Ol-4
Copyright © 1999 IFAC 14th Triennial World Congress, Beijing, P.R. China
DESIGN OF WALKING MACHINES - CONTROL ASPECTS Friedrich PfeifFer, J osef Steuer
LehTstuhl B fUr' Mechanik} TU-Afunchen Boltzmannstrape 15, D-85748 Car-ehing Fax: 089/28915213 Tel. 089/28915200 e-mail: [email protected]
Abstract: The dynamics and control of walking machines are presented. This paper focusses on a methodology in designing control and gives two realized examples: a six-legged and an eight-legged machine. ~opyrighl © 1999 IFAC Keywords: 'Walking, Robot Control, Control System Design.
1. INTRODUCTION
Design of walking machines includes configuration design, dynamics and control aspects and sensor and actuator performance (Pfeiffer et al., 1997). The design procedure itself may be model-based, or it may include some learning aspects. The author has long experience in model-based design, with much success, and therefore the paper shall focus on that aspect.
I Layout Des ign
I
Real Mochlne
Hordware in the Loop
Performance, Coofirmo:ions, improvements
a control concept can be evaluated. The concept is designed with the help of a reduced dynamics model and tested and improved with a full scale dynamic model (fig. 2).
I
I
Fig. 1. Realizing a Walking Machine Model-based means various iterative steps (fig. 1): choice of the walking machine's configuration: establishment of a simulation model including all realistic error sources like sensor noise, actuatortemperature beha"iour and the like: simulation of all important design data like loads and the choice of motors and gears; redesign of the concept and so on. On the basis of such a converging model
Fig. 2. Layout and Design With the outcome of such a model-based and iterative procedure a harctware-in-the-loop Lest is designed (fig. 3) to evaluate and then improve the drawbacks of the concepts. 897
Copyright 1999 IF AC
ISBN: 008 0432484
DESIGN OF WALKING MACHINES - CONTROL ASPECTS ...
Dynam,cs & Control
Ha rdvva re Componlens
1 I
14th World Congress ofTFAC
1
J
Simulation :.
Tests
I
I
I
! Hardvvare-in-the-Loop Agorthms
1
Fig. 5. The Six-Legged Wa.lking Machine "MAX"
Performance,
Loads Fig. 3. Hardware-in-the-Loop Tests As a final step follows the reali:mtion of the walking machine (fig. 1).
I
Simulation
err
I
1 Virtual Sensors
I
Real Sensors
1
I
Comparisons, Feedback to Design i
1 Machine Evaluation
Fig. 4. Machine Realization Control of walking IlIeans generat.ing one or more gait patternl-l, controlling these gait patterns even under unforeseen events like obstacles and to control the kinematics and the kinetics, the torques and forces of the legs. Furthermore it includes the control of contact forces on the ground, or in a tube (RoBmann and Pfeiffer, 1996: RoBmann; 1998). On a higher level walking means L'Cceiving piclur'es from the environment, evalua.ting these pictures and making deeisious with respect to the walking process. This paper shall focus on aspects of generating gaits, managing obstacles and controlling forces, torques and r.:ontacts.
