Personal Mobility with Assistive Walker

User Interface Design for Vehicle Mode

Masahiro Onozawa

, Sho Yokota

, Daisuke Chugo

and Hiroshi Hashimoto

Department of Mechanical Engineering, Toyo University, Kawagoe, Japan

School of Science and Technology, Kwansei Gakuin University, Sanda, Japan

Advanced Institute of Industrial Technology, Shinagawa, Japan

Keywords:

Personal Mobility, Assist, Walking, Electric Cart, Walker.

Abstract:

This paper presents Personal Mobility with assistive Walker (PM-W). PM-W is consists of handle, handle post,

position adjustable saddle, and two inwheel motors. The feature of PM-W is two mobile modes: ”walking

assist mode” and ”vehicle mode”. In ”walking assist mode”, PM-W assists user’s walking by driving two

wheels while user saddles on the seat and walks. In ”vehicle mode”, user controls PM-W like an electric bike

or conventional senior cart. Thus, a certain control interface is needed in ”vehicle mode”. Therefore, this

paper develops the interface for ”vehicle mode”. The proposed interface uses the bending strain at the root

of the handle post, because the handle should not have any moving parts on it such as throttle lever or stick

controller for supporting elderly’s stable posture. In order to verify the operability of this interface, the basic

experiment was conducted. From the results of the experiment, it is appropriate that the interface is adopted

for ”vehicle mode”, because it can be regarded that the operability of proposed interface is equal to or more

than the conventional interface.

1 INTRODUCTION

It is important to take a customary exercise for keep-

ing and promoting elderly’s health (Penedo, 2005). In

particular, the activities with going out leads to main-

tain both mental and physical health (Transportation

Research Board, 2005). However, there is the fact that

the weak of elderly’s physical ability makes the range

of activity narrow (Chou CH, 2012). Therefore, it is

desired that a certain support system expanding the

activity range with controlling the physical weakness.

There are some solutions for this problem. Honda

Walking Assist (Honda Motor Co., LTD., 2017) as-

sists to keep the user’s walking rhythm and stride

by power assisting for the swing of the legs while

measuring the hip joint movement. This system en-

ables user’s walking, however it requires pre-settings

or preparation for walking because this is the wear-

able power assists device. The other walking sup-

port system is the electric cart with pedaling unit (Jin-

hua She, 2013). The feature of the system is the pedal

unit which is the user interface of the electric cart,

which is one of the Personal Mobilities, and it is able

to control the load of it with adjusting user’s physi-

cal conditions. The rotational speed of the pedal is

used for controlling the velocity of the cart. User can

take exercise while moving by actuating the pedal.

In addition, this system adopts the electric cart be-

ing kinds of personal motilities. Thus it can expand

the activity range more than actual user physical abil-

ity, and can contribute to promote the activities with

going out. However, the exercise in this system is

different movement from natural human walking be-

cause it is only legs movements. It, therefore, is difﬁ-

cult to take whole body exercise like human walking.

And walking exercise is recommended for keeping el-

derlies health (American College of Sports Medicine,

2009).

On the other hand, the walker with a saddle

such as ”Raku-walk” (Kikuchi Seisakusho Co., LTD.,

2017), ”AR-5” (Ai Label Co., LTD., 2017) and

”KW200” (Kishi Engineering Co., LTD., 2017) have

been developed. These walkers do not use any ac-

tuators. The saddle in the walker partially supports

user’s weight, user can move forward by moving their

legs with sitting on it. Therefore it is expected to ex-

pand the user’s activities range with reducing phys-

ical load, because the leg strength is made full use

for moving forward not for supporting their weight.

However, it is difﬁcult to realize natural walking, be-

Onozawa, M., Yokota, S., Chugo, D. and Hashimoto, H.

Personal Mobility with Assistive Walker - User Interface Design for Vehicle Mode.

