15-09-2016, 02:26 PM
1454647686-IEEEPaper2013.pdf (Size: 1.37 MB / Downloads: 5)
Abstract—An intelligent cane robot is designed for aiding the
elderly and handicapped people’s walking. The robot consists of
a stick, a group of sensors, and an omnidirectional basis driven
by three Swedish wheels. Recognizing the user’s walking intention
plays an important role in the motion control of our cane robot.
To quantitatively describe the user’s walking intention, a concept
called “intentional direction (ITD)” is proposed. Both the state
model and the observation model of ITD are obtained by enumerating
the possible walking modes and analyzing the relationship
between the human–robot interaction force and the walking intention.
From these two models, the user’s walking intention can
be online inferred using the Kalman filtering technique. Based on
the estimated intention, a new admittance motion control scheme
is proposed for the cane robot. Walking experiments aided by the
cane robot on a flat ground and slope are carried out to validate the
proposed control approach. The experimental results show that the
user feels more natural and comfortable when our intention-based
admittance control is applied
INTRODUCTION
AS many countries step into the aging society rapidly, more
and more elders suffer from deficits of motor function or
disability of the limbs, which are usually caused by neurological
problems or lack of muscle strength. In addition, the growing
elderly population causes the shortage of people for nursing
care. Therefore, there is a great need to develop rehabilitation
robots that can partially replace the nurses and the therapists.
Currently, plentiful studies on rehabilitation robots can be found,
including the applications for the upper limb [1]–[5], for the
lower limb [6]–[9], and for the assisting or training of the whole
body [10]–[13].
In daily life, the walking is one of the most important human
activities. To improve the walking ability of the elderly the walker-type rehabilitation robot has become a popular research
topic over the last decade. There have been many intelligent
walker-type robots comprising active or passive wheels
and supporting frame. A novel assistive robotic walker called
“JAIST active robotic walker (JARoW)” to provide potential
users with sufficient ambulatory capability in an efficient, costeffective
way was presented in [14]. The Hitomi system to help
the blind in outdoor environment was proposed in [15]. A powerassisted
walker for physical support during walking was developed
in [16]. The Care-O-bot and Nursebot are developed as
personal service robots for elderly and disables [17], [18]. The
personal aid for mobility and monitoring (PAMM) system to
provide mobility assistance and user health status monitoring
was proposed in [19]. A new intelligent walker based on passive
robotics to assist the elderly, handicapped people, and the
blind was proposed in [20].
There are still many deficiencies in the present walker systems.
First, many walkers are designed for the indoor environment.
Second, most of them are big in size and/or heavy
in weight. An indoor robot is often restricted within limited
places. Big size makes it impossible to be used in narrow space
and heavy weight restricts the maneuverability. Many elders and
patients are not so weak that they have to be nursed carefully.
Nevertheless, sufficient support, such as a cane or stick, is necessary
to help them take a walk outside, which enables them
to realize high-quality lives or accelerate the rehabilitation. In
these cases, an intelligent cane system may be more useful than
walkers due to its flexibility and handiness. In [19], a SmartCane
system is also proposed, which has a relative smaller
size and nonholonomic constraint in kinematics. The nonholonomic
constraint is useful for moving along a path stably, but
reduces the maneuverability of the system. In [21] and [22], the
“GuideCane” and a robotic cane “Roji” are proposed for blind
or visually impaired pedestrians to navigate safely and quickly
through obstacles and other hazards. In the living environment,
including the narrow space, the cane system is expected to be
movable in omnidirections. Thus, omnidirectional mobile platform
is needed in the robot design. This kind of platform has
been considered in some applications [23], [24], whereas their
designs are special and not commercially available. Particularly,
they are proposed for walker systems but cane systems, which
are much smaller in size. Recently, commercial omniwheels
are applied in the area of walker systems [25]. The problem
that slender rollers of omniwheels have limited load capacities
is partly solved by the modern technology. In addition, a
small omnidirectional platform can be constructed by this kind
of wheels. In our previous study, an intelligent cane system
was designed based on a commercially available three-wheeled omnidirectional platform [26]. We also investigated the falldetection
and fall-prevention function of the cane robot systems
in [27].
The recognition of user’s walking intention plays an important
role in the study of the walker-type rehabilitation robots.
From the viewpoint of the control system of robot, the walking
intention provides a real-time reference trajectory for the robot
motion controller. Therefore, the more accurately the walking
intention is inferred, the more satisfactory the control performance
of the robot may be obtained. A manufacturing system
that is controlled based on the human intention/desire was proposed
in [28]. The dance partner robot, which estimates the intention
of human dancer, was proposed in [29] and [30]. When
we pay attention to the walker-type walking support system,
similar researches can be found in [31] and [32]. In the study
of motion–intention-recognizing approaches, the EMG-based
methods are widely applied [2], [3]. However, the EMG signals
are easily influenced by the location of electrodes, the thickness
of fattiness, the body temperature, and the perspiration. Meanwhile,
the information of the EMG signals is so large that a
complicated preprocessing procedure is required before using
them as the control input.
In this paper, we improved our former intelligent cane system
and studied its motion control problems in several situations
based on estimating human walking intention. The human walking
behavior is described by switching walking modes. To model
the human walking intention, an important concept called “intentional
direction” (ITD) is proposed, as well as its dynamic
model during human walking. Without knowing the ITD accurately,
it is not an easy task to design a motion controller of the
cane robot for an elderly or a handicapped user. Normally, these
people cannot walk along their ITD clearly and smoothly due
to their weak or handicapped lower limbs. For instance, even an
elderly intends to walk straightforwardly, he/she might finally
walk along a zigzag trajectory because of stumbling. Therefore,
the interactive forces measured by the cane robot consist of plentiful
user’s unintentional walking information, which is part of
the observation noise in the dynamic model of ITD. Comparing
with a young healthy subject, apparently this observation
noise of ITD is much bigger. Thus, it is necessary to pick up
the ITD as accurately as possible from the noisy measurement.
After that we may design a robot motion controller based on
the ITD to aid the user’s walking in accordance with his actual
walking intention. Some filtering technologies are used to
online estimate the ITD, based on which a new force control
scheme called “intention-based admittance control (IBAC)” is
proposed to provide a natural and intuitive interface for elderly
users.