2. CONTROL OF A SIX-LEGGED MACHINE The six-legged machine MAX (Fig. 5) has a weight of 23 kg a.nd a length of about 1 m.
The main ideas for the control system of this six-legged robot were ta.ken from neurobiological research with stick insects (Cruse, 1976; Eltze, 1994; Weidemann, 199;3). The technical realization follows in its performance very closely to biological principles. A global leg coordination module (LCM) is an information level where each leg informs its neighbouring legs about its state, influencing the decision functions of each single leg controller (SLC). The leg coordinat.ion module (LCM) is responsible for setting the landing and lifting poiuts of each leg (In the following AEP = anterior extreme. position and PEP = posterior extreme position). By cont1'01ling these points the goba\ behaviour of the walking process can be influenced. Although this level is doing a global ta.sk, the control mechanism works locally. In Figure 6 this mp.chanism is depicted. It can be seen, that neighbouring \egH can shift the AEPs and PEPs by small amounts. Thus legs can inhibit. adjoining legs from lifting of the ground in postponing their PEPs. Each leg gets specific informa.tion from the other legs namely the walking phase, the vp.locity and the AEP- and PEP-values. This information is sufficient for each LC:\-1 for computing its new AEP and PEP. These values are sent to the middle control level. There is no central supervision. The control inHuenccs used in this approach have heen measured and isolated by neurobiologists. Up to eight control mechanisms can be implemented in the LC.:vl, the principle of the two most important mechanisms numbered I and II are shortly explained in the following: Given that. the rostrally neighboured leg is not yet in STANCE phase, the machanism I inhibits the lifting of the leg in shifting back the PEP hy Cl. certain increment. Similarly, the mechanism II inhibits a. s1.art of the lifting phase when the contralateral adjacent leg is lIOt. yet. back in STA::\CE phase. The Single leg controller (SLC) is the heart of thp. leg motion performing all decisi.ons necessary to move the leg and to control the varioul-l phases. The order of the differp.nt phases in a normal 898
Copyright 1999 IF AC
ISBN: 008 0432484
DESIGN OF WALKING MACHINES - CONTROL ASPECTS ...
14th World Congress ofTFAC
Pt'lOses
f SLC
O___~
Fig. 6. Control Concept of the Six-Legged ¥lalking Machine "MAX" step is STANCE, PROTRACT, SWING and RETRACT (see Fig. 6). The SLC switches between the phases in dependency of the AEP, the PEP and some specific events (e.g. hitting an obstacle). It performs some on-line path planning at the beginning of the PROTRACT phase. Moreover the SLC adds some local intelligence which is needed especially for managing obstacles, impacts or other unforeseen events.
TRACT, SWiNG, RETRACT and RESWING) resembles a manipulator controller with on-line path planning. The controller for the AIR and STANCE phases differ in the controlled coordinates.
The single leg controller detects and surpasses obstacles, controls body height and corrects slippage effects. The capability of obstacle avoidance is achieved by means of a special detection mechanism and a different approach to general path planning. During SWING phase the SLC monitors the bending load in the leg segments. Whenever the corresponding strain gauge signa.l exceeds a certain threshold value the obstacle avoidance mechanism is activated. A short RES\IIlING phase is executed followed by a new S\VING phase trying to pass the obstacle.
Tube systems differ in their pipe diameters, , lengths, the mediums inside, the complexity of the tube arrangement etc. Different kinds of robots have been developed for inspecting and repairing tubes from the inside (Neubauer, 1994; Ro13mann, 1998). They are driven by wheels or chains or t.hey float with the medium. All types of robots have their specific difficulties, for example problems of traction or low flexibility and do not satisfy all requirements expected by the users.
In addition to the two upper levels the leg needs a low level.control system which typically, and again Hear the biological performance, consists in a feedforward non linear decoupling scheme combined with a feedback linear controller. The low level controller for the AIR phase (which indlldes PRO-
3. CONTROL OF AN EIGHT-LEGGED MACHINE
The aim of this project is the development of a robot moving forward by feet to study the possibilities and difficulties of legged locomotions in contrast t.o other systems. The higher flexibility of legs can be used to extend the technical possibilities of moving in tube systems (Fig. 7). The robot shown in Figure 7 has eight legs arranged like two stars. The attachments of the eight 899
Copyright 1999 IF AC
ISBN: 008 0432484
DESIGN OF WALKING MACHINES - CONTROL ASPECTS ...
