As shown in Fig. 7, we set five motion primitives for
compliance motion of the mobile platform for walking
assistance. The compliance model (2) is used to transform
detected force/torque to velocity to assist user walking.
where r M , r I are the parameters of nominal mass and
inertia, and r B is the nominal damping coefficient. These
values are selected according to the user’s preference.
B. Obstacle avoidance controller
To prevent from colliding with objects in the environment
during walking assist, we add an obstacle avoidance function
to the mobile platform. The distance data are provided from
laser range finder equipped on the front side of the robot. We
define nine scanning ranges. Each range covers 20 degree.
The segmentation of laser sensing area is shown in Fig.8. We
set the robot heading in range 4 to range 6 and use the judge
method shown in (3) to calculate the angle velocity of
obstacle avoidance. The distance data of range 3 and range7
are used to judge the direction of turning. As shown in (4), the
angular velocity of the mobile robot is determined by the
acquired distance data from laser scanner and the amount of
turning is related to the distance and heading relative to the
obstacle. Note that only the angular velocity is controlled in
obstacle avoidance, the linear velocity, which is determined
by the compliance controller of the robot, is not influenced.