DOI: 10.5220/0006474604650470

In Proceedings of the 14th International Conference on Informatics in Control, Automation and Robotics (ICINCO 2017) - Volume 2, pages 465-470

ISBN: Not Available

465

700

[mm]

690~780

200

Figure 1: Overview of the PM-W.

cause user’s pelvis leans during walking and sitting on

these walking apparatus. It is also difﬁcult to make an

active power assist and remarkable expansion of ac-

tivity range, because they do not have any actuators.

For this problem, we think that the problem can

be solved by compensating each weak points in above

mentioned walker and electric cart by combining the

electric mobile function of PM and walker’s exercise

function. This research, therefore, proposes ”Per-

sonal mobility with assistive walker: PM-W” which

has two mobile modes being ”walking assist mods”

and ”vehicle mode”. Here, ”walking assist mode”

assists user’s walking exercise, and ”Vehicle mode”

offers driving electric cart experience for users. In

particular, this paper introduces the designed proto-

type of PM-W and studies the interface in the ”vehicle

mode”.

2 OVERVIEW OF PM-W

The prototype of PM-W is shown in Fig. 1. PM-

W is consists of handle, handle post, position ad-

justable saddle, and two inwheel motors. And PM-

W is conventional two wheeled system shown in Fig.

2. In this ﬁgure, W[m] means width of PM-W and

R[m] is radius of wheel. The control inputs of the

system are Velecity v

ref

[m/s] and angular velocity

ref

[rad/s], and relationship between the wheel rota-

tional speeds ω

, ω

[rad/s] and control inputs are de-

signed by (1)(2).

ref

R(ω

+ ω

)

(1)

ref

R(ω

− ω

)

(2)

From (1) and (2), wheel speeds realizing v

ref

, ω

ref

are described as:

ref

+ ω

ref

(3)

ref

− ω

ref

. (4)

Thus, the role of the developed interface is to gen-

erate the control inputs v

ref

, ω

ref

from user’s con-

trol intention, and transmits these inputs to the servo

driver of PM-W. Here the speciﬁcation of PM-W is

summarized in Table. 1.

ref

Figure 2: The coordinate system of the PM-W.

Table 1: Speciﬁcations of the PM-W.

Width 700 mm

Wheel Radius 4 inch

Weight capacity 80 kg

Actuator DC brushless motor (90W x 2)

Power source Ni-MH battery (22.2[V] 4.0[Ahr])

2.1 Two Mobile Modes

The feature of PM-W is that it has two mobile modes:

”walking assist mode” and ”vehicle mode” shown in

Fig. 3.

(a) Walking Assists mode

(b) Vehicle mode

Figure 3: Vehicle mode and Walking assist mode.

2.1.1 Walking Assist Mode

In ”walking assists mode”, the height of the saddle is

adjusted to the user’s hip position in an upright stance.

The saddle partially support user’s weight in the up-

right stance. User walks with grasping the handle. At

this time, the handle functions as a handrail to support

user posture not to fall down while walking. Here, the

relative position of the saddle and the footrest was de-

ICINCO 2017 - 14th International Conference on Informatics in Control, Automation and Robotics

466

designed so as not to inhibit natural user’s walking by

experimentally investigating the stride. Because there

is a possibility of collision between the footrest and

user’s foots, if the relative position between saddle

and footrest is not appropriately designed. In addi-

tion, the handle and handle post were rigidly mounted

on the body frame of PM-W without any moving

parts, because the handle is important part for user’s

safety for preventing from falling down. The system

control the motor speed and assists user’s walk with

adapting users movements. In addition, by adjusting

motor torque, the assists level can be adopted to user’s

physical ability. From the above, user can exercises

while he/she moves with meeting own physical abil-

ity.

2.1.2 Vehicle Mode

In the ”vehicle mode”, the saddle position is lowered,

and user completely sits on the saddle. Then user

places own foots on the front section of PW-W as

a foot rest. And user can move while two inwheel

motors are controlled by the developed interface. By

this mode, PM-W can be used as conventionalelectric

cart, therefore, user can move and easily expand own

activity range.

By switching above mentioned two mobile modes

appropriately, user can exercise with moving and can

expand own activity range.