14th World Congress ofTFAC
The rohot is controlled by five Siemens microcontrollers 80C167 CAN, which are installed on the crawler itself. One controller acts as a central unit. Each ofthe remaining four units controls t.wo opposite legs. The controllers are able to communicate over a CAN bus system. Each leg has two potentiometers for measuring the joint angles and two tachometer generators for the angular velocity of the motors. In order to determine the contact forces to the pipe a special lightweight sensor was developed. With its five axes it does not depend on the exact contact configuration. For future extensions the electronic architecture allows the implementation of further sensors like inclination meters. The presented control strnctnre enables the robot to move through straight and curved pipes independently of the position inside the tube or the inclination of the tube (from horizontal up to vertical pipes). Considering the experiences with the six-legged walking machine a structure was chosen which is divided into two hierarchical levels. The upper level encloses the mechanism of coordination. The lower level controls the position and forces (it executes operating functions). Based on this division it is possible to realize a function orientated structure and to leave the solution of problems to the concerned components. Local Coordination
Central Coordination
(Leg Plane I)
Local Operating Level (Leg Plane 2)
LegPI.n~ inStanv
Fig. 7. Construction of the Pipe Crawling Robot "MORITZ" legs are located in two planes that intersect at the longitudinal axis of the central body. These planes are called leg planes. Each leg has two active joints, which are driven by DC-motors. Their axes of rotation are orthogonal to the leg planes. This provides each leg with a full planar mobility. The leg is mounted on the central body with an additional passive joint, which allow small compensating movernents in nonnal direction. The crawler has a length of about 0.75 m and is able to work in pipes with a diameter of 60 - 70 cm. In each of the eight legs, the distance between the two active joints (hip and knee) is 15 cm and the length of the last leg segment (from knee to foot) is 17 cm. The highest possible torque of the hip joint is 78 Nm short term and 40 Nm permanent. The corresponding values of the knee are 78 Nm and 20 Nm. In stretched out position a leg is able to calTY 6.5 times its own weight (less than 2 kg) permanently and 12 times for short time operations.
Fig. 8. Level of COOl"dination Local Op€rating Level
Central Operating
(Leg Plane I ;n Stunce)
Level
Local Operating Level (Leg Plane 2 '" Stance)
Fig. 9. Operating Level The gait pattern influences the dependencies between the legs and thus affects the coordination 900
Copyright 1999 IF AC
ISBN: 008 0432484
DESIGN OF WALKING MACHINES - CONTROL ASPECTS ...
Central Body Controller
<..
···········x·········
14th World Congress ofTFAC
Single Leg Controllers
............................;.
T,
T,
Robot Motion Observer F
Fig. 10. Overall Control Concept for "MORITZ" which are estimated from the joint angles of the legs. This is done by changing the leg forces to achieve accclerations for correcting the control errors. For this purpose the local operating level is used_ It receives the corresponding setpoint commands. These commands must be created with respect to restrictions like satisfying the condition of sticking or the limitations of the eletrical and mechanical components. • The local operating level controls the applied forces during the contact phase and the motions of a single leg during the different air phases. In contrast to the last ones, which are pure local problems (legs without contact can be assumed as decoupled), the forces of legs touching the environment are strongly coupled and therefore a strictly local realization cannot consider all effects in each configuration. Therefore local means as local as possible.
and the control structure. llecause of the limited leg mobility, a load shift is only feasible from the legs of one leg plane to the legs of the other leg plane. This provides the crawler with full mobility in this plane. Three dimensional movements must be approximated by acting in orthogonal spaces. In other cases the crawler is able to move straight on only (except for special contact positions). The diagrams of figures 8 and 9 show the principles of the coordination level and the operating level for the load phase. • The central coordination level coordinates the phase characteristics of the two leg planes. Decisions on switching of the legs under load al'e ll13de by this cOlnponent. The legs do not have any autonomy with the advantage of higher safety from falling. With respect to this property the concept differs from other solutions (Weidemi-inn, 1993). Furthermore, the problems which can only be mastered by a reaction of the whole robot should be solved in this level (e.g. the legs of one plane can not find any contact). • The local coor'd'inat'ion level controls the step circle of a single leg, especially the sequence of leg motion phases (stance, protract, swing, retract). It also reacts to disturbances induced by small obstacles. • The central operating level controls the position and the velocity of the central body
Figure 10 depicts the overall concept for controlling ":MORITZ". For force control a feedback linearization procedure was applied coming out with six compensating torques To to T 5 , which act in the sense of a kind of feedforward decoupling for the walking process. The machine is equipped with force sensors, angular encoders, tachometers, and in addition the power consumption of the motors and thus the torques are measured. Neverthe-
901
Copyright 1999 IF AC
ISBN: 008 0432484
DESIGN OF WALKING MACHINES - CONTROL ASPECTS ...
