Manuel A. Armada received his Ph. D. in Physics from the University of Valladolid (Spain) in 1979. He has been involved since 1976 in research activities related to Automatic Control and Robotics. He has been working in more than 40 RTD projects (including international ones like EUREKA, ESPRIT, BRITE/EURAM, GROWTH, INCO-COPERNICUS and others abroad the EU, especially with Latin America (CYTED) and Russia (ESSIDAC EC project and Joint Projects between RAS-CSIC). He also carried out an important activity in several EC Thematic Networks, like CLAWAR and ROBMAR. Dr. Armada was the Chairman of the TELEMAN Evaluation Panel set up by the EC for evaluation of that Programme, and was one of the independent external reviewers for EURON Network of Excellence. Dr. Armada owns several patents, and has published over 250 papers (including contributions to several books, monographs, journals, international congresses and workshops). He is currently Vice director of the Instituto de Automatica Industrial of the Spanish Council for Scientific Research (CSIC), being his main research direction concentrated in robot design and control, with especial emphasis in new fields like flexible robots and walking and climbing machines. Dr. Armada has been very active, in the last fifteen years, in the preparation and management of RTD projects and has been also a reviewer for projects funded by the European Commission and by the Spanish, Italian, Netherlands and Argentinean administrations. Dr. Armada is member of the editorial board and reviewer of several international journals. Dr. Armada is presently the National Contact Person for Spain at the IARP (International Advanced Robotics Programme).
Abstract: An approach for achieving reliable, built-in, high-accuracy force
sensing for legged robots is presented, based on direct exploitation
of the properties of a robotâs mechanical structure. The proposed
methodology relies on taking account of force-sensing requirements
at the robot design stage, with a view to embedding force-sensing capability
within the mechanical structure of the robot itself. The test
case is ROBOCLIMBER, a bulky, quadruped climbing and walking
machine whose weighty legs enable it to carry out heavy-duty
drilling operations. The paper shows that, with finite-element analysis
of ROBOCLIMBERâs mechanical configuration during the design
stage, candidate positions can be selected for the placement of force
transducers to measure indirectly the contact forces between the feet
and the ground. Force sensors are then installed at the theoretically
best positions on the mechanical structure, and several experiments
are carried out to calibrate all sensors within their operational range
of interest. After calibration, the built-in sensors are subjected to experimental
performance evaluation, and the final best sensor option
is found. The built-in force-sensing capability thus implemented is
subjected to its first test of usability when it is employed to compute
the actual centre of gravity of ROBOCLIMBER. The method
is shown to be useful for determining variation during a gait (due
to the non-negligible weight of the legs). Afterwards the force sensors
are shown to be useful for controlling footâground interaction,
and several illustrative experiments confirm the high sensitivity, reliability
and accuracy of the selected approach. Lastly, the built-in
sensors are used to measure ground-reaction forces and to compute the zero-moment point for ROBOCLIMBER in real time, both while
standing and while executing a dynamically balanced gait.
Abstract: The Dual Smart Drive is a specially designed nonlinear actuator intended for use in climbing andwalking
legged robots. It features a continuously changing transmission ratio and dual properties and is very suitable for
situations where the same drive is required to perform two different types of start-stop motions of a mobile link.
Then, the associated control problem to this nonlinear actuator is established and a backstepping design strategy
adopted to develop Lyapunov-based nonlinear controllers that ensure asymptotic tracking of the desired laws of
motion, which have been properly selected using time-optimal control. The approach is extended for bounded
control inputs. Both simulation and experimental results are presented to show the effectiveness and feasibility of
the proposed nonlinear control methods for the Dual Smart Drive.
Abstract: In this paper we consider the postural stability problem for nonlinearly
actuated quasi-static biped robots, both with respect to the joint
angular positions and also with reference to the gripping effect between
the foot/feet against the ground during robot locomotion. Zero
moment point based mathematical models are developed to establish
a relationship between the robot state variables and the stability
margin of the foot (feet) contact surface and the supporting ground.