2.2 Lifting Saddle Mechanism

The height of the saddle is different in these two

modes. In the ”Walking assist mode”, the saddle po-

sition should be high and should have enough gap to

foot rest depicted in Fig. 4 to realize natural walk-

ing. On the other hand, in ”Vehicle mode”, the sad-

dle position should be low and close to foot rest. To

do so, the saddle needs not only lifting function but

also adjusting function in a back and forth position.

In order to realize this movement by one actuator, the

lifting and sliding mechanism using electric cylinder

and parallel links is adopted shown in Fig.6. Thanks

to this mechanism, the saddle can be adjusted in both

height and back/forth position by one actuator.

3 USER INTERFACE FOR

VEHICLE MODE

The general user interface represented by a stick con-

troller or a throttle lever has the mechanical moving

part, and user operates a machines by adjust its range.

High

Long

Figure 4: Saddle height at

Walking Assist mode.

Short

low

Figure 5: Saddle height at

Vehicle mode.

Parallel link

Trajectory of the saddle

Figure 6: Lifting saddle mechanism.

In our PM-W, the contacting point between user

and the machine is the handle. Therefore the han-

dle should also be utilized as the interface in ”vehicle

mode”. However, the handle is used as a handrail for

preventing from falling down or supporting elderlies

stable posture. Thus, the handle should not have the

moving part on it such as throttle lever or stick con-

troller.

3.1 Input Information to Interface

When user grasps the handle and adds forces on it,

the small strain occurs in the handle post (the handle

is shown in Fig. 1). This system adopts the strains

around the handle post as an interface inputs. By this

method, it will be possible to obtain user’s control in-

tention form the rigid part.

Personal Mobility with Assistive Walker - User Interface Design for Vehicle Mode

467

3.2 FEM at Post of Handle Bar

In order to realize the interface using strains, the at-

taching portion of strain gauges is studied by Finite

Element Method. The purpose of this FEM is as:

• Selection of the attaching part for controlling for-

ward and backward movement of PM-W

• Selection of the attaching part for steering control

of PM-W.

Here, there are two types of strain being bending and

twisting. It is natural to use the vertical bending strain

being generated when the user puts the handle in for-

ward and backward for controlling forward and back-

ward movements.

The candidates for steering control are following

two parts:

• The bending strain in left and right direction,

when user applies the forces on the handle in lat-

eral direction to realize steering control of PM-W

• The twisting strain, when user twists handle to re-

alize steering control of PM-W.

In these two strains, the part showing big strain is

preferable to be used for inputs to the interface even

if small forces are applied, because the resolution of

control forces from user becomes high. For this rea-

son, the part showing big strain is used for steering

control of PM-W.

FEM is conducted under following conditions: the

material handle post is aluminum A6063, the hollow

pipe (inner diameter 24mm, outer diameter 28mm).

the following three kinds of forces are applied to the

handle.

FEM1 The bending strain when the forces are ap-

plied to the both endpoints of handle parallel to y

axis in Fig. 7. ( forward / backward operation )

FEM2 The bending strain when the forces are ap-

plied to the both endpoints of handle parallel to x

axis in Fig. 8. ( Steering operation )

FEM3 The twistin strain when the forces are applied

to the both endpoints of handle parallel to y axis

in Fig. 9 in oposit direction each other. ( Steering

operation )

The result of FEM1 is shown in 7. From this ﬁg-

ure, it is turned out that vertical bending strain is gen-

erated at root of handle post by applying forward /

backward forces to both endpoints of handle. Thus,

the vertical strain at the root of the handle post can be

used for forward / backward control of PM-W.

The result of FEM2 and FEM are shown in Fig.

8 and Fig. 9 respectively. Form these ﬁgures, it is

turned out that both bending and twisting strain are

generated at the root of the handle post. These strains

are summarized in Table 2. In this table, FEM2 shows

bigger strain than FEM3. From this thing, the hor-

izontal bending strain is more suitable than twisting

strain for the steering control of PM-W.