14th World Congress ofTFAC
less measurements are not complete. Therefore, three observers generate additional informat.ion about friction in the gears, gravity influence and machine kinemat.ics. The system works without any problems.
4. REFERENCES Cruse, H. (1976). The function of the legs in the free walking stick insect, Carausius Morosus. Journal of Com.parative Physiology p. 112. Eltze, J. (1994). Biologisch orientierte Entwicklung einer sechsbeinigen Laufmaschine. VoL 110 of Fortschrittsberichte VDl, Reihe 17. VDI-Verlag. Diisseldorf. Neubaucr, W. (1994). A spider-like robot t.hat climbes vertically in ducts. In: Froc. of the 1994 IEEE/RSJ Int. Conf. on Intelligent Robots and Systems. Munich. pp. 1178-1185. Pfciffer, F. and H. Cruse (1994). Bionik des Laufens - technische Urnsetzung biologischcn Wissens. Konstruktion pp. 261-266. Pfeiffer, F., T. Roi3mann and J. Steuer (1997). Theory and practice of walking machines. In: Human and lVlachine Locomotion. CISM. RoBmann, T. (1998). Eine Laufmaschine fUr Rohre. Dissert.ation. Lehrstuhl B fiir :::Vlechanik, TU-Miinchen. RoBmann, T. and F. Pfeifl"er (1996). Control and design of a pipe crawling robot. In: Proc, of the 13th World Congr'css of Automatic Control (LF., Ed.). San Francisco, USA. Weidemann, H.-J. (1993), Dynamik und Regelung von scchsbeinigen RolJotern und natiirlichen Hexapoden. Vol. 362 of Fodschrittsberichte VD!, Reihe 8. VDI-Verlag. Diisseldorf.
902
Copyright 1999 IF AC
ISBN: 008 0432484
14th World Congress ofTFAC
B-le-Ol-4
Copyright © 1999 IFAC 14th Triennial World Congress, Beijing, P.R. China
DESIGN OF WALKING MACHINES - CONTROL ASPECTS Friedrich PfeifFer, J osef Steuer
LehTstuhl B fUr' Mechanik} TU-Afunchen Boltzmannstrape 15, D-85748 Car-ehing Fax: 089/28915213 Tel. 089/28915200 e-mail: [email protected]
Abstract: The dynamics and control of walking machines are presented. This paper focusses on a methodology in designing control and gives two realized examples: a six-legged and an eight-legged machine. ~opyrighl © 1999 IFAC Keywords: 'Walking, Robot Control, Control System Design.
1. INTRODUCTION
Design of walking machines includes configuration design, dynamics and control aspects and sensor and actuator performance (Pfeiffer et al., 1997). The design procedure itself may be model-based, or it may include some learning aspects. The author has long experience in model-based design, with much success, and therefore the paper shall focus on that aspect.
I Layout Des ign
I
Real Mochlne
Hordware in the Loop
Performance, Coofirmo:ions, improvements
a control concept can be evaluated. The concept is designed with the help of a reduced dynamics model and tested and improved with a full scale dynamic model (fig. 2).
I
I
Fig. 1. Realizing a Walking Machine Model-based means various iterative steps (fig. 1): choice of the walking machine's configuration: establishment of a simulation model including all realistic error sources like sensor noise, actuatortemperature beha"iour and the like: simulation of all important design data like loads and the choice of motors and gears; redesign of the concept and so on. On the basis of such a converging model
Fig. 2. Layout and Design With the outcome of such a model-based and iterative procedure a harctware-in-the-loop Lest is designed (fig. 3) to evaluate and then improve the drawbacks of the concepts. 897
Copyright 1999 IF AC
ISBN: 008 0432484
DESIGN OF WALKING MACHINES - CONTROL ASPECTS ...
Dynam,cs & Control
Ha rdvva re Componlens
1 I
14th World Congress ofTFAC
1
J
Simulation :.