Then, in correspondence with the developed dynamical model and
its associated uncertainty, and in the presence of non-modeled robot
mechanical structure vibration modes, we propose a robust control
architecture that uses two cascade regulators. The overall robust
control system consists of a nonlinear robust variable structure controller
in an inner feedback loop for joint trajectory tracking, and
anHâ linear robust regulator in an outer, direct zero moment point
feedback loop to ensure the footâground contact stability. The effectiveness
of this cascade controller is evaluated using a simplified
prototype of a nonlinearly actuated biped robot in double support
placed on top of a one-degree-of-freedom mobile platform and subjected
to external disturbances. The achieved experimental results
have revealed that the simplified prototype is successfully stabilized.
Abstract: The perspective of using humanoid robots in practical applications is attracting an important research effort and the late steps forward in robot technology shows many remarkable achievements where design aspects, control systems and software evolution regarding humanoid machines have been realised. However, and although most of humanoid robots are intended to offer by some means a good degree of autonomy so that they could deploy all their intrinsic capabilities to perform âusefulâ tasks, the current state of this technology is being still far from allowing to attain such a goal. Among all the concerned features, and aiming to improve humanoid robot overall performance, it is required that they could work for
a long time spending minimum energy without losing their kinematical skills. In this
direction, a new kind of non-linear actuator, SMART, based on quasi-resonance principle, have been developed by the Industrial Automation Institute to improve the overall performance of biped locomotion. One of the major advantages provided by SMART is their inherent low energy consumption in comparison to classical transmission ratio actuators. However, due to their intrinsically non- linear characteristic, the practical use of these actuators poses some complications, in particular when it is necessary to realise advanced control schemes. To overcome this problem it is demonstrated how force sensing, implemented in one of the four-bar linkages of the mechanism, along with angular position measurement and motor current sensing, provides, throughout a sensor fusion strategy, an extra sensitivity to the non-linear actuator, resulting in an enhanced responsiveness. By using this approach the foreseen theoretical SMART actuator properties are fully demonstrated and experimental results are used to verify the practical advantages of the proposed method.
Abstract: In the last two decades in particular, climbing and walking robots
have been the subject of important research activity worldwide. However,
the practical use of these robots is still limited and only a few
are in actual use in live situations. In the general framework of the
CLAWAR Thematic Network, several working groups have been established
to formulate requirements, to define specifications and to
investigate those aspects of climbing and walking robot technology
that are more relevant with respect to selected application domains.
The aim of this paper is to present an overview of the investigations
carried out by the CLAWAR network, and to show various realizations
that could offer a good picture of how to rise above the barriers
to exploit this innovative class of robotic systems.
Abstract: Improving walking robots overall performance requires
that they could work for a long time spending minimum
energy without losing their kinematic skills. In this
paper the efficiency of locomotion mechanisms when
designing humanoid robots is considered, and it is
shown that proficient application of non-linear drives in
selected joints allows extending humanoid robot
functioning. A novel humanoid robot, SILO2, is
employed for the experimental confirmation of the
proposed approach .
Abstract: In this paper it is presented the undertaken research effort oriented to provide with self-tuning control capabilities to two walking robots : 1) a self propelling robot that is able to reach a desired remote location by means of special adhering devices attached to both ends of its kinematic chain; and, 2) a four legged locomotion robot. The digital control tasks are performed by specific hardware which is based on specialized microcontrollers what frees the host processor for implementing other tasks, like the plant transfer function identification and the digital filter tuning. On first place the hardware set-up is briefly introduced. It follows the identification sections showing how the several elements of the electromechanical chain (ZOH, amplifier, dc motor and load, encoder) are identified as a whole, in a very practical way, and considerations about stability of the algorithm, selection of the sampling rate, nonlinearities influence, ..., are made. In the third section it is shown how a pseudo deadbeat control is designed so that it suits to the digital filter hardwared implemented and then it is employed succesfully to the self-tuning of the different walking robots joints. Some experimental results are then reported.
Abstract: This paper describes the design and control concepts of a
wall-climbing robot. It has an hexapod configuration and
it is able to manoeuvre on vertical surfaces carrying high
payloads. Configuration and leg design criteria specific
for climbing tasks are discussed. The controller
architecture showing decentralised parallel control and
hard real-time performance is outlined. New stability
criteria for wall locomotion are introduced and a
climbing gait using force distribution shows the working
of our control scheme for wall gait generation. We call
this four phase discontinuous sawing gait. This prototype
is an example of a climber specifically tailored for
industrial applications.