STRAIN

Figure 7: Bending strain at the root of handle post by for-

ward force on the handle.

STRAIN

Figure 8: Bending strain at the root of handle post by lateral

force on the handle.

STRAIN

Figure 9: Bending strain at the root of handle post by twist-

ing force on the handle.

ICINCO 2017 - 14th International Conference on Informatics in Control, Automation and Robotics

468

Table 2: Rsults of FEMs.

Drection of applied force Strain

FEM1 Forward 1.09× 19

−4

FEM2 Right 1.48× 19

−4

FEM3 Twisting Right 9.72× 19

−5

From the above, this system adopts the bending

strains in horizontal and vertical direction at the root

of handle post as the inputs for the interface in ”vehi-

cle mode”.

3.3 Sequence of the Control

The vertical bending strain is used for forward / back-

ward control. The horizontal bending strain is used

for steering control. Here, the vertical strain is ex-

pressed by ε

, and the horizontal strain is shown in

. The both initial strains are expressed by ε

, ε

ω0

The control inputs v

ref

(1),(2) are generated based

on these strain as follows:



ref





0 a



− ε

ω0



(5)

Here, a

, a

are conversion coefﬁcients from strain

to control inputs, and their values are set by cut & try.

The total system conﬁguration including the interface

is shown in Fig. 10. The bending strains at the root

of the handle post are measured by the strain gauges,

and measured strains are ampliﬁed and are sent to the

Aruduino Mega. Aruduino Mega generates the con-

trol inputs v

ref

, ω

ref

by using (5), and calculates the

each motor speed ω

, ω

by using (3) and (4). The mo-

tor drivers (Hibot, 1BLDC PowerModule

) receives

, ω

, and then the motors are driven and PM-W is

activated. The open loop control is adopted for mo-

tors.

Baery

Strain Gauge Amp.

( )

Strain Gauge Amp.

( )

Arduino Mega

2560

Wheel (Le)

Brushless DC Motor

Motor driver

Strain

Gauges

Wheel (Right)

Brushless DC Motorω

hall sensor

Figure 10: System conﬁguration of the mobile platform.

4 TEST RUN

In the ”Vehicle mode”, since the interface using mov-

ing parts cannot be used for user safety, this system

adopts the bending strains at the root of the handle

post. In order to verify the operability of this inter-

face, the basic experiment was conducted. In this

experiment, we prepared the stick controller as the

control to proposed interface. The stick controller is

widely used for electric wheelchair and man-machine

interface.

The subject operated PM-W by using two inter-

faces along with the course shown in Fig. 11. At

that time, the lap times by two interfaces were mea-

sured. If the lap time by the proposed interface would

be same as it by the stick controller, it could be said

that the proposed interface has equivalent operability

to the conventional interface.

1.5䡉

5䡉

Start

Finish

Figure 11: Shape of a course

Figure 11: Shape of a course.

In this experiment, 10 subjects were recruited,

their age are 20s. They were lectured on how to drive

the PM-W in advance, and were given 3 minute prac-

tice for each interface. After that, the lap time was

measured.

䡉

Figure 12: Lap time by Proposed and Control interface

Figure 12: Lap time by Proposed and Control interface.

Time [s]

-1

-0.5

0.5

0.05

0.1

0.15

0.2

0 2 4 6 8 10 12 14 16 18 20

Interface by Strain

-1

-0.5

0.5

-0.05

0.05

0.1

0.15

0.2

0 2 4 6 8 10 12 14 16 18 20

S!ck Controller

ref

Figure 13: Transition of Control Inputs of both inputs

Figure 13: Transition of Control Inputs of both inputs.

Personal Mobility with Assistive Walker - User Interface Design for Vehicle Mode

469

Fig. 12 shows the average of lap time by each

interface. t test of Fig. 12 shows p = 0.17 > 0.05,

therefore there is no signiﬁcant difference between

two lap times. Thus, it can be said that proposed in-

terface has equivalent operability to the conventional

interface. In addition, let us focus on the transition

of the control inputs during experiment in both inter-

face shown in Fig. 13. From this ﬁgure, it is observed

that the control inputs by the stick controller are os-

cillated while the control inputs by propose interface

smoothly changed. From this oscillation, the stick

controller needs ﬁne correction operation. In particu-

lar, transition of v

ref

is largely oscillated. This means

the velocity of the PM-W is also oscillated. This kinds

of motion cannot provide comfortable ride or easy op-

eration.