Tests
I
I
I
! Hardvvare-in-the-Loop Agorthms
1
Fig. 5. The Six-Legged Wa.lking Machine "MAX"
Performance,
Loads Fig. 3. Hardware-in-the-Loop Tests As a final step follows the reali:mtion of the walking machine (fig. 1).
I
Simulation
err
I
1 Virtual Sensors
I
Real Sensors
1
I
Comparisons, Feedback to Design i
1 Machine Evaluation
Fig. 4. Machine Realization Control of walking IlIeans generat.ing one or more gait patternl-l, controlling these gait patterns even under unforeseen events like obstacles and to control the kinematics and the kinetics, the torques and forces of the legs. Furthermore it includes the control of contact forces on the ground, or in a tube (RoBmann and Pfeiffer, 1996: RoBmann; 1998). On a higher level walking means L'Cceiving piclur'es from the environment, evalua.ting these pictures and making deeisious with respect to the walking process. This paper shall focus on aspects of generating gaits, managing obstacles and controlling forces, torques and r.:ontacts.
2. CONTROL OF A SIX-LEGGED MACHINE The six-legged machine MAX (Fig. 5) has a weight of 23 kg a.nd a length of about 1 m.
The main ideas for the control system of this six-legged robot were ta.ken from neurobiological research with stick insects (Cruse, 1976; Eltze, 1994; Weidemann, 199;3). The technical realization follows in its performance very closely to biological principles. A global leg coordination module (LCM) is an information level where each leg informs its neighbouring legs about its state, influencing the decision functions of each single leg controller (SLC). The leg coordinat.ion module (LCM) is responsible for setting the landing and lifting poiuts of each leg (In the following AEP = anterior extreme. position and PEP = posterior extreme position). By cont1'01ling these points the goba\ behaviour of the walking process can be influenced. Although this level is doing a global ta.sk, the control mechanism works locally. In Figure 6 this mp.chanism is depicted. It can be seen, that neighbouring \egH can shift the AEPs and PEPs by small amounts. Thus legs can inhibit. adjoining legs from lifting of the ground in postponing their PEPs. Each leg gets specific informa.tion from the other legs namely the walking phase, the vp.locity and the AEP- and PEP-values. This information is sufficient for each LC:\-1 for computing its new AEP and PEP. These values are sent to the middle control level. There is no central supervision. The control inHuenccs used in this approach have heen measured and isolated by neurobiologists. Up to eight control mechanisms can be implemented in the LC.:vl, the principle of the two most important mechanisms numbered I and II are shortly explained in the following: Given that. the rostrally neighboured leg is not yet in STANCE phase, the machanism I inhibits the lifting of the leg in shifting back the PEP hy Cl. certain increment. Similarly, the mechanism II inhibits a. s1.art of the lifting phase when the contralateral adjacent leg is lIOt. yet. back in STA::\CE phase. The Single leg controller (SLC) is the heart of thp. leg motion performing all decisi.ons necessary to move the leg and to control the varioul-l phases. The order of the differp.nt phases in a normal 898
Copyright 1999 IF AC
ISBN: 008 0432484
DESIGN OF WALKING MACHINES - CONTROL ASPECTS ...
14th World Congress ofTFAC
Pt'lOses
f SLC
O___~
Fig. 6. Control Concept of the Six-Legged ¥lalking Machine "MAX" step is STANCE, PROTRACT, SWING and RETRACT (see Fig. 6). The SLC switches between the phases in dependency of the AEP, the PEP and some specific events (e.g. hitting an obstacle). It performs some on-line path planning at the beginning of the PROTRACT phase. Moreover the SLC adds some local intelligence which is needed especially for managing obstacles, impacts or other unforeseen events.
TRACT, SWiNG, RETRACT and RESWING) resembles a manipulator controller with on-line path planning. The controller for the AIR and STANCE phases differ in the controlled coordinates.
The single leg controller detects and surpasses obstacles, controls body height and corrects slippage effects. The capability of obstacle avoidance is achieved by means of a special detection mechanism and a different approach to general path planning. During SWING phase the SLC monitors the bending load in the leg segments. Whenever the corresponding strain gauge signa.l exceeds a certain threshold value the obstacle avoidance mechanism is activated. A short RES\IIlING phase is executed followed by a new S\VING phase trying to pass the obstacle.
Tube systems differ in their pipe diameters, , lengths, the mediums inside, the complexity of the tube arrangement etc. Different kinds of robots have been developed for inspecting and repairing tubes from the inside (Neubauer, 1994; Ro13mann, 1998). They are driven by wheels or chains or t.hey float with the medium. All types of robots have their specific difficulties, for example problems of traction or low flexibility and do not satisfy all requirements expected by the users.
In addition to the two upper levels the leg needs a low level.control system which typically, and again Hear the biological performance, consists in a feedforward non linear decoupling scheme combined with a feedback linear controller. The low level controller for the AIR phase (which indlldes PRO-
3. CONTROL OF AN EIGHT-LEGGED MACHINE
The aim of this project is the development of a robot moving forward by feet to study the possibilities and difficulties of legged locomotions in contrast t.o other systems. The higher flexibility of legs can be used to extend the technical possibilities of moving in tube systems (Fig. 7). The robot shown in Figure 7 has eight legs arranged like two stars. The attachments of the eight 899
Copyright 1999 IF AC
ISBN: 008 0432484
DESIGN OF WALKING MACHINES - CONTROL ASPECTS ...
14th World Congress ofTFAC
The rohot is controlled by five Siemens microcontrollers 80C167 CAN, which are installed on the crawler itself. One controller acts as a central unit. Each ofthe remaining four units controls t.wo opposite legs. The controllers are able to communicate over a CAN bus system. Each leg has two potentiometers for measuring the joint angles and two tachometer generators for the angular velocity of the motors. In order to determine the contact forces to the pipe a special lightweight sensor was developed. With its five axes it does not depend on the exact contact configuration. For future extensions the electronic architecture allows the implementation of further sensors like inclination meters. The presented control strnctnre enables the robot to move through straight and curved pipes independently of the position inside the tube or the inclination of the tube (from horizontal up to vertical pipes). Considering the experiences with the six-legged walking machine a structure was chosen which is divided into two hierarchical levels. The upper level encloses the mechanism of coordination. The lower level controls the position and forces (it executes operating functions). Based on this division it is possible to realize a function orientated structure and to leave the solution of problems to the concerned components. Local Coordination
Central Coordination
(Leg Plane I)
Local Operating Level (Leg Plane 2)
LegPI.n~ inStanv
Fig. 7. Construction of the Pipe Crawling Robot "MORITZ" legs are located in two planes that intersect at the longitudinal axis of the central body. These planes are called leg planes. Each leg has two active joints, which are driven by DC-motors. Their axes of rotation are orthogonal to the leg planes. This provides each leg with a full planar mobility. The leg is mounted on the central body with an additional passive joint, which allow small compensating movernents in nonnal direction. The crawler has a length of about 0.75 m and is able to work in pipes with a diameter of 60 - 70 cm. In each of the eight legs, the distance between the two active joints (hip and knee) is 15 cm and the length of the last leg segment (from knee to foot) is 17 cm. The highest possible torque of the hip joint is 78 Nm short term and 40 Nm permanent. The corresponding values of the knee are 78 Nm and 20 Nm. In stretched out position a leg is able to calTY 6.5 times its own weight (less than 2 kg) permanently and 12 times for short time operations.
Fig. 8. Level of COOl"dination Local Op€rating Level
Central Operating
(Leg Plane I ;n Stunce)
Level
Local Operating Level (Leg Plane 2 '" Stance)
Fig. 9. Operating Level The gait pattern influences the dependencies between the legs and thus affects the coordination 900
Copyright 1999 IF AC
ISBN: 008 0432484
DESIGN OF WALKING MACHINES - CONTROL ASPECTS ...
Central Body Controller
<..
···········x·········
14th World Congress ofTFAC
Single Leg Controllers
............................;.