Thus, from the viewpointsof lap time and smooth-

ness of the control inputs, it is turned out that the pro-

posed interface provides smooth operation compared

to the conventional interface though there is no dif-

ferences between the lap times by the proposed and

conventional interface.

From the above, it is appropriate that the interface

using strain at the root of handle post is adopted for

”vehicle mode”, because it can be regarded that the

operability of proposed interface is equal to or more

than the conventional interface.

5 CONCLUSION

This paper presented Personal Mobility with assis-

tive Walker (PM-W). PM-W has two mobile modes:

”walking assist mode” and ”vehicle mode”. In ”walk-

ing assist mode”, PM-W assists user’s walking by

driving two wheels while user saddles on the seat and

walks. In ”vehicle mode”, user controls PM-W like

an electric bike or conventional senior cart. There-

for a certain control interface is needed in ”vehicle

mode”. Thus, this paper develops the interface for

”vehicle mode”.

The main contact point between user and PM-W

is the handle. The handle is used as a handrail for

elderlies safe in both modes. Therefore the handle

should not have any moving parts on it such as throttle

lever or stick controller. From this things, the bend-

ing strain at the root of the handle post is focused on

for the interface. When user applies own forces to the

handle, the handle post has a strain. In order to realize

the interface using bending strain at the handle post,

FEM was conducted to ﬁnd the suitable strain part for

the interface. As a result, the vertical and horizontal

strain at the root of the handle post is used for the in-

puts of the interface. In order to verify the operability

of this interface, the basic experiment was conducted.

From the results of the experiment, it is appropriate

that the interface using strain at the root of handle

post is adopted for ”vehicle mode”, because it can be

regarded that the operability of proposed interface is

equal to or more than the conventional interface.

As the future works, we will develop the control

scheme for walking assists mode including the stabil-

ity of the user.

REFERENCES

Ai Label Co., LTD. (2017). Ar-5. In http://www.ailabel.org

(Last accessed on April 25 2017.).

American College of Sports Medicine (2009). Exercise and

physical activity for older adults. In Medicine & Sci-

ence in Sports & Exercise.

Chou CH, Hwang CL, W. Y. (2012). Effect of exercise on

physical function, daily living activities, and quality

of life in the frail older adults: a meta-analysis. In

Arch Phys Med Rehabil., pages 237–44.

Honda Motor Co., LTD. (2017). Honda walking assist de-

vice. In http://world.honda.com/Walking-Assist/ (Last

accessed on June 6 2017.).

Jinhua She, Sho Yokota, Y. E. D. (2013). Automatic heart-

rate-based selection of pedal load and control system

for electric cart. In Mechatronics, volume 23, pages

279–288. Elsevier.

Kikuchi Seisakusho Co., LTD. (2017). Raku-walk. In

http://www.kikuchiseisakusho.co.jp/ mechatro2/ im-

ages/MED

RO-03L.pdf (Last accessed on April 25

2017.).

Kishi Engineering Co., LTD. (2017). Walker with saddle :

Kw200. In http://www.kishieng.co.jp/product/walker/

(Last accessed on April 25 2017.).

Penedo, Frank Ja; Dahn, J. R. (2005). Exercise and well-

being: a review of mental and physical health beneﬁts

associated with physical activity. In Current Opinion

in Psychiatry, volume 18, pages 189–193.

Transportation Research Board (2005). Trb special report

does the built environment inﬂuence physical activ-

ity? examining the evidence 282. In Transportation

Research Board Institute of Medicine of The National

Academies, pages 19–22.

ICINCO 2017 - 14th International Conference on Informatics in Control, Automation and Robotics

470