T,
T,
Robot Motion Observer F
Fig. 10. Overall Control Concept for "MORITZ" which are estimated from the joint angles of the legs. This is done by changing the leg forces to achieve accclerations for correcting the control errors. For this purpose the local operating level is used_ It receives the corresponding setpoint commands. These commands must be created with respect to restrictions like satisfying the condition of sticking or the limitations of the eletrical and mechanical components. • The local operating level controls the applied forces during the contact phase and the motions of a single leg during the different air phases. In contrast to the last ones, which are pure local problems (legs without contact can be assumed as decoupled), the forces of legs touching the environment are strongly coupled and therefore a strictly local realization cannot consider all effects in each configuration. Therefore local means as local as possible.
and the control structure. llecause of the limited leg mobility, a load shift is only feasible from the legs of one leg plane to the legs of the other leg plane. This provides the crawler with full mobility in this plane. Three dimensional movements must be approximated by acting in orthogonal spaces. In other cases the crawler is able to move straight on only (except for special contact positions). The diagrams of figures 8 and 9 show the principles of the coordination level and the operating level for the load phase. • The central coordination level coordinates the phase characteristics of the two leg planes. Decisions on switching of the legs under load al'e ll13de by this cOlnponent. The legs do not have any autonomy with the advantage of higher safety from falling. With respect to this property the concept differs from other solutions (Weidemi-inn, 1993). Furthermore, the problems which can only be mastered by a reaction of the whole robot should be solved in this level (e.g. the legs of one plane can not find any contact). • The local coor'd'inat'ion level controls the step circle of a single leg, especially the sequence of leg motion phases (stance, protract, swing, retract). It also reacts to disturbances induced by small obstacles. • The central operating level controls the position and the velocity of the central body
Figure 10 depicts the overall concept for controlling ":MORITZ". For force control a feedback linearization procedure was applied coming out with six compensating torques To to T 5 , which act in the sense of a kind of feedforward decoupling for the walking process. The machine is equipped with force sensors, angular encoders, tachometers, and in addition the power consumption of the motors and thus the torques are measured. Neverthe-
901
Copyright 1999 IF AC
ISBN: 008 0432484
DESIGN OF WALKING MACHINES - CONTROL ASPECTS ...
14th World Congress ofTFAC
less measurements are not complete. Therefore, three observers generate additional informat.ion about friction in the gears, gravity influence and machine kinemat.ics. The system works without any problems.
4. REFERENCES Cruse, H. (1976). The function of the legs in the free walking stick insect, Carausius Morosus. Journal of Com.parative Physiology p. 112. Eltze, J. (1994). Biologisch orientierte Entwicklung einer sechsbeinigen Laufmaschine. VoL 110 of Fortschrittsberichte VDl, Reihe 17. VDI-Verlag. Diisseldorf. Neubaucr, W. (1994). A spider-like robot t.hat climbes vertically in ducts. In: Froc. of the 1994 IEEE/RSJ Int. Conf. on Intelligent Robots and Systems. Munich. pp. 1178-1185. Pfciffer, F. and H. Cruse (1994). Bionik des Laufens - technische Urnsetzung biologischcn Wissens. Konstruktion pp. 261-266. Pfeiffer, F., T. Roi3mann and J. Steuer (1997). Theory and practice of walking machines. In: Human and lVlachine Locomotion. CISM. RoBmann, T. (1998). Eine Laufmaschine fUr Rohre. Dissert.ation. Lehrstuhl B fiir :::Vlechanik, TU-Miinchen. RoBmann, T. and F. Pfeifl"er (1996). Control and design of a pipe crawling robot. In: Proc, of the 13th World Congr'css of Automatic Control (LF., Ed.). San Francisco, USA. Weidemann, H.-J. (1993), Dynamik und Regelung von scchsbeinigen RolJotern und natiirlichen Hexapoden. Vol. 362 of Fodschrittsberichte VD!, Reihe 8. VDI-Verlag. Diisseldorf.
902
Copyright 1999 IF AC
ISBN: 008 0432484